<?xml version="1.0" encoding="utf-8" ?>
<!DOCTYPE chapter SYSTEM "chapter.dtd">
<chapter>
<header>
<copyright>
<year>2004</year><year>2016</year>
<holder>Ericsson AB. All Rights Reserved.</holder>
</copyright>
<legalnotice>
Licensed under the Apache License, Version 2.0 (the "License");
you may not use this file except in compliance with the License.
You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software
distributed under the License is distributed on an "AS IS" BASIS,
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
See the License for the specific language governing permissions and
limitations under the License.
</legalnotice>
<title>ERTS Release Notes</title>
<prepared>otp_appnotes</prepared>
<docno>nil</docno>
<date>nil</date>
<rev>nil</rev>
<file>notes.xml</file>
</header>
<p>This document describes the changes made to the ERTS application.</p>
<section><title>Erts 7.3.1</title>
<section><title>Fixed Bugs and Malfunctions</title>
<list>
<item>
<p>
<c>process_info(Pid, last_calls)</c> did not work for
<c>Pid /= self()</c>.</p>
<p>
Own Id: OTP-13418</p>
</item>
<item>
<p>
Make sure to create a crash dump when running out of
memory. This was accidentally removed in the erts-7.3
release.</p>
<p>
Own Id: OTP-13419</p>
</item>
<item>
<p>
Schedulers could be woken by a premature timeout on
Linux. This premature wakeup was however harmless.</p>
<p>
Own Id: OTP-13420</p>
</item>
<item>
<p>
A process communicating with a port via one of the
<c>erlang:port_*</c> BIFs could potentially end up in an
inconsistent state if the port terminated during the
communication. When this occurred the process could later
block in a <c>receive</c> even though it had messages
matching in its message queue.</p>
<p>
This bug was introduced in erts version 5.10 (OTP R16A).</p>
<p>
Own Id: OTP-13424 Aux Id: OTP-10336 </p>
</item>
<item>
<p>
The reference count of a process structure could under
rare circumstances be erroneously managed. When this
happened invalid memory accesses occurred.</p>
<p>
Own Id: OTP-13446</p>
</item>
<item>
<p>
Fix race between <c>process_flag(trap_exit,true)</c> and
a received exit signal.</p>
<p>
A process could terminate due to exit signal even though
<c>process_flag(trap_exit,true)</c> had returned. A very
specific timing between call to <c>process_flag/2</c> and
exit signal from another scheduler was required for this
to happen.</p>
<p>
Own Id: OTP-13452</p>
</item>
</list>
</section>
</section>
<section><title>Erts 7.3</title>
<section><title>Fixed Bugs and Malfunctions</title>
<list>
<item>
<p>
The '-path' flag to 'erl' has been documented. This flag
replaces the path specified in the boot script. It has
always existed, but was earlier only documented in SASL
(script).</p>
<p>
Own Id: OTP-13060</p>
</item>
<item>
<p>
The <c>call_time</c> tracing functionality internally
used a time based on OS system time in order to measure
call time which could cause erroneous results if OS
system time was changed during tracing.</p>
<p>
This functionality now use Erlang monotonic time in order
to measure time. Besides fixing the erroneous results due
to OS system time being used, the results are often also
better since Erlang monotonic time often has better
accuracy and precision.</p>
<p>
Own Id: OTP-13216</p>
</item>
<item>
<p>
Fix behaviour of -delay_write command line switch of
epmd, which is used for debugging - in some cases epmd
was sleeping twice the requested amount of time.</p>
<p>
Own Id: OTP-13220</p>
</item>
<item>
<p>
Fix race between timeout and exit signal that could cause
a process to ignore the exit signal and continue
execution. Bug exist since OTP 18.0.</p>
<p>
Own Id: OTP-13245</p>
</item>
<item>
<p>
Fix bug in <c>erlang:halt/1,2</c> for large exit status
values, causing either <c>badarg</c> (on 32-bit) or exit
with a crash dump and/or core dump (on 64-bit). Make
<c>erlang:halt/1,2</c> tolerate any non negative integer
as exit status and truncate high order bits if the OS
does not support it.</p>
<p>
Own Id: OTP-13251 Aux Id: ERL-49 </p>
</item>
<item>
<p>
<seealso
marker="kernel:gen_tcp#accept/2"><c>gen_tcp:accept/2</c></seealso>
was not <seealso
marker="erts:time_correction#Time_Warp_Safe_Code">time
warp safe</seealso>. This since it used the same time as
returned by <seealso
marker="erts:erlang#now/0"><c>erlang:now/0</c></seealso>
when calculating timeout. This has now been fixed.</p>
<p>
Own Id: OTP-13254 Aux Id: OTP-11997, OTP-13222 </p>
</item>
<item>
<p>
Fix faulty error handling when writing to a compressed
file.</p>
<p>
Own Id: OTP-13270</p>
</item>
<item>
<p>
Fix sendfile usage for large files on FreeBSD</p>
<p>
Own Id: OTP-13271</p>
</item>
<item>
<p>
Fix bug that could cause
<c>process_info(P,current_location)</c> to crash emulator
for hipe compiled modules.</p>
<p>
Own Id: OTP-13282 Aux Id: ERL-79 </p>
</item>
<item>
<p>
Out of memory errors have been changed to cause an exit
instead of abort.</p>
<p>
Own Id: OTP-13292</p>
</item>
<item>
<p>
When calling <c>garbage_collect/[1,2]</c> or
<c>check_process_code/[2,3]</c> from a process with a
higher priority than the priority of the process operated
on, the run queues could end up in an inconsistent state.
This bug has now been fixed.</p>
<p>
Own Id: OTP-13298 Aux Id: OTP-11388 </p>
</item>
<item>
<p>
A workaround for an issue with older gcc versions (less
than 5) and inline assembly on 32-bit x86 caused an
emulator crash when it had been compiled with a newer gcc
version. An improved <c>configure</c> test, run when
building OTP, now detects whether the workaround should
be used or not.</p>
<p>
Own Id: OTP-13326 Aux Id: ERL-80 </p>
</item>
</list>
</section>
<section><title>Improvements and New Features</title>
<list>
<item>
<p>Introduced new statistics functionality in order to
more efficiently retrieve information about run able and
active processes and ports. For more information see:</p>
<list> <item><seealso
marker="erlang#statistics_total_run_queue_lengths"><c>statistics(total_run_queue_lengths)</c></seealso></item>
<item><seealso
marker="erlang#statistics_run_queue_lengths"><c>statistics(run_queue_lengths)</c></seealso></item>
<item><seealso
marker="erlang#statistics_total_active_tasks"><c>statistics(total_active_tasks)</c></seealso></item>
<item><seealso
marker="erlang#statistics_active_tasks"><c>statistics(active_tasks)</c></seealso></item>
</list>
<p>
Own Id: OTP-13201</p>
</item>
<item>
<p>
Time warp safety improvements.</p>
<p>
Introduced the options <c>monotonic_timestamp</c>, and
<c>strict_monotonic_timestamp</c> to the trace,
sequential trace, and system profile functionality. This
since the already existing <c>timestamp</c> option is not
time warp safe.</p>
<p>
Introduced the option <c>safe_fixed_monotonic_time</c> to
<c>ets:info/2</c> and <c>dets:info/2</c>. This since the
already existing <c>safe_fixed</c> option is not time
warp safe.</p>
<p>
Own Id: OTP-13222 Aux Id: OTP-11997 </p>
</item>
<item>
<p>
Fix a register race where down nodes goes undetected in
epmd</p>
<p>
Own Id: OTP-13301</p>
</item>
<item>
<p>
Improved the gcc inline assembly implementing double word
atomic compare and exchange on x86/x86_64 so that it also
can be used when compiling with clang.</p>
<p>
Own Id: OTP-13336</p>
</item>
<item>
<p>
An optimization preventing a long wait for a scheduler
thread looking up information about a process executing
on another scheduler thread had unintentionally been lost
in erts-5.10 (OTP R16A). This optimization has now been
reintroduced.</p>
<p>
Own Id: OTP-13365 Aux Id: OTP-9892 </p>
</item>
</list>
</section>
</section>
<section><title>Erts 7.2.1</title>
<section><title>Fixed Bugs and Malfunctions</title>
<list>
<item>
<p>
Revert "Fix erroneous splitting of emulator path"</p>
<p>
Own Id: OTP-13202</p>
</item>
<item>
<p>
Fix HiPE enabled emulator for FreeBSD.</p>
<p>
Own Id: OTP-13204 Aux Id: pr926 </p>
</item>
</list>
</section>
</section>
<section><title>Erts 7.2</title>
<section><title>Fixed Bugs and Malfunctions</title>
<list>
<item>
<p>
Small documentation fixes</p>
<p>
Own Id: OTP-13017</p>
</item>
<item>
<p>
Fix memory corruption bug caused by disabling
distribution and then re-enable distribution with a node
name that has previously been used by a remote node.</p>
<p>
Own Id: OTP-13076 Aux Id: seq12959 </p>
</item>
<item>
<p>
Renamed variables with name bool as Visual Studio 2015
now treats this is a keyword.</p>
<p>
Own Id: OTP-13079</p>
</item>
<item>
<p><c>erl_prim_loader</c> has not supported custom
loaders for several releases. In the documentation for
<c>erl_prim_loader</c>, all references to custom loaders
have now been removed.</p>
<p>
Own Id: OTP-13102</p>
</item>
<item>
<p>
Fixed compilation of erts together with libc versions
that do not define __uint32_t.</p>
<p>
Own Id: OTP-13105</p>
</item>
<item>
<p>
erl -make now returns non-zero exit codes on failure</p>
<p>
Own Id: OTP-13107</p>
</item>
<item>
<p>
Fix crash on init:restart in embedded mode caused by
on_load handler process not being relaunched leading to
load failure for modules such as crypto and asn1rt_nif
that need it to be present for correct NIF loading.</p>
<p>
Own Id: OTP-13115</p>
</item>
<item>
<p>
Fix maps decode in erlang:binary_to_term/1</p>
<p>Decoding a term with a large (HAMT) map in an small
(FLAT) map could cause a critical error if the external
format was not produced by beam.</p>
<p>
Own Id: OTP-13125</p>
</item>
<item>
<p>
Fix very rare bug in GC when big maps with a lot of hash
collisions from a remote node are waiting in inner
message queue.</p>
<p>
Own Id: OTP-13146</p>
</item>
<item>
<p>
Fixed a bug that could cause a crash dump to become
almost empty.</p>
<p>
Own Id: OTP-13150</p>
</item>
</list>
</section>
<section><title>Improvements and New Features</title>
<list>
<item>
<p> Updated the xmllint target to just check the xml
files with real documentation content.<br/> Corrected
some errors and added some missing target in the DTD's.
</p>
<p>
Own Id: OTP-13026</p>
</item>
<item>
<p>
Add function enif_getenv to read OS environment variables
in a portable way from NIFs.</p>
<p>
Own Id: OTP-13147</p>
</item>
</list>
</section>
</section>
<section><title>Erts 7.1</title>
<section><title>Fixed Bugs and Malfunctions</title>
<list>
<item>
<p>
Fix bug in ETS that could cause stray objects marked for
deletion to occasionally be missed by the cleanup done by
<c>safe_fixtable(_,false)</c>.</p>
<p>
Own Id: OTP-12870</p>
</item>
<item>
<p>
Fixed VM crash that could occur if a trace port was
linked to a process, and the trace port terminated
abnormally while handling a trace message. This bug has
always existed in the runtime system with SMP support.</p>
<p>
Own Id: OTP-12901</p>
</item>
<item>
<p>
Instead of aborting, the vm now creates a crash dump when
a system process is terminated.</p>
<p>
Own Id: OTP-12934</p>
</item>
<item>
<p>
Fixed a rare emulator dead lock that occurred when
erlang:process_flag(priority,...) was called by a process
that was also scheduled for an internal system activity.</p>
<p>
Own Id: OTP-12943</p>
</item>
<item>
<p>
The runtime system on various posix platforms (except for
Linux and Solaris) could crash when large amounts of
file-descriptors were in use.</p>
<p>
Own Id: OTP-12954</p>
</item>
<item>
<p>
A beam file compiled by hipe for an incompatible runtime
system was sometimes not rejected by the loader, which
could lead to vm crash. This fix will also allow the same
hipe compiler to be used by both normal and debug-built
vm.</p>
<p>
Own Id: OTP-12962</p>
</item>
<item>
<p>
Fix bug in <c>maps:merge/2</c> when called by hipe
compiled code that could cause vm crash. Bug exists since
erts-7.0 (OTP 18.0).</p>
<p>
Own Id: OTP-12965</p>
</item>
<item>
<p>
When tracing with <c>process_dump</c> option, the VM
could abort if there was an ongoing binary match
somewhere in the call stack of the traced process.</p>
<p>
Own Id: OTP-12968</p>
</item>
<item>
<p>
Fixed possible output deadlock in tty driver when hitting
"CTRL-C" in a non-smp emulator shell on unix.</p>
<p>
Own Id: OTP-12987 Aux Id: Seq12947 </p>
</item>
<item>
<p>
Fix <c>binary_to_integer</c> to throw badarg for "+" and
"-" similar to <c>list_to_integer</c>.</p>
<p>
Own Id: OTP-12988</p>
</item>
<item>
<p>
Suppress warning of unused argument when using macro
enif_make_pid.</p>
<p>
Own Id: OTP-12989</p>
</item>
</list>
</section>
<section><title>Improvements and New Features</title>
<list>
<item>
<p>
Changed default clock source used for OS system time on
MacOS X to <c>gettimeofday()</c> in order to improve
performance. The system can be configured during build to
use the previously used higher resolution clock source by
passing the switch <seealso
marker="doc/installation_guide:INSTALL#Advanced-configuration-and-build-of-ErlangOTP_Configuring"><c>--with-clock-resolution=high</c></seealso>
when configuring the build.</p>
<p>
Own Id: OTP-12945 Aux Id: OTP-12892 </p>
</item>
<item>
<p>
Added the <c>configure</c> option <seealso
marker="doc/installation_guide:INSTALL#Advanced-configuration-and-build-of-ErlangOTP_Configuring"><c>--disable-saved-compile-time</c></seealso>
which disables saving of compile date and time in the
emulator binary.</p>
<p>
Own Id: OTP-12971</p>
</item>
</list>
</section>
</section>
<section><title>Erts 7.0.3</title>
<section><title>Fixed Bugs and Malfunctions</title>
<list>
<item>
<p>
Fixed a binary memory leak when printing to shell using
the tty driver (i.e. not -oldshell).</p>
<p>
Own Id: OTP-12941</p>
</item>
<item>
<p>
Fix a bug where the standard error port sometimes crashes
with eagain as the reason.</p>
<p>
Own Id: OTP-12942</p>
</item>
<item>
<p>
When tracing with <c>process_dump</c> option, the VM
could abort if there was an ongoing binary match
somewhere in the call stack of the traced process./</p>
<p>
Own Id: OTP-12968</p>
</item>
</list>
</section>
</section>
<section><title>Erts 7.0.2</title>
<section><title>Fixed Bugs and Malfunctions</title>
<list>
<item>
<p>
A process could end up in an inconsistent half exited
state in the runtime system without SMP support. This
could occur if the processes was traced by a port that it
also was linked to, and the port terminated abnormally
while handling a trace message for the process.</p>
<p>
This bug has always existed in the runtime system without
SMP support, but never in the runtime system with SMP
support.</p>
<p>
Own Id: OTP-12889 Aux Id: seq12885 </p>
</item>
<item>
<p>
Removed unnecessary copying of data when retrieving
corrected Erlang monotonic time.</p>
<p>
Own Id: OTP-12894</p>
</item>
<item>
<p>
Changed default OS monotonic clock source chosen at build
time. This in order to improve performance. The behavior
will now on most systems be that (both OS and Erlang)
monotonic time stops when the system is suspended.</p>
<p>
If you prefer that monotonic time elapse during suspend
of the machine, you can pass the command line argument
<c>--enable-prefer-elapsed-monotonic-time-during-suspend</c>
to <c>configure</c> when building Erlang/OTP. The
configuration stage will try to find such a clock source,
but might not be able to find it. Note that there might
be a performance penalty associated with such a clock
source.</p>
<p>
*** POTENTIAL INCOMPATIBILITY ***</p>
<p>
Own Id: OTP-12895</p>
</item>
<item>
<p>
<c>erlang:system_info(end_time)</c> returned a faulty
value on 32-bit architectures.</p>
<p>
Own Id: OTP-12896</p>
</item>
</list>
</section>
<section><title>Improvements and New Features</title>
<list>
<item>
<p>
The <c>configure</c> command line argument
<c>--enable-gettimeofday-as-os-system-time</c> has been
added which force usage of <c>gettimeofday()</c> for OS
system time. This will improve performance of
<c>os:system_time()</c> and <c>os:timestamp()</c> on
MacOS X, at the expense of worse accuracy, resolution and
precision of Erlang monotonic time, Erlang system time,
and OS system time.</p>
<p>
Own Id: OTP-12892</p>
</item>
</list>
</section>
</section>
<section><title>Erts 7.0.1</title>
<section><title>Fixed Bugs and Malfunctions</title>
<list>
<item>
<p>
Fix a rare hanging of the VM seen to happen just after
emulator start. Bug exists since R14.</p>
<p>
Own Id: OTP-12859 Aux Id: seq12882 </p>
</item>
</list>
</section>
</section>
<section><title>Erts 7.0</title>
<section><title>Fixed Bugs and Malfunctions</title>
<list>
<item>
<p>
Fix issuing with spaces and quoting in the arguments when
using erlang:open_port spawn_executable on windows. The
behavior now mimics how unix works. This change implies a
backwards incompatibility for how spawn_executable works
on windows.</p>
<p>
*** POTENTIAL INCOMPATIBILITY ***</p>
<p>
Own Id: OTP-11905</p>
</item>
<item>
<p>
Fix global call trace when hipe compiled code call beam
compiled functions. Tracing of beam functions should now
alway work regardless who the caller is.</p>
<p>
Own Id: OTP-11939</p>
</item>
<item>
<p>
Correct cache alignment for ETS <c>write_concurrency</c>
locks to improve performance by reduced false sharing.
May increase memory footprint for tables with
<c>write_concurrency</c>.</p>
<p>
Own Id: OTP-11974</p>
</item>
<item>
<p>
All possibly blocking operations in the fd/spawn and
terminal driver have been converted to non-blocking
operations. Before this fix it was possible for the VM to
be blocked for a long time if the entity consuming
stdout/stderr did not consume it fast enough.</p>
<p>
Own Id: OTP-12239</p>
</item>
<item>
<p>
Add missing overhead for offheap binaries created from
external format. This fix can improve the garbage
collection of large binaries originating from
<c>binary_to_term</c> or messages from remote nodes.</p>
<p>
Own Id: OTP-12554</p>
</item>
<item>
<p>
Ensure hashing of zero is consistent</p>
<p> Erlang treats positive and negative zero as
equal:</p>
<p>
<c>true = 0.0 =:= 0.0/-1</c></p>
<p>However, Erlangs hash functions: hash, phash and
phash2 did not reflect this behaviour. The hash values
produced by the different hash functions would not be
identical for positive and negative zero.</p> <p>This
change ensures that hash value of positive zero is always
produced regardless of the signedness of the zero float,
i.e.,</p>
<p>
<c>true = erlang:phash2(0.0) =:=
erlang:phash2(0.0/-1)</c></p>
<p>
Own Id: OTP-12641</p>
</item>
<item>
<p>
Ensure NIF term creation disallows illegal floating point
values and too long atoms. Such values will cause a NIF
to throw badarg exception when it returns.</p>
<p>
Own Id: OTP-12655</p>
</item>
<item>
<p>
Fixed building of Map results from match_specs</p>
<p>
A faulty "box-value" entered into the heap which could
cause a segmentation fault in the garbage collector if it
was written on a heap fragment.</p>
<p>
Own Id: OTP-12656</p>
</item>
<item>
<p>
Fix hipe bug when matching a "writable" binary. The bug
has been seen to sometimes cause a failed binary matching
of a correct utf8 character, but other symptoms are also
possible.</p>
<p>
Own Id: OTP-12667</p>
</item>
<item>
<p>
Keep dirty schedulers from waking other schedulers.</p>
<p>
Own Id: OTP-12685</p>
</item>
<item>
<p>
Disable floating point exceptions if the VM is compiled
by clang/llvm. This is a known long-standing problem in
clang/llvm.</p>
<p>
Own Id: OTP-12717</p>
</item>
<item>
<p>
Fix bug in <c>file:sendfile</c> for FreeBSD causing not
the entire file to be sent.</p>
<p>
Own Id: OTP-12720</p>
</item>
<item>
<p>
Fix the broken Android support in erl_child_setup.c</p>
<p>
Own Id: OTP-12751</p>
</item>
<item>
<p>
Faulty statistics reported by the <c>fix_alloc</c>
allocator.</p>
<p>
Own Id: OTP-12766</p>
</item>
<item>
<p>
Fix two erts_snprintf() calls to correct sizes.</p>
<p>
- run_erl.c (ose): Use the size of the signal type, not
its pointer. - erl_node_tables.c: Use the size of the
_BUFFER in erts_snprintf() to make sure we can use the
full space.</p>
<p>
Own Id: OTP-12771</p>
</item>
<item>
<p>
Delayed memory allocations could be delayed an
unnecessarily long time.</p>
<p>
Own Id: OTP-12812</p>
</item>
<item>
<p>
Make sure that timeouts on a pool of acceptors are
released in the correct order.</p>
<p>
Own Id: OTP-12817</p>
</item>
<item>
<p>
Fix segmentation fault in module_info for deleted modules</p>
<p>
Own Id: OTP-12820</p>
</item>
<item>
<p>Fix garbage collection of literals in code purge</p>
<p>During code purging and check_process_code, the
checking of the binary reference embedded in the match
binary state was omitted for the tracing tests. This
would cause the binary match state to reference
deallocated memory.</p>
<p>
Own Id: OTP-12821</p>
</item>
<item>
<p>
A bug has been corrected for gen_tcp:close so when
{linger,{true,0}} is in effect it does not wait for data
in the driver queue to transfer out before closing the
port. Bug fix by Rory Byrne.</p>
<p>
Own Id: OTP-12840</p>
</item>
<item>
<p>
The documentation of the driver callback <seealso
marker="driver_entry#start"><c>start()</c></seealso>
erroneously stated that a return value of
<c>ERL_DRV_ERROR_ERRNO</c> caused the error value to be
passed via <c>erl_errno</c> when it should have been
<c>errno</c>.</p>
<p>
Own Id: OTP-12855</p>
</item>
</list>
</section>
<section><title>Improvements and New Features</title>
<list>
<item>
<p>
Add <c>md5</c> and <c>module</c> entries to
<c>?MODULE:module_info/0/1</c> and remove obsolete entry
'import'.</p>
<p>
*** POTENTIAL INCOMPATIBILITY ***</p>
<p>
Own Id: OTP-11940</p>
</item>
<item>
<p>
Debug function <c>erlang:display/1</c> shows content of
binaries and bitstrings, not only the length.</p>
<p>
Own Id: OTP-11941</p>
</item>
<item>
<p>The time functionality of Erlang has been extended.
This both includes a <seealso
marker="time_correction#The_New_Time_API">new
API</seealso> for time, as well as <seealso
marker="time_correction#Time_Warp_Modes">time warp
modes</seealso> which alters the behavior of the system
when system time changes. <em>You are strongly encouraged
to use the new API</em> instead of the old API based on
<seealso
marker="erlang#now/0"><c>erlang:now/0</c></seealso>.
<c>erlang:now/0</c> has been deprecated since it is and
forever will be a scalability bottleneck. For more
information see the <seealso
marker="time_correction">Time and Time
Correction</seealso> chapter of the ERTS User's
Guide.</p>
<p>Besides the API changes and time warp modes a lot of
scalability and performance improvements regarding time
management has been made internally in the runtime
system. Examples of such improvements are scheduler
specific timer wheels, scheduler specific BIF timer
management, parallel retrieval of monotonic time and
system time on systems with primitives that are not
buggy.</p>
<p>
Own Id: OTP-11997</p>
</item>
<item>
<p><c>erlang:function_exported(M, F, A)</c> will now
return <c>true</c> if <c>M:F/A</c> refers to a BIF.</p>
<p>
*** POTENTIAL INCOMPATIBILITY ***</p>
<p>
Own Id: OTP-12099</p>
</item>
<item>
<p>
New BIF: <c>erlang:get_keys/0</c>, lists all keys
associated with the process dictionary. Note:
<c>erlang:get_keys/0</c> is auto-imported.</p>
<p>
*** POTENTIAL INCOMPATIBILITY ***</p>
<p>
Own Id: OTP-12151 Aux Id: seq12521 </p>
</item>
<item>
<p>
Make distributed send of large messages yield to improve
real-time characteristics.</p>
<p>
Own Id: OTP-12232</p>
</item>
<item>
<p>
Use high accuracy poll timeouts</p>
<p>
Where available, use poll/select API's that can handle
time resolutions less than 1ms. In the cases where such
API's are not available the timeout is rounded up to the
nearest ms.</p>
<p>
Own Id: OTP-12236</p>
</item>
<item>
<p>
The internal group to user_drv protocol has been changed
to be synchronous in order to guarantee that output sent
to a process implementing the user_drv protocol is
printed before replying. This protocol is used by the
standard_output device and the ssh application when
acting as a client. </p>
<p>
This change changes the previous unlimited buffer when
printing to standard_io and other devices that end up in
user_drv to 1KB.</p>
<p>
*** POTENTIAL INCOMPATIBILITY ***</p>
<p>
Own Id: OTP-12240</p>
</item>
<item>
<p>The previously introduced "eager check I/O" feature is
now enabled by default.</p>
<p>Eager check I/O can be disabled using the <c>erl</c>
command line argument: <seealso
marker="erl#+secio"><c>+secio false</c></seealso></p>
<p>Characteristics impact compared to previous
default:</p> <list> <item>Lower latency and smoother
management of externally triggered I/O operations.</item>
<item>A slightly reduced priority of externally triggered
I/O operations.</item> </list>
<p>
Own Id: OTP-12254 Aux Id: OTP-12117 </p>
</item>
<item>
<p>
Properly support maps in match_specs</p>
<p>
Own Id: OTP-12270</p>
</item>
<item>
<p>
The notice that a crashdump has been written has been
moved to be printed before the crashdump is generated
instead of afterwords. The wording of the notice has also
been changed.</p>
<p>
*** POTENTIAL INCOMPATIBILITY ***</p>
<p>
Own Id: OTP-12292</p>
</item>
<item>
<p>
New function <c>ets:take/2</c>. Works the same as
<c>ets:delete/2</c> but also returns the deleted
object(s).</p>
<p>
Own Id: OTP-12309</p>
</item>
<item>
<p>
Tracing with cpu_timestamp option has been enabled on
Linux.</p>
<p>
Own Id: OTP-12366</p>
</item>
<item>
<p>
ets:info/1,2 now contains information about whether
write_concurrency or read_concurrency is enabled.</p>
<p>
Own Id: OTP-12376</p>
</item>
<item>
<p>
Improved usage of <c>gcc</c>'s builtins for atomic memory
access. These are used when no other implementation of
atomic memory operations is available. For example, when
compiling for ARM when <c>libatomic_ops</c> is not
available.</p>
<p>
The largest improvement will be seen when compiling with
a <c>gcc</c> with support for the <c>__atomic_*</c>
builtins (using a <c>gcc</c> of at least version 4.7),
but also when only the legacy <c>__sync_*</c> builtins
are available (using a <c>gcc</c> of at least version
4.1) an improvement can be seen.</p>
<p>
For more information see the "<seealso
marker="doc/installation_guide:INSTALL#Advanced-configuration-and-build-of-ErlangOTP_Configuring_Atomic-Memory-Operations-and-the-VM">Atomic
Memory Operations and the VM</seealso>" section of
<c>$ERL_TOP/HOWTO/INSTALL.md</c>.</p>
<p>
Own Id: OTP-12383</p>
</item>
<item>
<p>
Introduce <c>math:log2/1</c> function to math module.</p>
<p>
Own Id: OTP-12411</p>
</item>
<item>
<p>The documentation of the Abstract Format (in the ERTS
User's Guide) has been updated with types and
specification. (Thanks to Anthony Ramine.) </p> <p> The
explicit representation of parentheses used in types of
the abstract format has been removed. Instead the new
functions <c>erl_parse:type_inop_prec()</c> and
<c>erl_parse:type_preop_prec()</c> can be used for
inserting parentheses where needed. </p>
<p>
Own Id: OTP-12492</p>
</item>
<item>
<p>
Remove perfctr support</p>
<p>
Development of perfctr in the linux kernel ceased in
2010. The perfctr support code in the Erlang VM is thus
effectively dead code and therefor removed.</p>
<p>
Own Id: OTP-12508</p>
</item>
<item>
<p><c>zlib:inflateChunk/2</c> has been added. It works
like <c>zlib:inflate/2</c>, but decompresses no more data
than will fit in the buffer configured by
<c>zlib:setBufSize/2</c>.</p>
<p>
Own Id: OTP-12548</p>
</item>
<item>
<p>
Use linear search for small select_val arrays</p>
<p>
Own Id: OTP-12555</p>
</item>
<item>
<p>
New BIF ets:update_counter/4 with a default object as
argument, which will be inserted in the table if the key
was not found.</p>
<p>
Own Id: OTP-12563</p>
</item>
<item>
<p>
Export missing types from zlib module</p>
<p>
Own Id: OTP-12584</p>
</item>
<item>
<p>
Use persistent hashmaps for large Maps</p>
<p>Maps will use a
persistent hashmap implementation when the number of
pairs in a Map becomes sufficiently large. The change
will occur when a Map reaches 33 pairs in size but this
limit might change in the future.</p>
<p>The most significant impact for the user by this
change is speed, and to a lesser degree memory
consumption and introspection of Maps. Memory consumption
size is probalistic but lesser than <c>gb_trees</c> or
<c>dict</c> for instance. Any other impacts will be
transparent for the user except for the following
changes.</p>
<p>Semantics of Maps have changed in two incompatible
ways compared to the experimental implementation in OTP
17:</p> <list> <item>Hashing of maps is done different by
<c>erlang:phash2/1,2</c>, <c>erlang:phash/1</c> and
<c>erlang:hash/2</c>.</item> <item>Comparing two maps
with ==, /=, =<, <, >= and >, is done
different if the keys contain floating point
numbers.</item> </list>
<p>
*** POTENTIAL INCOMPATIBILITY ***</p>
<p>
Own Id: OTP-12585</p>
</item>
<item>
<p>
Scalability improvement for <seealso
marker="erlang#make_ref/0">erlang:make_ref/0</seealso>,
and other functionality that create references. Each
scheduler now manage its own set of references. By this
no communication at all is needed when creating
references.</p>
<p>
Previous implementation generated a strictly
monotonically increasing sequence of references
corresponding to creation time on the runtime system
instance. This is <em>not</em> the case with current
implementation. You can only expect reference to be
unique. The Erlang/OTP documentation has never mentioned
anything else but the uniqueness property, so this change
<em>is</em> fully compatible. The only reason we've
marked this as a potential incompatibility is since an
early draft for an Erlang specification mentions strict
monotonicity as a property.</p>
<p>
If you need to create data with a strict monotonicity
property use <seealso
marker="erlang#unique_integer/1">erlang:unique_integer([monotonic])</seealso>.
Do <em>not</em> use the deprecated <seealso
marker="erlang#now/0">erlang:now()</seealso>.</p>
<p>
*** POTENTIAL INCOMPATIBILITY ***</p>
<p>
Own Id: OTP-12610</p>
</item>
<item>
<p>
Enable different abort signal from heart</p>
<p>By using environment variable HEART_KILL_SIGNAL, heart
can now use a different signal to kill the old running
Erlang.</p>
<p>By default the signal is SIGKILL but SIGABRT may also
be used by setting environment variable:
HEART_KILL_SIGNAL=SIGABRT</p>
<p>
Own Id: OTP-12613 Aux Id: seq12826 </p>
</item>
<item>
<p>
Update autconf to latest version 2015-03-04</p>
<p>
Own Id: OTP-12646</p>
</item>
<item>
<p>
Optimization of timers internally in the VM. This include
process timers (<c>receive ... after</c>), port timers
(<c>driver_set_timer()</c>) as well as BIF timers
(<c>erlang:send_after()</c>/<c>erlang:start_timer()</c>).</p>
<p>
Each scheduler thread now has its own lock-free timer
service instead of one locked central service. This
dramatically improves performance of timer management on
systems with a large amount of schedulers and timers.</p>
<p>
The timer service internal data structure has also been
optimized to be able to handle more timers than before.
That is, each timer service is by its self able to handle
more timers without dramatic performance loss than the
old centralized timer service.</p>
<p>
The API of BIF timers has also been extended. Timeout
values are for example no longer limited to 32-bit
integers. For more information see the documentation of
<seealso
marker="erlang#start_timer/4"><c>erlang:start_timer/4</c></seealso>,
<seealso
marker="erlang#send_after/4"><c>erlang:send_after/4</c></seealso>,
<seealso
marker="erlang#cancel_timer/2"><c>erlang:cancel_timer/2</c></seealso>,
and <seealso
marker="erlang#read_timer/2"><c>erlang:read_timer/2</c></seealso>.</p>
<p>
Characteristics impact: Calls to the synchronous versions
of <c>erlang:cancel_timer()</c>, and
<c>erlang:read_timer()</c> may take substantially longer
time to complete than before. This occur when the timer
that is accessed is managed by a remote scheduler. You
typically want to use the new asynchronous option in
order to avoid blocking the calling process.</p>
<p>
Own Id: OTP-12650 Aux Id: OTP-11997 </p>
</item>
<item>
<p>
Specialize instructions from common assembler patterns</p>
<p>Specialize common instructions of <c>rem</c>,
<c>band</c>, <c>minus</c> and <c>plus</c> in the beam
loader. This will reduce the number of fetches and thus
lessen the instruction dispatch pressure during runtime
and speed up those operations in some common cases.</p>
<p>Specialize move patterns from x-registers to the stack
with a new <c>move_window</c> instruction. This change
will reduce instruction dispatch pressure.</p>
<p>
Own Id: OTP-12690</p>
</item>
<item>
<p>
Fix cross compilation for Android.</p>
<p>
Own Id: OTP-12693</p>
</item>
<item>
<p>
Fix incorrect use of autoconf macro AC_EGREP_CPP, which
could cause faulty configuration if run from a path
containing the string 'yes'.</p>
<p>
Own Id: OTP-12706</p>
</item>
<item>
<p>
Minimal Java version is now 1.6</p>
<p>
Own Id: OTP-12715</p>
</item>
<item>
<p>
Send format and args on process exit to error_logger</p>
<p>
Previously, the emulator would generate a whole string
with values and call the error_logger passing
<c>"~s~n"</c>. This changes it to a format string
containing <c>~p</c> with the respective values as
arguments.</p>
<p>
Own Id: OTP-12735</p>
</item>
<item>
<p>
Map error logger warnings to warning messages by default.</p>
<p>
Own Id: OTP-12755</p>
</item>
<item>
<p>
Configure architecture ppc64le architecture as a ppc64</p>
<p>
Own Id: OTP-12761</p>
</item>
<item>
<p>
Add function <c>enif_raise_exception</c> to allow a NIF
to raise an error exception with any type of reason.</p>
<p>
Own Id: OTP-12770</p>
</item>
<item>
<p>
Optimized node table statistics retrieval.</p>
<p>
Own Id: OTP-12777</p>
</item>
<item>
<p>
Map beam error logger warnings to warning messages by
default. Previously these messages were mapped to the
error channel by default.</p>
<p>
Own Id: OTP-12781</p>
</item>
<item>
<p>
gen_tcp:shutdown/2 is now asynchronous</p>
<p>
This solves the following problems with the old
implementation:</p>
<p>
It doesn't block when the TCP peer is idle or slow. This
is the expected behaviour when shutdown() is called: the
caller needs to be able to continue reading from the
socket, not be prevented from doing so.</p>
<p>
It doesn't truncate the output. The current version of
gen_tcp:shutdown/2 will truncate any outbound data in the
driver queue after about 10 seconds if the TCP peer is
idle of slow. Worse yet, it doesn't even inform anyone
that the data has been truncated: 'ok' is returned to the
caller; and a FIN rather than an RST is sent to the TCP
peer.</p>
<p>
*** POTENTIAL INCOMPATIBILITY ***</p>
<p>
Own Id: OTP-12797</p>
</item>
<item>
<p>
Introduced delayed node table GC. This in order to avoid
oscillation of entries in and out of the tables. The
oscillation caused unnecessary lock contention on the
table locks. The delay length can be set by passing the
<seealso marker="erl#+zdntgc"><c>+zdntgc</c></seealso>
command line argument.</p>
<p>
Characteristics impact: The tables can grow to very large
sizes with unused entries if the node is get huge amounts
of short lived connections from other nodes. This problem
can be alleviated by shortening the length of the delay
using the <c>+zdntgc</c> command line argument.</p>
<p>
Own Id: OTP-12802</p>
</item>
<item>
<p>Improved implementation of <seealso
marker="erlang#statistics/1"><c>erlang:statistics</c></seealso><c>(io)</c>
in order to reduce contention between schedulers.</p>
<p>Characteristics impact: The actual call to
<c>erlang:statistics(io)</c> takes longer time to
complete, but the overall impact on the system is
improved.</p>
<p>
Own Id: OTP-12842</p>
</item>
<item>
<p>
There are many cases where user code needs to be able to
distinguish between a socket that was closed normally and
one that was aborted. Setting the option
{show_econnreset, true} enables the user to receive
ECONNRESET errors on both active and passive sockets.</p>
<p>
Own Id: OTP-12843</p>
</item>
<item>
<p>
Do not preallocate too large event pool</p>
<p>
A default pool size of 4000 is too excessive for the
common case. This corresponds directly to the number of
threads in the system. Change
ERTS_TS_EV_ALLOC_DEFAULT_POOL_SIZE to 2048. Change
ERTS_TS_EV_ALLOC_POOL_SIZE to 32.</p>
<p>
Own Id: OTP-12849</p>
</item>
</list>
</section>
</section>
<section><title>Erts 6.4.1.5</title>
<section><title>Fixed Bugs and Malfunctions</title>
<list>
<item>
<p>
Fixed a bug that could cause a crash dump to become
almost empty.</p>
<p>
Own Id: OTP-13150</p>
</item>
</list>
</section>
</section>
<section><title>Erts 6.4.1.4</title>
<section><title>Fixed Bugs and Malfunctions</title>
<list>
<item>
<p>
The 'raw' socket option could not be used multiple times
in one call to any e.g gen_tcp function because only one
of the occurrences were used. This bug has been fixed,
and also a small bug concerning propagating error codes
from within inet:setopts/2.</p>
<p>
Own Id: OTP-11482 Aux Id: seq12872 </p>
</item>
</list>
</section>
</section>
<section><title>Erts 6.4.1</title>
<section><title>Fixed Bugs and Malfunctions</title>
<list>
<item>
<p>
The VTS mode in Common Test has been modified to use a
private version of the Webtool application (ct_webtool).</p>
<p>
Own Id: OTP-12704 Aux Id: OTP-10922 </p>
</item>
</list>
</section>
</section>
<section><title>Erts 6.4</title>
<section><title>Fixed Bugs and Malfunctions</title>
<list>
<item>
<p>
Fix missing quotation in the <c>LM_FIND_EMU_CC</c>
<c>autoconf</c> macro which could cause build failures.</p>
<p>
Own Id: OTP-12388</p>
</item>
<item>
<p>
Fix erroneous printout of monitors in crashdump file.</p>
<p>
Own Id: OTP-12537</p>
</item>
<item>
<p>
The runtime system without SMP support could crash in the
BIF <c>port_control/3</c> if the port that was being
accessed died during the call to the BIF.</p>
<p>
Own Id: OTP-12544 Aux Id: Seq12777 </p>
</item>
<item>
<p>
Avoid corrupt oversized integer to be created from binary
matching. Instead throw system_limit exception which is
the correct behavior. A peculiar symptom of this bug was
that bitwise operations (band, bor, bxor) on such
oversized integers could return the empty list [].
Credit: Mikael Pettersson, Nico Kruber</p>
<p>
Own Id: OTP-12556</p>
</item>
<item>
<p>
A race condition when calling <c>port_info/1</c> could
cause a memory fault has been fixed.</p>
<p>
Own Id: OTP-12587</p>
</item>
<item>
<p>
Fix comparison of exact terms. An overflow that could
cause faulty comparisons has been fixed. Comparison of
exact terms is exclusively used within Maps.</p>
<p>
Own Id: OTP-12623</p>
</item>
<item>
<p>
Fix bug in <c>list_to_integer/1</c> for very long lists
that could cause VM crash.</p>
<p>
Own Id: OTP-12624</p>
</item>
</list>
</section>
<section><title>Improvements and New Features</title>
<list>
<item>
<p>
Introduced a runtime system internal 64-bit API for
atomic memory operations.</p>
<p>
Own Id: OTP-12351</p>
</item>
<item>
<p>
Add command line argument option for the initial size of
process dictionaries.</p>
<p>
Use '+hpds <size>' to set initial process
dictionary size for spawned processes.</p>
<p>
Own Id: OTP-12535 Aux Id: seq12809 </p>
</item>
<item>
<p>
Fix documentation on $char for Unicode</p>
<p>
Own Id: OTP-12545</p>
</item>
</list>
</section>
</section>
<section><title>Erts 6.3.1</title>
<section><title>Fixed Bugs and Malfunctions</title>
<list>
<item>
<p>
Fix getifaddrs realloc pointer error</p>
<p>
When a buffer was exhausted and subsequently reallocated,
we could get an unsafe pointer pointing to faulty memory.</p>
<p>
For this to occur we would need to have a large number of
interfaces and a reallocation of memory to a lower
addresses.</p>
<p>
The symptom would be garbage returned from
erlang:port_control(Port, 25, [])
(prim_inet:getifaddrs(Port) resulting in a badarg) or a
segmentation fault.</p>
<p>
Own Id: OTP-12445</p>
</item>
<item>
<p>
Don't close all file descriptors twice in child_setup</p>
<p>
The commit c2b4eab25c907f453a394d382c04cd04e6c06b49
introduced an error in which child_setup erroneously
tried to close all file descriptors twice.</p>
<p>
Use closefrom() if available when closing all file
descriptors.</p>
<p>
The function closefrom() was only used in the vfork()
case before but is now also used in the fork() case if
available.</p>
<p>
Own Id: OTP-12446</p>
</item>
<item>
<p>
During a crashdump all file descriptors are closed to
ensure the closing of the epmd port and to reserve a file
descriptor for the crashdump file.</p>
<p>
If a driver (third party library) cannot handle closing
of sockets this could result in a segmentation fault in
which case a crashdump would not be produced. This is now
fixed by only closing inets sockets via an emergency
close callback to the driver and thus closing the epmd
socket.</p>
<p>
Own Id: OTP-12447</p>
</item>
</list>
</section>
</section>
<section><title>Erts 6.3</title>
<section><title>Fixed Bugs and Malfunctions</title>
<list>
<item>
<p>
Fix HiPE debug lock checking on OS X 64bit</p>
<p>
Position-independent code is mandatory on OS X. We use
r11 as an intermediate register to fill
BIF_P->hipe.bif_callee. This fixes the following error
when doing `make debug FLAVOR=smp`:</p>
<p>
clang -cc1as: fatal error: error in backend: 32-bit
absolute addressing is not supported in 64-bit mode</p>
<p>
Own Id: OTP-12188</p>
</item>
<item>
<p>
Fix race bug that could cause VM crash in
<c>erlang:port_get_data/1</c> if the port was closed by a
concurrent process. Also fix fatal bug if
<c>port_set_data/2</c> is called with a non-immediate
data term. Both bugs exist since R16B01.</p>
<p>
Own Id: OTP-12208</p>
</item>
<item>
<p>
Correct make variable SSL_DED_LD_RUNTIME_LIBRARY_PATH
when erl_xcomp_sysroot ends with a slash.</p>
<p>
Own Id: OTP-12216 Aux Id: seq12700 </p>
</item>
<item>
<p>
Fix two cases of unreachable code caused by false use of
assigment operators.</p>
<p>
Own Id: OTP-12222</p>
</item>
<item>
<p>
Fix bug when hipe compiled code makes tail call to a BIF
that disables GC while trapping (sush as binary_to_list,
list_to_binary, binary_to_term, term_to_binary).</p>
<p>
Own Id: OTP-12231</p>
</item>
<item>
<p>
Fix bug when a migrated empty memory carrier is reused
just before it should be destroyed by the thread that
created it.</p>
<p>
Own Id: OTP-12249</p>
</item>
<item>
<p>
Prevents compile-time errors in NIFs, when the compiler
is instructed to treat missing field initializers as
errors, by adding an initializer for the new options
field which was added to ErlNifEntry for 17.3.</p>
<p>
Own Id: OTP-12266</p>
</item>
<item>
<p>
Fixed CPU topology detection on FreeBSD systems where
Erlang/OTP is compiled by new C compilers (including, but
possibly not limited to, gcc 4.9 and clang).</p>
<p>
Own Id: OTP-12267</p>
</item>
<item>
<p>
Use C99 function isfinite() instead of finite() when
available on non GCC compilers.</p>
<p>
Own Id: OTP-12268</p>
</item>
<item>
<p>
Fix bug on windows where an incorrect number of links
could be returned when doing file:read_file_info on a
directory.</p>
<p>
Own Id: OTP-12269</p>
</item>
<item>
<p>
Fix rare bug when purging module on VM started with
+Meamin.</p>
<p>
Own Id: OTP-12273</p>
</item>
<item>
<p>
Repair run_erl terminal window size adjustment sent from
to_erl. This was broken in OTP 17.0 which could lead to
strange cursor behaviour in the to_erl shell.</p>
<p>
Own Id: OTP-12275 Aux Id: seq12739 </p>
</item>
<item>
<p>
Fixed bug on windows causing gen_tcp/udp to return an
error when given an fd to work with.</p>
<p>
Own Id: OTP-12289</p>
</item>
<item>
<p>
Fix various internal erts issues where negating a signed
integer in C would trigger undefined behavior. This fixes
issues when dividing with bignums and list_to_integer.</p>
<p>
Own Id: OTP-12290</p>
</item>
<item>
<p>
When flushing output to stdout on windows, the emulator
could sometimes hang indefinitely waiting for the flush
to complete. This has been fixed.</p>
<p>
Own Id: OTP-12291</p>
</item>
<item>
<p>
Fix so that non-smp emulators with dirty scheduler
support shows the correct number of dirty schedulers when
calling erlang:system_info(system_version).</p>
<p>
Own Id: OTP-12295</p>
</item>
<item>
<p>
Add <c>nif_version</c> to <c>erlang:system_info/1</c> in
order to get the NIF API version of the runtime system in
a way similar to <c>driver_version</c>.</p>
<p>
Own Id: OTP-12298</p>
</item>
<item>
<p>
Fix bug that could cause the return value from dirty NIF
with zero arity to be treated as garbage, leading to VM
crash.</p>
<p>
Own Id: OTP-12300</p>
</item>
<item>
<p>
Improve allocation carrier migration search logic. This
will reduce the risk of failed migrations that could lead
to excess memory consumption. It will also improve smp
performance due to reduced memory contention on the
migration pool.</p>
<p>
Own Id: OTP-12323</p>
</item>
</list>
</section>
<section><title>Improvements and New Features</title>
<list>
<item>
<p>Introduced support for eager check I/O.</p>
<p>By default eager check I/O will be disabled, but this
will most likely be changed in OTP 18. When eager check
I/O is enabled, schedulers will more frequently check for
I/O work. Outstanding I/O operations will however not be
prioritized to the same extent as when eager check I/O is
disabled.</p>
<p>Eager check I/O can be enabled using the <c>erl</c>
command line argument: <seealso
marker="erl#+secio"><c>+secio true</c></seealso></p>
<p>Characteristics impact when enabled:</p> <list>
<item>Lower latency and smoother management of externally
triggered I/O operations.</item> <item>A slightly reduced
priority of externally triggered I/O operations.</item>
</list>
<p>
Own Id: OTP-12117</p>
</item>
<item>
<p>
Fix erts .app-file</p>
<p>
Own Id: OTP-12189</p>
</item>
<item>
<p>
Add configure option --with-ssl-incl=PATH to support
OpenSSL installations with headers and libraries at
different places.</p>
<p>
Own Id: OTP-12215 Aux Id: seq12700 </p>
</item>
<item>
<p>
Optimization of atomic memory operations with release
barrier semantics on 32-bit PowerPC when using the
implementation included in OTP.</p>
<p>
Own Id: OTP-12250</p>
</item>
<item>
<p>
Minor adjustment of scheduler activation code making sure
that an activation of a scheduler is not prevented by its
run-queue being non-empty. (Thanks to Songlu Cai)</p>
<p>
Own Id: OTP-12287</p>
</item>
<item>
<p>
Improved support for atomic memory operations provided by
the <url
href="https://github.com/ivmai/libatomic_ops/">libatomic_ops</url>
library. Most importantly support for use of native
double word atomics when implemented by
<c>libatomic_ops</c> (for example, implemented for ARM).</p>
<p>
The <seealso
marker="doc/installation_guide:INSTALL#Advanced-configuration-and-build-of-ErlangOTP_Configuring_Atomic-Memory-Operations-and-the-VM"><c>$ERL_TOP/HOWTO/INSTALL.md</c></seealso>
document now also more clearly describes when you want to
build together with a <c>libatomic_ops</c> installation.</p>
<p>
Own Id: OTP-12302</p>
</item>
<item>
<p>
Add configure option --with-ssl-rpath to control which
runtime library path to use for dynamic linkage toward
OpenSSL.</p>
<p>
Own Id: OTP-12316 Aux Id: seq12753 </p>
</item>
<item>
<p>
Added systemd notify support to epmd</p>
<p>
Own Id: OTP-12321</p>
</item>
</list>
</section>
</section>
<section><title>Erts 6.2.1</title>
<section><title>Fixed Bugs and Malfunctions</title>
<list>
<item>
<p>
Fix bug when an migrated empty memory carrier is reused
just before it should be destroyed by the thread that
created it.</p>
<p>
Own Id: OTP-12249</p>
</item>
<item>
<p>
Repair run_erl terminal window size adjustment sent from
to_erl. This was broken in OTP 17.0 which could lead to
strange cursor behaviour in the to_erl shell.</p>
<p>
Own Id: OTP-12275 Aux Id: seq12739 </p>
</item>
</list>
</section>
</section>
<section><title>Erts 6.2</title>
<section><title>Fixed Bugs and Malfunctions</title>
<list>
<item>
<p>
General documentation updates.</p>
<p>
Own Id: OTP-12052</p>
</item>
<item>
<p>A bug in the VM code implementing sending of signals
to ports could cause the receiving port queue to remain
in a busy state forever. When this state had been
reached, processes sending command signals to the port
either got suspended forever, or, if the <c>nosuspend</c>
feature was used, always failed to send to the port. This
bug was introduced in ERTS version 5.10.</p>
<p>In order for this bug to be triggered on a port, one
had to at least once utilize the <c>nosuspend</c>
functionality when passing a signal to the port. This by
either calling</p> <list> <item> <seealso
marker="erlang#port_command/3"><c>port_command(Port,
Data, [nosuspend | Options])</c></seealso>, </item>
<item> <seealso
marker="erlang#send/3"><c>erlang:send(Port, {PortOwner,
{command, Data}}, [nosuspend | Options])</c></seealso>,
</item> <item> <seealso
marker="erlang#send_nosuspend/2"><c>erlang:send_nosuspend(Port,
{PortOwner, {command, Data}})</c></seealso>, or </item>
<item> <seealso
marker="erlang#send_nosuspend/3"><c>erlang:send_nosuspend(Port,
{PortOwner, {command, Data}}, Options)</c></seealso>.
</item> </list>
<p>Thanks Vasily Demidenok for reporting the issue, and
Sergey Kudryashov for providing a testcase.</p>
<p>
Own Id: OTP-12082 Aux Id: OTP-10336 </p>
</item>
<item>
<p>
Fix size overflow bug at memory allocation. A memory
allocation call, with an insane size close to the entire
address space, could return successfully as if it had
allocated just a few bytes. (Thanks to Don A. Bailey for
reporting)</p>
<p>
Own Id: OTP-12091</p>
</item>
<item>
<p>
Fix various issues where negating a signed integer would
trigger undefined behaviour. This fixes issues in the
enif_make_int64 interface and some edge cases inside the
erlang runtime system.</p>
<p>
Own Id: OTP-12097</p>
</item>
<item>
<p>
The documentation erroneously listed the <seealso
marker="erl#+swct"><c>+swct</c></seealso> command line
argument under <c>+sws</c>.</p>
<p>
Own Id: OTP-12102 Aux Id: OTP-10994 </p>
</item>
<item>
<p>
Profiling messages could be delivered out of order when
profiling on <c>runnable_procs</c> and/or
<c>runnable_ports</c> using <seealso
marker="erlang#system_profile/2"><c>erlang:system_profile/2</c></seealso>.
This bug was introduced in ERTS version 5.10.</p>
<p>
Own Id: OTP-12105 Aux Id: OTP-10336 </p>
</item>
<item>
<p>
Various logging fixes, including: Add run queue index to
the process dump in crash dumps.<br/> Add thread index to
enomem slogan when crashing.<br/> Remove error logger
message for sending messages to old instances of the same
node.</p>
<p>
Own Id: OTP-12115</p>
</item>
<item>
<p>
Fix compiler warnings reported by LLVM</p>
<p>
Own Id: OTP-12138</p>
</item>
<item>
<p>
Correct conversion of <c>MIN_SMALL</c> by
<c>list_to_integer/1</c> and <c>binary_to_integer/1</c>.
The bug produced an unnormalized bignum which can cause
strange behavior such as comparing different to a correct
<c>MIN_SMALL</c> integer. The value <c>MIN_SMALL</c> is
<c>-(1 bsl 27) = -134217728</c> on a 32-bit VM and <c>-(1
bsl 59) = -576460752303423488</c> on a 64-bit VM. (Thanks
to Jesper Louis Andersen, Mikael Pettersson and Anthony
Ramine for report, patch and optimization suggestion)</p>
<p>
Own Id: OTP-12140</p>
</item>
<item>
<p>
Fix bug in <c>term_to_binary</c> that reallocates binary
with inconsistent size information. Bug has never been
confirmed to be the cause of any faulty behavior.</p>
<p>
Own Id: OTP-12141</p>
</item>
<item>
<p>
Real_path method used while prim loading archive files
was not taking into account the fact that windows
directory symlinks can be across different drives.</p>
<p>
Own Id: OTP-12155</p>
</item>
</list>
</section>
<section><title>Improvements and New Features</title>
<list>
<item>
<p>
Add log2 histogram to lcnt for lock wait time</p>
<p>
Own Id: OTP-12059</p>
</item>
<item>
<p>
Introduced <seealso
marker="erl_nif#enif_schedule_nif"><c>enif_schedule_nif()</c></seealso>
to the NIF API.</p>
<p>
The <c>enif_schedule_nif()</c> function allows a
long-running NIF to be broken into separate NIF
invocations without the help of a wrapper function
written in Erlang. The NIF first executes part of the
long-running task, then calls <c>enif_schedule_nif()</c>
to schedule a NIF for later execution to continue the
task. Any number of NIFs can be scheduled in this manner,
one after another. Since the emulator regains control
between invocations, this helps avoid problems caused by
native code tying up scheduler threads for too long.</p>
<p>
The <c>enif_schedule_nif()</c> function also replaces the
<c>enif_schedule_dirty_nif()</c> in the experimental
dirty NIF API. Note that the only incompatible changes
made are in the experimental dirty NIF API.</p>
<p>
See the <seealso marker="erl_nif">NIF
documentation</seealso> for more information.</p>
<p>
Thanks to Steve Vinoski.</p>
<p>
*** POTENTIAL INCOMPATIBILITY ***</p>
<p>
Own Id: OTP-12128</p>
</item>
</list>
</section>
</section>
<section><title>Erts 6.1.2</title>
<section><title>Fixed Bugs and Malfunctions</title>
<list>
<item>
<p>
OTP-11850 fixed filelib:wildcard/1 to work with broken
symlinks. This correction, however, introduced problems
since symlinks were no longer followed for functions like
filelib:ensure_dir/1, filelib:is_dir/1,
filelib:file_size/1, etc. This is now corrected.</p>
<p>
Own Id: OTP-12054 Aux Id: seq12660 </p>
</item>
</list>
</section>
</section>
<section><title>Erts 6.1.1</title>
<section><title>Fixed Bugs and Malfunctions</title>
<list>
<item>
<p>
Fixed ETHR_FORCE_INLINE which caused the build to break
on some platforms without adequate thread support
(VxWorks).</p>
<p>
Own Id: OTP-12010</p>
</item>
</list>
</section>
</section>
<section><title>Erts 6.1</title>
<section><title>Fixed Bugs and Malfunctions</title>
<list>
<item>
<p>The documentation for <c>spawn_opt/5</c> now has a
note mentioning that the <c>monitor</c> option is not
supported.</p>
<p>
Own Id: OTP-11849</p>
</item>
<item>
<p>
Fix broken system monitoring of <c>large_heap</c> for
non-smp VM. No message for <c>large_heap</c> was ever
sent on non-smp VM. Bug exist since R16B.</p>
<p>
Own Id: OTP-11852</p>
</item>
<item>
<p>
The emulator without SMP support crashed when passing a
message to a process without enough heap space for the
message. This bug was introduced in <c>erts-6.0</c>.</p>
<p>
Own Id: OTP-11887 Aux Id: OTP-11388 </p>
</item>
<item>
<p>
Fix race between ETS table deletion and unfixation that
could cause VM crash. The race could happen between a
terminating process that does not own the table but has a
fixation on it and another process that deletes the table
(maybe the owner terminating) at the same time. Bug
existed since R15B02.</p>
<p>
Own Id: OTP-11892</p>
</item>
<item>
<p>The string following the <c>-eval</c> option when
invoking <c>erl</c> would not be properly translated from
UTF-8 to a list of Unicode characters (as would the
arguments for <c>-run</c>).</p>
<p>That bug would cause the build of Erlang/OTP to fail
when building in a directory whose pathname contained
non-US ASCII characters encoded in UTF-8. (Thanks to Eric
Pailleau for reporting this bug.)</p>
<p>
Own Id: OTP-11916</p>
</item>
<item>
<p>
Fix erts_debug:size/1 to handle Map sizes</p>
<p>
Own Id: OTP-11923</p>
</item>
<item>
<p>
Removed <c>erlang:bitstr_to_list/1</c> and
<c>erlang:list_to_bitstr/1</c>. They were added by
mistake, and have always raised an <c>undefined</c>
exception when called.</p>
<p>
Own Id: OTP-11942</p>
</item>
<item>
<p>
Fixed compilation using mingw-w64 on Windows.</p>
<p>
Thanks to Jani Hakala.</p>
<p>
Own Id: OTP-11945</p>
</item>
<item>
<p>
The git sha is no longer printed in the shell start
header when erlang is built from a tagged git release.</p>
<p>
Own Id: OTP-11961</p>
</item>
<item>
<p>
Fixed a bug where <c>send</c> trace events were
erroneously dropped when the send was done to a
registered process. This bug was introduced in R16B.</p>
<p>
Own Id: OTP-11968</p>
</item>
</list>
</section>
<section><title>Improvements and New Features</title>
<list>
<item>
<p>The following native functions now bump an appropriate
amount of reductions and yield when out of
reductions:</p> <list>
<item><c>erlang:binary_to_list/1</c></item>
<item><c>erlang:binary_to_list/3</c></item>
<item><c>erlang:bitstring_to_list/1</c></item>
<item><c>erlang:list_to_binary/1</c></item>
<item><c>erlang:iolist_to_binary/1</c></item>
<item><c>erlang:list_to_bitstring/1</c></item>
<item><c>binary:list_to_bin/1</c></item> </list>
<p>Characteristics impact:</p> <taglist>
<tag>Performance</tag> <item>The functions converting
from lists got a performance loss for very small lists,
and a performance gain for very large lists.</item>
<tag>Priority</tag> <item>Previously a process executing
one of these functions effectively got an unfair priority
boost. This priority boost depended on the input size.
The larger the input was, the larger the priority boost
got. This unfair priority boost is now lost. </item>
</taglist>
<p>
Own Id: OTP-11888</p>
</item>
<item>
<p>
The systemd features of epmd have been removed from epmd
by default. To enable them you have to build erlang with
the configure option --enable-systemd.</p>
<p>
Own Id: OTP-11921</p>
</item>
<item>
<p>
Removed Erlang wrapper code used when calling
<c>binary_to_term/1</c>, and <c>binary_to_term/2</c>.
This improves the performance of these BIFs especially
when they are called with small binaries as input.</p>
<p>
Own Id: OTP-11931</p>
</item>
<item>
<p>
Add erlang:system_info(tolerant_timeofday), an API to
check whether compensation for sudden changes of system
time is enabled or not.</p>
<p>
Own Id: OTP-11970</p>
</item>
</list>
</section>
</section>
<section><title>Erts 6.0.1</title>
<section><title>Fixed Bugs and Malfunctions</title>
<list>
<item>
<p>
Fix broken system monitoring of <c>large_heap</c> for
non-smp VM. No message for <c>large_heap</c> was ever
sent on non-smp VM. Bug exist since R16B.</p>
<p>
Own Id: OTP-11852</p>
</item>
<item>
<p>
Fixed type spec of <c>erlang:system_info/1</c>.</p>
<p>
Own Id: OTP-11859 Aux Id: OTP-11615 </p>
</item>
</list>
</section>
</section>
<section><title>Erts 6.0</title>
<section><title>Fixed Bugs and Malfunctions</title>
<list>
<item>
<p>
The option dupnames did not work as intended in re. When
looking for names with {capture, [Name, ...]}, re:run
returned a random instance of the match for that name,
instead of the leftmost matching instance, which was what
the documentation stated. This is now corrected to adhere
to the documentation. The option {capture,all_names}
along with a re:inspect/2 function is also added to
further help in using named subpatterns.</p>
<p>
Own Id: OTP-11205</p>
</item>
<item>
<p>
Allow loading of NIF library with unicode path name</p>
<p>
Own Id: OTP-11408</p>
</item>
<item>
<p>
Allow loading of driver with unicode path name</p>
<p>
Own Id: OTP-11549</p>
</item>
<item>
<p>
Fixed a bug where starting Erlang without having an open
stdin on fd 0 would sometimes deadlock the emulator when
terminating.</p>
<p>
Own Id: OTP-11558</p>
</item>
<item>
<p>
The option '-names' in epmd now works on Windows (Thanks
to Johannes Weißl)</p>
<p>
Own Id: OTP-11565</p>
</item>
<item>
<p>
Correction of the examples in escript documentation.
(Thanks to Pierre Fenoll).</p>
<p>
Own Id: OTP-11577</p>
</item>
<item>
<p>
Fix bs_get_integer instruction</p>
<p>
The instruction bs_get_integer could unnecessarily
trigger a garbage collection in failure cases which is
unwanted or outright dangerous.</p>
<p>
Ex:</p>
<p>
<<X:Sz,_/bits>> = <<"some
binary">></p>
<p>
Previously, if Sz induced X to a bignum it would reserved
memory size this on the heap via a garbage collection
before checking if the size could actually match.</p>
<p>
It will now check the binary size before triggering a
collection.</p>
<p>
Own Id: OTP-11581</p>
</item>
<item>
<p>
Remove heap space overestimation in <c>binary_to_term</c>
(and remote message reception) for integers in the
intervals [-2147483648,-1] and [256,2147483647] on 64-bit
emulators.</p>
<p>
Own Id: OTP-11585</p>
</item>
<item>
<p>
Add support for detecting the separate tinfo library from
ncurses (Thanks to Dirkjan Ochtman)</p>
<p>
Own Id: OTP-11590</p>
</item>
<item>
<p>
Deprecation warning for system_flag(cpu_topology) has
been extended for removal in OTP 18 (Thanks to Steve
Vinoski for the update)</p>
<p>
Own Id: OTP-11602</p>
</item>
<item>
<p>
Documentation improvement regarding some awkward wording
around the +spp flag. (Thanks to Brian L. Troutwine )</p>
<p>
Own Id: OTP-11607</p>
</item>
<item>
<p>
Fixed bug where sendfile would return the wrong error
code for a remotely closed socket if the socket was in
passive mode. (Thanks to Vincent Siliakus for reporting
the bug.)</p>
<p>
Own Id: OTP-11614</p>
</item>
<item>
<p>
Increase garbage collection tenure rate</p>
<p>The garbage collector tries to maintain the previous
heap block size during a minor gc, i.e. 'need' is not
utilized in determining the size of the new heap, instead
it relies on tenure and garbage to be sufficiently
large.</p>
<p>In instances during intense growing with exclusively
live data on the heap coupled with delayed tenure,
fullsweeps would be triggered directly after a minor gc
to make room for 'need' since the new heap would be
full.</p>
<p>To remedy this, the tenure of terms on the minor heap
will always happen (if it is below the high watermark)
instead of every other minor gc.</p>
<p>Characteristics Impact: Reduced CPU-time spent in
garbage collection but may infer delays in collecting
garbage from the heap. Tweak 'fullsweep_after' options to
increase gc pressure if needed.</p>
<p>
Own Id: OTP-11617</p>
</item>
<item>
<p>
Fix bug when comparing integers with floats larger than
2^992. The bug could potentially cause memory corruption
on 32-bit emulators.</p>
<p>
Own Id: OTP-11618</p>
</item>
<item>
<p>
Cross-compilation fixes for TileraMDE-3.0.1.125620</p>
<p>
Own Id: OTP-11635</p>
</item>
<item>
<p>
sendfile no longer uses async threads by default</p>
<p>
This has been done because a slow client attack is
possible if the async thread pool is used. The scenario
is:</p>
<p>
Client does a request for a file and then slowly receives
the file one byte at a time. This will eventually fill
the async thread pool with blocking sendfile operations
and thus starving the vm of all file operations.</p>
<p>
If you still want to use the async threads pool for
sendfile an option to enable it has been introduced.</p>
<p>
Thanks to Christopher Faulet for identifying this
vulnerability.</p>
<p>
*** POTENTIAL INCOMPATIBILITY ***</p>
<p>
Own Id: OTP-11639</p>
</item>
<item>
<p>
Do proper rollback of calls to
<c>enif_open_resource_type</c> when load/upgrade
callbacks of NIF library return failure.</p>
<p>
Own Id: OTP-11722</p>
</item>
<item>
<p>
Changed the default configuration when configuring with
<c>$ERL_TOP/configure</c> to be the same as when
configuring with <c>$ERL_TOP/otp_build configure</c>.</p>
<p>
Previously floating point exceptions got enabled by
default on Linux when HiPE was enabled when configuring
with <c>$ERL_TOP/configure</c>, but not when configuring
with <c>$ERL_TOP/otp_build configure</c>. The default is
now in both cases not to use floating point exceptions
since there still exist unresolved issues with floating
point exceptions on Linux.</p>
<p>
For more information see <seealso
marker="doc/installation_guide:INSTALL"><c>$ERL_TOP/HOWTO/INSTALL.md</c></seealso>.</p>
<p>
*** POTENTIAL INCOMPATIBILITY ***</p>
<p>
Own Id: OTP-11723</p>
</item>
<item>
<p>
A comment in erl_db_tree.c no longer differ from the
code. (Thanks to Cobus Carstens)</p>
<p>
Own Id: OTP-11793</p>
</item>
<item>
<p>
Fix epmd debug functionality for VxWorks (Thanks to Jay
True)</p>
<p>
Own Id: OTP-11808</p>
</item>
<item>
<p>
Use closefrom/2 when available in child_setup (Thanks to
Rick Reed and Anthony Ramine)</p>
<p>
Own Id: OTP-11809</p>
</item>
<item>
<p>
Fix dtrace/systemtap bug where the probe arguments would
be concatenated due to faulty length calculation. </p>
<p>
Thanks to Michal Ptaszek and Scott Lystig Fritchie</p>
<p>
Own Id: OTP-11816</p>
</item>
<item>
<p>
It is now better documented that the <c>+fn*</c> flags to
<c>erl</c> also affect how command line parameters and
environment variables are read. (Thanks to Vlad
Dumitrescu)</p>
<p>
Own Id: OTP-11818</p>
</item>
</list>
</section>
<section><title>Improvements and New Features</title>
<list>
<item>
<p>
Options to set match_limit and match_limit_recursion are
added to re:run. The option report_errors is also added
to get more information when re:run fails due to limits
or compilation errors.</p>
<p>
Own Id: OTP-10285</p>
</item>
<item>
<p> Dialyzer's <c>unmatched_return</c> warnings have been
corrected. </p>
<p>
Own Id: OTP-10908</p>
</item>
<item>
<p>
A common case is to wrap an argument to
<c>list_to_binary/1</c> in a list to ensure conversion
can happen even though the argument may already be a
binary. Take special care of this case and do not copy
binary.</p>
<p>
Impact: May cause incompatibility since a single binary
is no longer copied. Use <c>binary:copy/1,2</c> instead.</p>
<p>
*** POTENTIAL INCOMPATIBILITY ***</p>
<p>
Own Id: OTP-11082</p>
</item>
<item>
<p>
Make erlang:open_port/2 spawn and spawn_executable handle
unicode.</p>
<p>
Own Id: OTP-11105</p>
</item>
<item>
<p>
Handle unicode (widestring) in erl, erlc, heart, etc on
windows.</p>
<p>
Own Id: OTP-11135</p>
</item>
<item>
<p>
The version of the PCRE library Used by Erlang's re
module is raised to 8.33 from 7.6. This means, among
other things, better Unicode and Unicode Character
Properties support. New options connected to PCRE 8.33
are also added to the re module (ucd, notempty_atstart,
no_start_optimize). PCRE has extended the regular
expression syntax between 7.6 and 8.33, why this imposes
a potential incompatibility. Only very complicated
regular expressions may be affected, but if you know you
are using obscure features, please test run your regular
expressions and verify that their behavior has not
changed.</p>
<p>
*** POTENTIAL INCOMPATIBILITY ***</p>
<p>
Own Id: OTP-11204</p>
</item>
<item>
<p>Filenames containing UTF-8 encoded characters can now
be handled by erlc.</p>
<p>If you have set the <c>ERLC_EMULATOR</c> environment
variable, note that <c>erlc</c> in OTP 17 will only work
with <c>erl</c> in OTP 17 since the protocol between the
<c>erlc</c> program and the <c>erl_compile</c> module has
changed.</p>
<p>
Own Id: OTP-11248</p>
</item>
<item>
<p>
By giving --enable-static-{nifs,drivers} to configure it
is now possible to statically linking of nifs and drivers
to the main Erlang VM binary. At the moment only the asn1
and crypto nifs of the Erlang/OTP nifs and drivers have
been prepared to be statically linked. For more details
see the Installation Guide in the System documentation.</p>
<p>
Own Id: OTP-11258</p>
</item>
<item>
<p>
Erlang/OTP has been ported to the realtime operating
system OSE. The port supports both smp and non-smp
emulator. For details around the port and how to started
see the User's Guide in the <seealso
marker="ose:ose_intro">ose</seealso> application. </p>
<p>
Note that not all parts of Erlang/OTP has been ported. </p>
<p>
Notable things that work are: non-smp and smp emulators,
OSE signal interaction, crypto, asn1, run_erl/to_erl,
tcp, epmd, distribution and most if not all non-os
specific functionality of Erlang.</p>
<p>
Notable things that does not work are: udp/sctp, os_mon,
erl_interface, binding of schedulers.</p>
<p>
Own Id: OTP-11334</p>
</item>
<item>
<p>
Add the {active,N} socket option for TCP, UDP, and SCTP,
where N is an integer in the range -32768..32767, to
allow a caller to specify the number of data messages to
be delivered to the controlling process. Once the
socket's delivered message count either reaches 0 or is
explicitly set to 0 with inet:setopts/2 or by including
{active,0} as an option when the socket is created, the
socket transitions to passive ({active, false}) mode and
the socket's controlling process receives a message to
inform it of the transition. TCP sockets receive
{tcp_passive,Socket}, UDP sockets receive
{udp_passive,Socket} and SCTP sockets receive
{sctp_passive,Socket}. </p>
<p>
The socket's delivered message counter defaults to 0, but
it can be set using {active,N} via any gen_tcp, gen_udp,
or gen_sctp function that takes socket options as
arguments, or via inet:setopts/2. New N values are added
to the socket's current counter value, and negative
numbers can be used to reduce the counter value.
Specifying a number that would cause the socket's counter
value to go above 32767 causes an einval error. If a
negative number is specified such that the counter value
would become negative, the socket's counter value is set
to 0 and the socket transitions to passive mode. If the
counter value is already 0 and inet:setopts(Socket,
[{active,0}]) is specified, the counter value remains at
0 but the appropriate passive mode transition message is
generated for the socket.</p>
<p>
Thanks to Steve Vinoski</p>
<p>
Own Id: OTP-11368</p>
</item>
<item>
<p>
A new optional scheduler utilization balancing mechanism
has been introduced. For more information see the
<seealso marker="erl#+sub"><c>+sub</c></seealso> command
line argument.</p>
<p>
Characteristics impact: None, when not enabled. When
enabled, changed timing in the system, normally a small
overhead due to measuring of utilization and calculating
balancing information. On some systems, such as old
Windows systems, the overhead can be quite substantial.
This time measurement overhead highly depend on the
underlying primitives provided by the OS.</p>
<p>
Own Id: OTP-11385</p>
</item>
<item>
<p>
A call to either the <c>garbage_collect/1</c> BIF or the
<c>check_process_code/2</c> BIF may trigger garbage
collection of another processes than the process calling
the BIF. The previous implementations performed these
kinds of garbage collections without considering the
internal state of the process being garbage collected. In
order to be able to more easily and more efficiently
implement yielding native code, these types of garbage
collections have been rewritten. A garbage collection
like this is now triggered by an asynchronous request
signal, the actual garbage collection is performed by the
process being garbage collected itself, and finalized by
a reply signal to the process issuing the request. Using
this approach processes can disable garbage collection
and yield without having to set up the heap in a state
that can be garbage collected.</p>
<p>
The <seealso
marker="erts:erlang#garbage_collect/2"><c>garbage_collect/2</c></seealso>,
and <seealso
marker="erts:erlang#check_process_code/3"><c>check_process_code/3</c></seealso>
BIFs have been introduced. Both taking an option list as
last argument. Using these, one can issue asynchronous
requests.</p>
<p>
<c>code:purge/1</c> and <c>code:soft_purge/1</c> have
been rewritten to utilize asynchronous
<c>check_process_code</c> requests in order to
parallelize work.</p>
<p>
Characteristics impact: A call to the
<c>garbage_collect/1</c> BIF or the
<c>check_process_code/2</c> BIF will normally take longer
time to complete while the system as a whole wont be as
much negatively effected by the operation as before. A
call to <c>code:purge/1</c> and <c>code:soft_purge/1</c>
may complete faster or slower depending on the state of
the system while the system as a whole wont be as much
negatively effected by the operation as before.</p>
<p>
Own Id: OTP-11388 Aux Id: OTP-11535, OTP-11648 </p>
</item>
<item>
<p>
Cleanup 'Buckets' and 'Time left' fields in crashdump to
ease parsing.</p>
<p>
Own Id: OTP-11419</p>
</item>
<item>
<p>
Add sync option to file:open/2.</p>
<p>
The sync option adds the POSIX O_SYNC flag to the open
system call on platforms that support the flag or its
equivalent, e.g., FILE_FLAG_WRITE_THROUGH on Windows. For
platforms that don't support it, file:open/2 returns
{error, enotsup} if the sync option is passed in. Thank
to Steve Vinoski and Joseph Blomstedt</p>
<p>
Own Id: OTP-11498</p>
</item>
<item>
<p>
erlang:binary_to_term will now cost an appropriate amount
of reductions and will interrupt (yield) for reschedule
if the term is big. This avoids too long schedules when
binary_to_term is used. (Thanks to Svante Karlsson for
the original patch)</p>
<p>
Impact: Programs running binary_to_term on large binaries
will run more smoothly, but rescheduling will impact the
single process performance of the BIF. Single threaded
benchmarks might show degraded performance of the BIF,
while general system behaviour will be improved.</p>
<p>
Own Id: OTP-11535 Aux Id: OTP-11388 </p>
</item>
<item>
<p>
Added high resolution icon for windows. (Thanks to Daniel
Goertz for the inspiration.)</p>
<p>
Own Id: OTP-11560</p>
</item>
<item>
<p>
Migration of memory carriers has been enabled by default
on all ERTS internal memory allocators based on the
<seealso
marker="erts_alloc#alloc_util"><c>alloc_util</c></seealso>
framework except for <c>temp_alloc</c>. That is, <seealso
marker="erts_alloc#M_acul"><c>+M<S>acul
de</c></seealso> is default for these allocators. Note
that this also implies changed allocation strategies for
all of these allocators. They will all now use the
"address order first fit carrier best fit" strategy.</p>
<p>
By passing <c>+Muacul 0</c> on the command line, all
configuration changes made by this change will be
reverted.</p>
<p>
Characteristics impact: Improved memory characteristics
with a smaller memory footprint at the expense of a quite
small performance cost.</p>
<p>
Own Id: OTP-11604 Aux Id: OTP-10279 </p>
</item>
<item>
<p>A clarification has been added to the documentation of
<c>-on_load()</c> in the Reference Manual that it is only
recommended for loading NIF libraries.</p>
<p>
Own Id: OTP-11611</p>
</item>
<item>
<p><c>+fnaw</c> is now default when starting the
emulator; it used to be <c>+fnl</c>.</p>
<p>
*** POTENTIAL INCOMPATIBILITY ***</p>
<p>
Own Id: OTP-11612</p>
</item>
<item>
<p>
EEP43: New data type - Maps</p>
<p>
With Maps you may for instance:</p>
<taglist>
<tag/> <item><c>M0 = #{ a => 1, b => 2}, % create
associations</c></item>
<tag/><item><c>M1 = M0#{ a := 10 }, % update values</c></item>
<tag/><item><c>M2 = M1#{ "hi" =>
"hello"}, % add new associations</c></item>
<tag/><item><c>#{ "hi" := V1, a := V2, b := V3} = M2.
% match keys with values</c></item>
</taglist>
<p>
For information on how to use Maps please see Map Expressions in the
<seealso marker="doc/reference_manual:expressions#map_expressions">
Reference Manual</seealso>.</p>
<p>
The current implementation is without the following
features:</p>
<taglist>
<tag/><item>No variable keys</item>
<tag/><item>No single value access</item>
<tag/><item>No map comprehensions</item>
</taglist>
<p>
Note that Maps is <em>experimental</em> during OTP 17.0.</p>
<p>
Own Id: OTP-11616</p>
</item>
<item>
<p>
The previously deprecated driver API function
<c>driver_async_cancel()</c> has been removed. Due to
this, the driver API version has been bumped to 3.0.</p>
<p>
Thanks to Steve Vinoski.</p>
<p>
*** POTENTIAL INCOMPATIBILITY ***</p>
<p>
Own Id: OTP-11628</p>
</item>
<item>
<p>
Experimental "dirty scheduler" functionality has been
introduced. In order to try the functionality out, you
need to pass the command line argument
<c>--enable-dirty-schedulers</c> to <c>configure</c> when
building the system.</p>
<p>
Dirty schedulers can currently only be used by NIFs on a
system with SMP support. More information can be found in
the <seealso
marker="erl_nif#dirty_nifs"><c>erl_nif(3)</c></seealso>
documentation, the <seealso
marker="erl"><c>erl(1)</c></seealso> documentation, and
in the git commit comment of commit
'c1c03ae4ee50e58b7669ea88ec4d29c6b2b67c7b'.</p>
<p>
Note that the functionality is <em>experimental</em>, and
<em>not supported</em>. This functionality <em>will</em>
be subject to backward incompatible changes. You should
<em>not</em> enable the dirty scheduler functionality on
production systems. It is only provided for testing.</p>
<p>
Thanks to Steve Vinoski.</p>
<p>
Own Id: OTP-11629</p>
</item>
<item>
<p>
Improve reduction cost and yielding of
<c>term_to_binary</c>. The reduction cost is increased
and garbage collection is disabled during yield.</p>
<p>
Impact: Improves system responsiveness when
<c>term_to_binary</c> is called with large terms without
significant degradation of single threaded performance.</p>
<p>
Own Id: OTP-11648 Aux Id: OTP-11388 </p>
</item>
<item>
<p>
By default, the system's version of zlib will be used,
provided its version is 1.2.4 or higher; otherwise the
built-in zlib will be used. The built-in version of zlib
has been bumped to 1.2.8. (Use the
<c>--enable-builtin-zlib</c> option to <c>configure</c>
to force the use of the built-in zlib.)</p>
<p>
Own Id: OTP-11669</p>
</item>
<item>
<p>
The default float encoding in binary_to_term and
external_size has been changed to use minor_mode 1
instead of 0.</p>
<p>
*** POTENTIAL INCOMPATIBILITY ***</p>
<p>
Own Id: OTP-11738</p>
</item>
<item>
<p>
Introduced the <c>configure</c> option
<c>--with-assumed-cache-line-size=SIZE</c>. For more
information see <seealso
marker="doc/installation_guide:INSTALL"><c>$ERL_TOP/HOWTO/INSTALL.md</c></seealso>.</p>
<p>
Own Id: OTP-11742</p>
</item>
<item>
<p>
Halfword emulator is marked as deprecated. It still works
as before but is planned to be removed in a future major
release.</p>
<p>
Own Id: OTP-11777</p>
</item>
<item>
<p>
The external format for Maps has changed in a way that is
not compatible with the format used in OTP 17.0-rc1 and
OTP 17.0-rc2.</p>
<p>
*** POTENTIAL INCOMPATIBILITY ***</p>
<p>
Own Id: OTP-11782</p>
</item>
<item>
<p>
Fixed faulty make dependency that would make some make
versions fail while building gen_git_version.mk.</p>
<p>
Own Id: OTP-11784</p>
</item>
<item>
<p>
Introduced functionality for allowing old drivers and NIF
libraries to be loaded during a transition period. For
more information see <seealso
marker="erts:erl_driver#version_management">the version
management section in the <c>erl_driver(3)</c>
documentation</seealso> and <seealso
marker="erts:erl_nif#version_management">the version
management section in the <c>erl_nif(3)</c>
documentation</seealso>.</p>
<p>
Own Id: OTP-11799</p>
</item>
<item>
<p>
Support file paths longer than 259 characters on Windows.
Long absolute paths are automatically converted to UNC
format with a <c>\\?\</c> prefix which is the only way to
represent long paths. The 259 character limit still
applies for individual file names, relative paths and the
current working directory.</p>
<p>
Own Id: OTP-11813</p>
</item>
<item>
<p>
Document that escript:create/2 also accepts a 3-elements
tuple containing files and zip:create/3 options to build
a zip file.</p>
<p>
Thanks to Pierre Fenoll</p>
<p>
Own Id: OTP-11827</p>
</item>
<item>
<p>
Add systemd socket activation for epmd.</p>
<p>
Thanks to Matwey V. Kornilov</p>
<p>
Own Id: OTP-11829</p>
</item>
</list>
</section>
</section>
<section><title>Erts 5.10.4.1</title>
<section><title>Known Bugs and Problems</title>
<list>
<item>
<p>
When using gen_tcp:connect and the <c>fd</c> option with
<c>port</c> and/or <c>ip</c>, the <c>port</c> and
<c>ip</c> options were ignored. This has been fixed so
that if <c>port</c> and/or <c>ip</c> is specified
together with <c>fd</c> a bind is requested for that
<c>fd</c>. If <c>port</c> and/or <c>ip</c> is not
specified bind will not be called.</p>
<p>
Own Id: OTP-12061</p>
</item>
</list>
</section>
</section>
<section><title>Erts 5.10.4</title>
<section><title>Fixed Bugs and Malfunctions</title>
<list>
<item>
<p>
When normalizing paths, erl_prim_loader would always
convert backslash to forward slash. This is correct on
Windows, but not on other operating systems.
erl_prim_loader now checks which OS is running before
performing this conversion.</p>
<p>
Own Id: OTP-11170</p>
</item>
<item>
<p>
Fixed syslog defines and defined LOG_ERR for systems
without syslog.h. Thanks to Matt Lewandowsky.</p>
<p>
Own Id: OTP-11349</p>
</item>
<item>
<p>
Check all pattern arguments passed to binary:matches/2.
Thanks to Mike Sassak.</p>
<p>
Own Id: OTP-11350</p>
</item>
<item>
<p>
Fix two small silent rules omissions. Thanks to Anthony
Ramine.</p>
<p>
Own Id: OTP-11351</p>
</item>
<item>
<p>
Teach configure to detect if posix_memalign cannot align
to more than the system page size. </p>
<p>
For cross-compiled systems a new environment variable
called erl_xcomp_posix_memalign has been introduced to
indicate whether posix_memalign should be used.</p>
<p>
Own Id: OTP-11371</p>
</item>
<item>
<p>
Fix bsr bug occurring when shifting a huge number a huge
number of bits to the right. Thanks to Lars Hesel
Christensen.</p>
<p>
Own Id: OTP-11381</p>
</item>
<item>
<p>
Fix memory leak for distributed monitors</p>
<p>
Own Id: OTP-11410</p>
</item>
<item>
<p>
Fix various typos in erts, kernel and ssh. Thanks to
Martin Hässler.</p>
<p>
Own Id: OTP-11414</p>
</item>
<item>
<p>
Crashdumps initiated by out-of-memory on process spawn
could cause the beam to segfault during crashdump writing
due to invalid pointers.</p>
<p>
The pointers are invalid since the process creation never
finished. This fix removes these processes from the
printouts. Reported by Richard Carlsson.</p>
<p>
Own Id: OTP-11420</p>
</item>
<item>
<p>
Crash dumps from 64-bit Erlang machines would have all
memory addresses truncated to 32 bits, which could cause
trouble inspecting processes message queues and stacks in
the crashdump viewer.</p>
<p>
Own Id: OTP-11450</p>
</item>
<item>
<p>
Threads other than schedulers threads could make thread
unsafe accesses when support for migration of memory
carriers had been enabled, i.e., when the <seealso
marker="erts_alloc#M_acul"><c>+M<S>acul</c></seealso>
command line flag had been passed to <seealso
marker="erl"><c>erl</c></seealso>. This could cause
corruption of the VMs internal state.</p>
<p>
This bug was introduced in erts-5.10.2 when the support
for migration of memory carriers was introduced.</p>
<p>
Own Id: OTP-11456 Aux Id: OTP-10279 </p>
</item>
<item>
<p>
Fix bug in <c>binary_to_term</c> for invalid bitstrings
and very large binaries (>2Gb).</p>
<p>
Own Id: OTP-11479</p>
</item>
<item>
<p>
Under rare circumstances a process calling <seealso
marker="kernel:inet#close/1"><c>inet:close/1</c></seealso>,
<seealso
marker="kernel:gen_tcp#close/1"><c>gen_tcp:close/1</c></seealso>,
<seealso
marker="kernel:gen_udp#close/1"><c>gen_udp:close/1</c></seealso>,
or <seealso
marker="kernel:gen_sctp#close/1"><c>gen_sctp:close/1</c></seealso>
could hang in the call indefinitely.</p>
<p>
Own Id: OTP-11491</p>
</item>
<item>
<p>
Fix bug that could cause a 32-bit emulator to always
crash at start (since R16B01) depending on the alignment
of static data in the beam executable.</p>
<p>
Own Id: OTP-11496</p>
</item>
<item>
<p>
Fix benign bugs regarding bitstring compare. Only a
nuisance for debug and valgrind VM.</p>
<p>
Own Id: OTP-11501</p>
</item>
<item>
<p>
Silence warnings (Thanks to Anthony Ramine)</p>
<p>
Own Id: OTP-11517</p>
</item>
<item>
<p>
The default wordsize of the emulator (beam) is now
determined by compiler default on Mac OSX (Darwin). This
was previously forced to 32bits by the configure script
unless otherwise specified.</p>
<p>
Own Id: OTP-11521</p>
</item>
</list>
</section>
<section><title>Improvements and New Features</title>
<list>
<item>
<p>
A new memory allocation feature called "super carrier"
has been introduced. The super carrier feature can be
used in different ways. It can for example be used for
pre-allocation of all memory that the runtime system
should be able to use.</p>
<p>
By default the super carrier is disabled. It is enabled
by passing the <seealso
marker="erts:erts_alloc#MMscs"><c>+MMscs <size in
MB></c></seealso> command line argument. For more
information see the documentation of the <seealso
marker="erts:erts_alloc#MMsco"><c>+MMsco</c></seealso>,
<seealso
marker="erts:erts_alloc#MMscrfsd"><c>+MMscrfsd</c></seealso>,
<seealso
marker="erts:erts_alloc#MMscrpm"><c>+MMscrpm</c></seealso>,
<seealso
marker="erts:erts_alloc#MMscs"><c>+MMscs</c></seealso>,
<seealso
marker="erts:erts_alloc#Musac"><c>+MMusac</c></seealso>,
and, <seealso
marker="erts:erts_alloc#Mlpm"><c>+Mlpm</c></seealso>
command line arguments in the <seealso
marker="erts:erts_alloc"><c>erts_alloc(3)</c></seealso>
documentation.</p>
<p>
Since it is disabled by default there should be no impact
on system characteristics if not used.</p>
<p>
This change has been marked as a potential
incompatibility since the returned list when calling
<seealso
marker="erts:erlang#system_info_allocator_tuple"><c>erlang:system_info({allocator,
mseg_alloc})</c></seealso> now also include an
<c>{erts_mmap, _}</c> tuple as one element in the list.</p>
<p>
*** POTENTIAL INCOMPATIBILITY ***</p>
<p>
Own Id: OTP-11149</p>
</item>
<item>
<p>
Added erlang:system_info(ets_limit) to provide a way to
retrieve the runtime's maximum number of ETS tables.
Thanks to Steve Vinoski</p>
<p>
Own Id: OTP-11362</p>
</item>
<item>
<p>
Add new BIF os:unsetenv/1 which deletes an environment
variable. Thanks to Martin Hässler.</p>
<p>
Own Id: OTP-11446</p>
</item>
<item>
<p> Introduced a new guarantee regarding exit signals
from ports: </p><p> If the process calling one of the
synchronous port BIFs listed below is linked to the port
identified by the first argument, and the port exits
before sending the result of the port operation, the exit
signal issued due to this link will be received by the
processes before the BIF returns, or fail with an
exception due to the port not being open. </p><p> The
synchronous port BIFs are: </p> <list> <item><seealso
marker="erlang#port_close/1"><c>port_close/1</c></seealso></item>
<item><seealso
marker="erlang#port_command/2"><c>port_command/2</c></seealso></item>
<item><seealso
marker="erlang#port_command/3"><c>port_command/3</c></seealso></item>
<item><seealso
marker="erlang#port_connect/2"><c>port_connect/2</c></seealso></item>
<item><seealso
marker="erlang#port_control/3"><c>port_control/3</c></seealso></item>
<item><seealso
marker="erlang#port_call/3"><c>erlang:port_call/3</c></seealso></item>
<item><seealso
marker="erlang#port_info/1"><c>erlang:port_info/1</c></seealso></item>
<item><seealso
marker="erlang#port_info/2"><c>erlang:port_info/2</c></seealso></item>
</list> <p> Note that some ports under certain
circumstances unlink themselves from the calling process
before exiting, i.e. even though the process linked
itself to the port there might be no link triggering an
exit signal. </p> <p>Characteristics impact: The return
or exception from the synchronous port BIF will be
delayed if the port simultaneously exit due to some issue
unrelated to the outstanding synchronous port BIF call.
In all other cases characteristics are unchanged. </p>
<p>
Own Id: OTP-11489</p>
</item>
</list>
</section>
</section>
<section><title>Erts 5.10.3.1</title>
<section><title>Improvements and New Features</title>
<list>
<item>
<p>
Memory allocators will be able to create <c>sys_alloc</c>
carriers as fallback, if <c>mseg_alloc</c> cannot create
more carriers, on systems with <c>posix_memalign()</c>
support. This is similar to how it worked in pre-R16
releases.</p>
<p>
Windows systems will create carriers using
<c>_aligned_malloc()</c> and can by this use the new
optimized allocator header scheme introduced in R16 on
other platforms.</p>
<p>
Own Id: OTP-11318</p>
</item>
</list>
</section>
</section>
<section><title>Erts 5.10.3</title>
<section><title>Fixed Bugs and Malfunctions</title>
<list>
<item>
<p> The documentation of predefined types and built-in
types has been corrected. </p>
<p>
Own Id: OTP-11090</p>
</item>
<item>
<p>
Fix changing terminal parameters in to_erl</p>
<p>
Change the behaviour of to_erl to use TCSADRAIN instead
of TCSANOW when changing terminal parameters. This makes
the serial driver wait for the output queues to be empty
before applying the terminal parameter change. Thanks to
Stefan Zegenhagen.</p>
<p>
Own Id: OTP-11206</p>
</item>
<item>
<p>
The default value of {flush, boolean()} in erlang:halt/2
is documented to be 'true' if the status is an integer.
The implementation behaviour was reversed. The
Implementation is now corrected to adhere to the
documentation. Thanks to Jose Valim for reporting the
error.</p>
<p>
*** POTENTIAL INCOMPATIBILITY ***</p>
<p>
Own Id: OTP-11218</p>
</item>
<item>
<p>
Fix serious race bug in R16B01 that could cause PID
mix-ups when a lot of processes were spawned and
terminated in a very rapid pace on an SMP emulator with
at least two scheduler threads.</p>
<p>
Own Id: OTP-11225</p>
</item>
<item>
<p>
Validating a trace pattern with the option silent no
longer incorrectly enables/disables the silent option of
the calling process.</p>
<p>
Own Id: OTP-11232</p>
</item>
<item>
<p>
Fixed a bug where GCC 4.8 and later use a more aggressive
loop optimization algorithm that broke some previously
working code in the efile driver. Thanks to Tomas
Abrahamsson for reporting this issue.</p>
<p>
Own Id: OTP-11246</p>
</item>
<item>
<p>
Fixed bug when printing memory allocator acul option in
crash dump.</p>
<p>
Own Id: OTP-11264</p>
</item>
<item>
<p>
Opening a new compressed file on Windows could in rare
(random) cases result in {error,eisdir} or other error
codes although it should have succeeded. This is now
corrected.</p>
<p>
Own Id: OTP-11265</p>
</item>
<item>
<p>
Fixed a race condition when closing a trace port that
would cause the emulator to crash.</p>
<p>
Own Id: OTP-11290</p>
</item>
</list>
</section>
<section><title>Improvements and New Features</title>
<list>
<item>
<p>
There is a new somewhat experimental socket option
'netns' that can set the network namespace for a socket
on Linux:es where it is supported. See the documentation.</p>
<p>
Own Id: OTP-11157</p>
</item>
<item>
<p>
New allocator strategy <c>aoffcbf</c> (address order
first fit carrier best fit). Supports carrier migration
but with better CPU performance than <c>aoffcaobf</c>.</p>
<p>
Own Id: OTP-11174</p>
</item>
<item>
<p>
Introduced functionality for inspection of system and
build configuration.</p>
<p>
Own Id: OTP-11196</p>
</item>
<item>
<p>
Fix matching of floating point middle-endian machines.
Thanks to Johannes Weissl.</p>
<p>
Own Id: OTP-11201</p>
</item>
<item>
<p>
Fix compile error on ARM and GCC versions greater than
4.1.0. Thanks to Johannes Weissl.</p>
<p>
Own Id: OTP-11214</p>
</item>
<item>
<p>
run_erl: Redirect standard streams to /dev/null. Thanks
to Johannes Weissl.</p>
<p>
Own Id: OTP-11215</p>
</item>
<item>
<p>
Misc. corrections in documentation for erl_driver. Thanks
to Giacomo Olgeni.</p>
<p>
Own Id: OTP-11227</p>
</item>
<item>
<p>
Fix documentation regarding binary_part.</p>
<p>
Own Id: OTP-11239</p>
</item>
<item>
<p>
Make edlin understand a few important control keys.
Thanks to Stefan Zegenhagen.</p>
<p>
Own Id: OTP-11251</p>
</item>
<item>
<p>
Export type zlib:zstream/0. Thanks to Loic Hoguin.</p>
<p>
Own Id: OTP-11278</p>
</item>
<item>
<p>
Add erl option to set schedulers by percentages. </p>
<p>
For applications where measurements show enhanced
performance from the use of a non-default number of
emulator scheduler threads, having to accurately set the
right number of scheduler threads across multiple hosts
each with different numbers of logical processors is
difficult because the erl +S option requires absolute
numbers of scheduler threads and scheduler threads online
to be specified.</p>
<p>
To address this issue, add a +SP option to erl, similar
to the existing +S option but allowing the number of
scheduler threads and scheduler threads online to be set
as percentages of logical processors configured and
logical processors available, respectively. For example,
"+SP 50:25" sets the number of scheduler threads to 50%
of the logical processors configured, and the number of
scheduler threads online to 25% of the logical processors
available. The +SP option also interacts with any
settings specified with the +S option, such that the
combination of options "+S 4:4 +SP 50:50" (in either
order) results in 2 scheduler threads and 2 scheduler
threads online.</p>
<p>
Thanks to Steve Vinoski</p>
<p>
Own Id: OTP-11282</p>
</item>
<item>
<p>
Extend erl_driver interface with lock names</p>
<p>
Lock and thread names are already a feature in the driver
interface. This extension will let developers read these
names which eases debugging.</p>
<p>
Own Id: OTP-11303</p>
</item>
<item>
<p>
Fix incorrect values returned by integer_to_binary/2.
Thanks to Juan Jose Comellas.</p>
<p>
Own Id: OTP-11311</p>
</item>
<item>
<p>
Fix system_flag scheduling_statistics - disable . Thanks
to Steve Vinoski.</p>
<p>
Own Id: OTP-11317</p>
</item>
<item>
<p> The documentation of predefined types has been
corrected Thanks to Kostis Sagonas. </p>
<p>
Own Id: OTP-11321</p>
</item>
</list>
</section>
</section>
<section><title>Erts 5.10.2</title>
<section><title>Fixed Bugs and Malfunctions</title>
<list>
<item>
<p>
A bug in prim_inet has been corrected. If the port owner
was killed at a bad time while closing the socket port
the port could become orphaned hence causing port and
socket leaking. Reported by Fred Herbert, Dmitry Belyaev
and others.</p>
<p>
Own Id: OTP-10497 Aux Id: OTP-10562 </p>
</item>
<item>
<p>
Compilation fixes for NetBSD. Thanks to YAMAMOTO Takashi.</p>
<p>
Own Id: OTP-10941</p>
</item>
<item>
<p>
Fixed a race condition when using delayed_write when
writing to a file which would cause the same data to be
written multiple times.</p>
<p>
Own Id: OTP-10984</p>
</item>
<item>
<p>
Fix small memory leak from tracing with option
<c>meta</c>.</p>
<p>
Own Id: OTP-10997</p>
</item>
<item>
<p>
Correct typo in erlsrv usage. Thanks to Bryan Hunter</p>
<p>
Own Id: OTP-11002</p>
</item>
<item>
<p>
ct_run: delete unused function. Thanks to Tuncer Ayaz.</p>
<p>
Own Id: OTP-11003</p>
</item>
<item>
<p>
Corrections to run_erl/to_erl handshake behaviour.</p>
<p>
Own Id: OTP-11012</p>
</item>
<item>
<p>
Fix typo in type: erlang:process_info_item(). Thanks to
Andrew Tunnell-Jones.</p>
<p>
Own Id: OTP-11024</p>
</item>
<item>
<p>
Fix src/dest overlap issue in ttsl driver. Thanks to
Steve Vinoski.</p>
<p>
Own Id: OTP-11064</p>
</item>
<item>
<p>
When sending to a port using <c>erlang:send(Port, Msg,
[nosuspend])</c>, the send operation was performed
synchronously. This bug has now been fixed.</p>
<p>
Own Id: OTP-11076 Aux Id: OTP-10336 </p>
</item>
<item>
<p>
When converting a faulty binary to a list with
unicode:characters_to_list, the error return value could
contain a faulty "rest", i.e. the io_list of characters
that could not be converted was wrong. This happened only
if input was a sub binary and conversion was from utf8.
This is now corrected.</p>
<p>
Own Id: OTP-11080</p>
</item>
<item>
<p>
Runtime system could crash when reporting stale
<c>driver_select()</c>.</p>
<p>
Own Id: OTP-11084</p>
</item>
<item>
<p>
Fix lock order violation for memory instrumentation
(+Mim, +Mis, +Mit).</p>
<p>
Own Id: OTP-11085</p>
</item>
<item>
<p>
Fixed some compilation warnings on miscellaneous
platforms. Thanks to Anthony Ramine.</p>
<p>
Own Id: OTP-11086</p>
</item>
<item>
<p>
Fixed issue when flushing i/o during shutdown on windows
where the Erlang VM would sometime hang due to a race
condition.</p>
<p>
Own Id: OTP-11096</p>
</item>
<item>
<p>
Fixed issue where repeated calls to erlang:nodes() could
cause unnecessary contention in the dist_table lock.</p>
<p>
Own Id: OTP-11097</p>
</item>
<item>
<p>
Properly guard WIDE_TAG use with HAVE_WCWIDTH in
ttsl_drv. Thanks to Anthony Ramine</p>
<p>
Own Id: OTP-11106</p>
</item>
<item>
<p>
Fix some Makefile rules that didn't support silent rules.
Thanks to Anthony Ramine.</p>
<p>
Own Id: OTP-11111</p>
</item>
<item>
<p>
Fix receive support in erl_eval with a BEAM module.
Thanks to Anthony Ramine.</p>
<p>
Own Id: OTP-11137</p>
</item>
<item>
<p>
erlang:now() could suddenly jump ~24 days into the future
on Windows. This is now corrected. Thanks to Garret Smith
for reporting and testing fixes.</p>
<p>
Own Id: OTP-11146</p>
</item>
<item>
<p>
erlang:term_to_binary will now cost an appropriate amount
of reductions and will interrupt (yield) for reschedule
if the term is big. This avoids too long schedules when
term_to_binary is used. </p>
<p>
Impact: Programs running term_to_binary on large terms
will run more smothly, but rescheduling will impact the
single process performance of the BIF. Single threaded
benchmarks will show degraded performance of the BIF when
called with very large terms, while general system
behaviour will be improved. The overhead for allowing
restart and reduction counting also degrades local
performance of the BIF with between 5% and 10% even for
small terms.</p>
<p>
Own Id: OTP-11163</p>
</item>
</list>
</section>
<section><title>Improvements and New Features</title>
<list>
<item>
<p>
Replaced the lock protecting gathering of garbage
collection statistics with a lock-free solution.</p>
<p>
Own Id: OTP-10271 Aux Id: kunagi-108
[04c5410f-9cc4-4696-8639-36bf98686c7a-7] </p>
</item>
<item>
<p>Support for migration of memory carriers between
memory allocator instances has been introduced.</p>
<p>By default this feature is not enabled and do not
effect the characteristics of the system. When enabled it
has the following impact on the characteristics of the
system:</p> <list> <item>Reduced memory footprint when
the memory load is unevenly distributed between scheduler
specific allocator instances.</item> <item>Depending on
the default allocaton strategy used on a specific
allocator there might or might not be a slight
performance loss.</item> <item>When enabled on the
<c>fix_alloc</c> allocator, a different strategy for
management of fix blocks will be used.</item> <item>The
information returned from <seealso
marker="erlang:system_info_allocator_tuple"><c>erlang:system_info({allocator,
A})</c></seealso>, and <seealso
marker="erlang:system_info_allocator_sizes"><c>erlang:system_info({allocator_sizes,
A})</c></seealso> will be slightly different when this
feature has been enabled. An <c>mbcs_pool</c> tuple will
be present giving information about abandoned carriers,
and in the <c>fix_alloc</c> case no <c>fix_types</c>
tuple will be present. </item></list>
<p>For more information, see the documentation of the
<seealso
marker="erts_alloc#M_acul"><c>+M<S>acul</c></seealso>
command line argument.</p>
<p>
Own Id: OTP-10279</p>
</item>
<item>
<p>
Change specs for spawn_opt to use the process_level()
type declaration instead of re-defining it in various
places. Thanks to Kostis Sagonas.</p>
<p>
Own Id: OTP-11008</p>
</item>
<item>
<p> Postscript files no longer needed for the generation
of PDF files have been removed. </p>
<p>
Own Id: OTP-11016</p>
</item>
<item>
<p>Erlang source files with non-ASCII characters are now
encoded in UTF-8 (instead of latin1).</p>
<p>
Own Id: OTP-11041 Aux Id: OTP-10907 </p>
</item>
<item>
<p>
Optimization of simultaneous <c>inet_db</c> operations on
the same socket by using a lock free implementation.</p>
<p>
Impact on the characteristics of the system: Improved
performance.</p>
<p>
Own Id: OTP-11074</p>
</item>
<item>
<p>
The <c>high_msgq_watermark</c> and
<c>low_msgq_watermark</c> <c>inet</c> socket options
introduced in OTP-R16A could only be set on TCP sockets.
These options are now generic and can be set on all types
of sockets.</p>
<p>
Own Id: OTP-11075 Aux Id: OTP-10336 </p>
</item>
<item>
<p>A new better algorithm for management of the process,
and port tables has been introduced.</p>
<p>Impact on the characteristics of the system:</p>
<list> <item>The new algorithm ensures that both insert
and delete operations can be made in O(1) time
complexity. Previously used algorithm either caused
insert or delete to be O(N).</item> <item>The new
algorithm will also ensure that reuse of identifiers will
be less frequent than when the old algorithm was
used.</item> <item>Previously used algorithm ensured that
the latest created identifier compared as the largest
when comparing two identifiers of the same type that had
been created on the same node as long as no identifiers
had been reused. Since identifiers can be reused quite
fast, one has never been able to rely on this property.
Due to the introduction of this new algorithm this
property will not hold even if no identifiers has been
reused yet. This could be considered as an
incompatibility.</item> </list>
<p>Due to the above mensioned potential incompatibility,
it will still be possible to enable the old algorithm for
some time. The command line argument <seealso
marker="erl#+P"><c>+P legacy</c></seealso> will enable
the old algorithm on the process table, and <seealso
marker="erl#+Q"><c>+Q legacy</c></seealso> will do the
same for the port table. These command line arguments are
however deprecated as of their introduction and have been
scheduled for removal in OTP-R18.</p>
<p>
*** POTENTIAL INCOMPATIBILITY ***</p>
<p>
Own Id: OTP-11077</p>
</item>
<item>
<p>
Support wide characters in the shell through wcwidth().
Thanks to Anthony Ramine. Reported by Loïc Hoguin.</p>
<p>
Own Id: OTP-11088</p>
</item>
<item>
<p>
Added total used memory for each process in erlang crash
dumps.</p>
<p>
Own Id: OTP-11098</p>
</item>
<item>
<p>
Added support for hipe on Raspberry Pi (armv6l). Thanks
to Klaus Alfert.</p>
<p>
Own Id: OTP-11125</p>
</item>
<item>
<p>
Remove 'query' from the list of reserved words in docs.
Thanks to Matthias Endler and Loïc Hoguin.</p>
<p>
Own Id: OTP-11158</p>
</item>
<item>
<p>
Lift static limitation (FD_SETSIZE) for file descriptors
on Mac OS X. (Thanks to Anthony Ramine)</p>
<p>
Own Id: OTP-11159</p>
</item>
</list>
</section>
<section><title>Known Bugs and Problems</title>
<list>
<item>
<p>Miscellaneous native code in OTP misbehave either due
to lengthy execution, or due to not bumping reductions
properly. Problems typically occur when passing huge sets
of data to a misbehaving BIF. Fixing this has high
priority and is being worked on, but there will remain
issues like this for some time.</p>
<p>In order to alleviate problems with scheduling which
might occur when executing misbehaving native code, the
command line argument <seealso
marker="erl#+sfwi">+sfwi</seealso> has been
introduced.</p>
<p>By default this feature is disabled and you are
advised not to enable it if you do not encounter problems
with misbehaving native code.</p>
<p>When enabled it has the following impact on the
characteristics of the system:</p> <list> <item>Work will
always be distributed between schedulers even when
executing misbehaving native code.</item> <item>It may
cause an increased amount of processes and/or ports
bouncing between schedulers which in turn will cause a
performance loss.</item> <item>It may cause reduced
performance due to reduced or lost work compaction when
all schedulers do not execute under full load.</item>
<item>An increased contention on run queue locks.</item>
</list>
<p>
Own Id: OTP-11164</p>
</item>
</list>
</section>
</section>
<section><title>Erts 5.10.1.2</title>
<section><title>Fixed Bugs and Malfunctions</title>
<list>
<item>
<p>
A bug in the implementation of offline schedulers has
been fixed. The bug was introduced in OTP-R16A/ERTS-5.10,
and caused work-stealing between schedulers to fail. This
in turn, caused work to accumulate in some run-queues.
The bug was only triggered when there were offline
schedulers in the system, i.e., when the amount of online
schedulers was less than the total amount of schedulers.
The effect of the bug got more severe the larger amount
of offline schedulers the system had.</p>
<p>
Own Id: OTP-11022 Aux Id: OTP-9892 </p>
</item>
</list>
</section>
</section>
<section><title>Erts 5.10.1.1</title>
<section><title>Fixed Bugs and Malfunctions</title>
<list>
<item>
<p>
The BIF <seealso
marker="erlang#is_process_alive/1"><c>is_process_alive/1</c></seealso>
could prematurely return <c>false</c> while the process
being inspected was terminating. This bug was introduced
in ERTS-5.10.</p>
<p>
Own Id: OTP-10926</p>
</item>
<item>
<p>Fix a problem in <c>erlang:delete_element/2</c> where
the call could corrupt one word of stack if the heap and
stack met during call.</p>
<p>
Own Id: OTP-10932</p>
</item>
<item>
<p>
The <seealso
marker="erl#+sws"><c>+sws<value></c></seealso> and
<seealso
marker="erl#+swt"><c>+swt<value></c></seealso>
system flags failed if no white space were passed between
the parameter and value parts of the flags. Upon failure,
the runtime system refused to start.</p>
<p>
Own Id: OTP-11000</p>
</item>
</list>
</section>
<section><title>Improvements and New Features</title>
<list>
<item>
<p>
Scheduler threads will now by default be less eager
requesting wakeup due to certain cleanup operations. This
can also be controlled using the <seealso
marker="erl#+swct"><c>+swct</c></seealso> command line
argument of <seealso
marker="erl"><c>erl(1)</c></seealso>.</p>
<p>
Own Id: OTP-10994</p>
</item>
</list>
</section>
</section>
<section><title>Erts 5.10.1</title>
<section><title>Fixed Bugs and Malfunctions</title>
<list>
<item>
<p>
Threads created internally in the runtime system by
vanilla, fd, and spawn drivers on Windows systems could
make thread unsafe calls to <c>driver_select()</c>.</p>
<p>
Own Id: OTP-10802</p>
</item>
<item>
<p>
Threads created internally in the runtime system by the
vanilla, fd, and spawn drivers on Windows systems could
make unsafe memory accesses to driver data after port had
terminated.</p>
<p>
Own Id: OTP-10803</p>
</item>
<item>
<p>
The runtime system could crash when flushing data to
standard out or standard error on Windows.</p>
<p>
Own Id: OTP-10807</p>
</item>
<item>
<p>Bugs due to the port optimizations introduced in
erts-5.10/OTP-R16A have been fixed:</p> <list>
<item>Memory leak when terminating ports</item>
<item>Memory leak when reaching the system limit of
maximum amount of concurrently existing ports</item>
<item>Crashs due to missing, or late test of bad port
handle</item> <item>The newly introduced driver API
function <c>erl_drv_busy_msgq_limits()</c> could not be
used by dynamically linked in drivers on Windows</item>
</list>
<p>
Own Id: OTP-10809 Aux Id: OTP-10336 </p>
</item>
<item>
<p>
Fix <c>{packet,httph}</c> header capitalization for
unrecognized header fields longer than 20 charachters
such as <c>Sec-Websocket-Version</c>. The limit is simply
raised from 20 to 50 characters with the hope that valid
headers longer than 50 are not used.</p>
<p>
Own Id: OTP-10824</p>
</item>
<item>
<p>
Fix rounding issues in float_to_list/1,2. Thanks to Serge
Aleynikov</p>
<p>
Own Id: OTP-10837</p>
</item>
<item>
<p>
Fix memory leak in file driver introduced in R16A.</p>
<p>
Own Id: OTP-10841</p>
</item>
<item>
<p>A bug in an ERTS internal queue implementation could
cause the loss of a wake up signal to a consumer thread.
This has now been fixed.</p>
<p>The effect of this bug, when triggered, was often only
a small or even no delay of certain operations. This
since, threads often are woken due to other unrelated
reasons. However, if the consumer thread was not woken
due to other reasons when the bug was triggered, these
operations could be left hanging, potentially for ever.
Such effects seems to have been very rare, but we have on
at least one occasion gotten a report about such an
issue.</p>
<p>Operations potentially effected by this bug:</p>
<taglist> <tag>Inspection of memory allocation
status</tag> <item>The Erlang process calling
<c>erlang:memory/[0,1]</c>, or
<c>erlang:system_info({allocator|allocator_sizes, _})</c>
could potentially hang waiting for responses from
involved threads.</item> <tag>Async thread pool
jobs</tag> <item>An async thread pool job request and/or
reply could potentially be left hanging. In OTP this only
effected file operations, but user implemented drivers
using the async thread pool were also effected. In the
file operation case, this would typically translate into
an Erlang process potentially hanging on the file
operation.</item> <tag>Shutting down the runtime
system</tag> <item>Due to the issue with the async thread
pool mentioned above, flushing of I/O while terminating
the runtime system could also potentially hang.</item>
<tag>ETS memory deallocation</tag> <item>Scheduled jobs
handling deallocation of the main structure of an ETS
table could potentially hang. This more or less only
translates into minor memory leaks.</item> <tag>Shutting
down distribution</tag> <item>The distribution shutdown
phase used when manually shutting down the distribution,
i.e., when calling <c>net_kernel:stop()</c>, could
potentially hang.</item> </taglist>
<p>
Own Id: OTP-10854</p>
</item>
<item>
<p>
OS X Snow Leopard now only uses write, as writev does not
work properly on very large files.</p>
<p>
Own Id: OTP-10858</p>
</item>
<item>
<p>
Fixed a bug where line oriented file I/O using read_ahead
was very slow for files with very large difference in
line length.</p>
<p>
Own Id: OTP-10859</p>
</item>
<item>
<p>
In erts-5.10 (R16A) faulty hashvalues were produced for
non-ASCII atoms (characters in byte-range 128..255). This
has now been fixed and hashvalues conforms to previous
OTP releases.</p>
<p>
Own Id: OTP-10860</p>
</item>
<item>
<p>
Fixes of memory accesses that might be thread unsafe when
the runtime system has been linked against third-party
libraries for atomic memory operations during the build.
Most builds are uneffected by this bug. If triggered, the
runtime system will most likely crash more or less
immediately.</p>
<p>
Own Id: OTP-10875 Aux Id: OTP-10854 </p>
</item>
<item>
<p>
Fixed a bug where it was longer possible to give the +sws
proposal flag to non-smp emulators.</p>
<p>
Own Id: OTP-10881 Aux Id: seq12258 </p>
</item>
<item>
<p>
Faulty type to bytes read for ReadFile on Windows. This
could cause windows systems to misbehave. The correct
type is now used.</p>
<p>
Own Id: OTP-10890</p>
</item>
<item>
<p>
Change default max ports for Windows to 8192. Having a
too large value caused Windows to not be able to recover
properly. If you want to use another value, pass <c>+Q
Value</c> to erl or werl.</p>
<p>
Own Id: OTP-10892</p>
</item>
<item>
<p>
Fix rare crash on halfword vm during code loading.</p>
<p>
Own Id: OTP-10896</p>
</item>
</list>
</section>
<section><title>Improvements and New Features</title>
<list>
<item>
<p>
Tuple funs (deprecated in R15B) are no longer supported.</p>
<p>
*** POTENTIAL INCOMPATIBILITY ***</p>
<p>
Own Id: OTP-10170</p>
</item>
<item>
<p>
Added four new bifs, <c>erlang:binary_to_integer/1,2</c>,
<c>erlang:integer_to_binary/1</c>,
<c>erlang:binary_to_float/1</c> and
<c>erlang:float_to_binary/1,2</c>. These bifs work
similarly to how their list counterparts work, except
they operate on binaries. In most cases converting from
and to binaries is faster than converting from and to
lists. </p>
<p>
These bifs are auto-imported into erlang source files and
can therefore be used without the <c>erlang</c> prefix.</p>
<p>
Own Id: OTP-10300 Aux Id: kunagi-74 [74] </p>
</item>
<item>
<p>
The experimental support for packages has been removed.</p>
<p>
Own Id: OTP-10348 Aux Id: kunagi-316 [227] </p>
</item>
<item>
<p>The driver API function <seealso
marker="erl_driver#erl_drv_consume_timeslice"><c>erl_drv_consume_timeslice()</c></seealso>,
and the NIF API function <seealso
marker="erl_nif#enif_consume_timeslice"><c>enif_consume_timeslice()</c></seealso>
have been introduced.</p>
<p>These functions are provided in order to better
support co-operative scheduling, improve system
responsiveness, and to make it easier to prevent
misbehaviors of the VM due to a process or port
monopolizing a scheduler thread. They can be used when
dividing lengthy work into a number of repeated calls
without the need to use threads.</p>
<p>
Own Id: OTP-10810</p>
</item>
<item>
<p>
The list_to_integer/2 bif has been optimized when used
with bases other than 10.</p>
<p>
Own Id: OTP-10834 Aux Id: kunagi-74 [74] </p>
</item>
<item>
<p>
The git commit sha of the HEAD commit is now added to the
Erlang shell when compiling a non-released Erlang
version.</p>
<p>
Own Id: OTP-10838</p>
</item>
<item>
<p>Change caching policy for memory segment allocator.
For instance, prefer sbc segments over mbc segments,
caching policy is time-arrow aware, evicting older cached
segments to store newer segments. </p> <p>The default
number of cachable segment has been increased from five
to ten segments. This can be modified, same as before,
with the command line option <c>+MMmcs 5</c></p>
<p>Impact: Increased speed for processing on larger
objects, e.g. binaries. Slight increase of mapped and
resident memory. Tune your system with memory options to
<c>erl</c> for best performance.</p>
<p>
Own Id: OTP-10840</p>
</item>
<item>
<p>
Updated config.sub and config.guess to latest version
from gnu.org</p>
<p>
Own Id: OTP-10848</p>
</item>
<item>
<p>
Add an xcomp file for Blue Gene/Q. Thanks to Kostis
Sagonas.</p>
<p>
Own Id: OTP-10849</p>
</item>
<item>
<p>
Cleanup of documentation of the type language. Thanks to
Kostis Sagonas.</p>
<p>
Own Id: OTP-10850</p>
</item>
<item>
<p>
Change the return value of hipe_bifs:remove_refs_from/1.
Thanks to Kostis Sagonas.</p>
<p>
Own Id: OTP-10851</p>
</item>
<item>
<p>
As of ERTS-5.10/OTP-R16A node names passed in the EPMD
protocol are required to be encoded in UTF-8. Since EPMD
previously accepted latin1 encoded node names this is an
incompatibility. However, since Erlang nodes always have
required characters in node names to be 7-bit ASCII
characters (and still do require this), this
incompatibility should not effect anyone using EPMD as an
Erlang Port Mapper Daemon.</p>
<p>
*** POTENTIAL INCOMPATIBILITY ***</p>
<p>
Own Id: OTP-10872 Aux Id: OTP-10753 </p>
</item>
<item>
<p>
The +pc flag to erl can be used to set the range of
characters considered printable. This affects how the
shell and io:format("~tp",...) functionality does
heuristic string detection. More can be read in STDLIB
users guide.</p>
<p>
Own Id: OTP-10884</p>
</item>
<item>
<p>
Fix a number of type cast errors related to formatted
printing on Win64 that can potentially cause problem when
the Erlang VM exceeds 4 GB of ram. (Thanks to Blaine
Whittle for the original patch)</p>
<p>
Own Id: OTP-10887</p>
</item>
<item>
<p>
The effect of the deprecated environment variable
ERL_MAX_PORTS had been removed premeturely. It has now
been readded. Note that this is still scheduled to be
released in R17B.</p>
<p>
Own Id: OTP-10895</p>
</item>
</list>
</section>
</section>
<section><title>Erts 5.10</title>
<section><title>Fixed Bugs and Malfunctions</title>
<list>
<item>
<p>
Set new peeled off SCTP socket to nonblocking socket
(Thanks to Jonas Falkevik)</p>
<p>
Own Id: OTP-10491</p>
</item>
<item>
<p>
Fix various typos (thanks to Tuncer Ayaz)</p>
<p>
Own Id: OTP-10611</p>
</item>
<item>
<p>
Fix fd leak when using async thread pool</p>
<p> When using the async thread pool, if an erlang
process asks to open a file and it gets shutdown/killed
while the file:open/2 call hasn't returned, it's possible
to leak a file descriptor against the target file. This
has now been fixed. (Thanks to Filipe David Manana)</p>
<p>
Own Id: OTP-10677</p>
</item>
<item>
<p>
Use sys/types.h instead of string.h to pull ssize_t
definition to erl_driver.h. This fixes build issue on
NetBSD. (Thanks to Yamamoto Takashi).</p>
<p>
Own Id: OTP-10699</p>
</item>
<item>
<p>
Arguments given with the -run or -s flags to erl are now
translated according to the file name encoding mode of
the runtime system.</p>
<p>
Own Id: OTP-10702</p>
</item>
<item>
<p>
The octet counters in the gen_tcp/inet interface could
behave in unexpected ways on 64bit platforms. The
behaviour is now as expected.</p>
<p>
Own Id: OTP-10746</p>
</item>
<item>
<p>
Certain linux kernels, most notably in redhat and CentOS
distribution, had a bug in writev which generated an
infinite loop in the tcp code of the VM. The bug is now
worked around.</p>
<p>
Own Id: OTP-10747</p>
</item>
<item>
<p>
A process that got killed (got an exit signal) while
operating on a compresseed file, could cause a
segmentation fault in the VM. This is now corrected.
Thanks to Filipe David Manana for identifying the problem
and submitting a solution.</p>
<p>
Own Id: OTP-10748</p>
</item>
<item>
<p>
Windows previously used three digit exponent in
formatting which caused difference between platforms, as
can be seen by float_to_list/1. This has now been fixed.</p>
<p>
Own Id: OTP-10751</p>
</item>
</list>
</section>
<section><title>Improvements and New Features</title>
<list>
<item>
<p>
A boolean socket option 'ipv6_v6only' for IPv6 sockets
has been added. The default value of the option is OS
dependent, so applications aiming to be portable should
consider using <c>{ipv6_v6only,true}</c> when creating an
<c>inet6</c> listening/destination socket, and if
neccesary also create an <c>inet</c> socket on the same
port for IPv4 traffic. See the documentation.</p>
<p>
Own Id: OTP-8928 Aux Id: kunagi-193 [104] </p>
</item>
<item>
<p>It is now allowed to define stubs for BIFs, to allow
type specs to be written for BIFs. For example, if there
is BIF called <c>lists:member/2</c>, a dummy definition
of <c>lists:member/2</c> is now allowed.</p>
<p>
Own Id: OTP-9861</p>
</item>
<item>
<p>Process optimizations. The most notable:</p> <list>
<item>New internal process table implementation allowing
for both parallel reads as well as writes. Especially
read operations have become really cheap. This reduce
contention in various situations. For example when,
spawning processes, terminating processes, sending
messages, etc.</item> <item>Optimizations of run queue
management reducing contention.</item>
<item>Optimizations of process internal state changes reducing
contention.</item> </list> <p>These changes imply changes
of the characteristics the system. Most notable: changed
timing in the system.</p>
<p>
Own Id: OTP-9892 Aux Id: OTP-10167 </p>
</item>
<item>
<p>
Non-blocking code loading. Earlier when an Erlang module
was loaded, all other execution in the VM were halted
while the load operation was carried out in single
threaded mode. Now modules are loaded without blocking
the VM. Processes may continue executing undisturbed in
parallel during the entire load operation. The load
operation is completed by making the loaded code visible
to all processes in a consistent way with one single
atomic instruction. Non-blocking code loading will
improve realtime characteristics when modules are
loaded/upgraded on a running SMP system.</p>
<p>
Own Id: OTP-9974</p>
</item>
<item>
<p>In the SMP emulator, turning on and off tracing will
no longer take down the system to single-scheduling. </p>
<p>
Own Id: OTP-10122</p>
</item>
<item>
<p>Remove VxWorks support</p>
<p>
Own Id: OTP-10146</p>
</item>
<item>
<p>
Added a general framework for executing benchmarks of
Erlang/OTP. Benchmarks for the Erlang VM and mnesia have
been incorporated in the framework. </p>
<p>
For details about how to add more benchmarks see
$ERL_TOP/HOWTO/BENCHMARKS.md in the source distribution.</p>
<p>
Own Id: OTP-10156</p>
</item>
<item>
<p>
Optimized deletion of ETS-tables which significantly
improves performance when large amounts of temporary
tables are used.</p>
<p>
This change imply changes of the characteristics the
system. Most notable: changed timing in the system.</p>
<p>
Own Id: OTP-10167 Aux Id: OTP-9892 </p>
</item>
<item>
<p>
Tuple funs (deprecated in R15B) are no longer supported.</p>
<p>
*** POTENTIAL INCOMPATIBILITY ***</p>
<p>
Own Id: OTP-10170</p>
</item>
<item>
<p>
New internal header scheme for allocators</p>
<p>
Impact: Reduces size on object allocated in multiblock
carriers by one word</p>
<p>
Own Id: OTP-10273 Aux Id: kunagi-20 [20] </p>
</item>
<item>
<p>Major port improvements. The most notable:</p> <list>
<item>New internal port table implementation allowing for
both parallel reads as well as writes. Especially read
operations have become really cheap.This reduce
contention in various situations. For example when,
creating ports, terminating ports, etc. </item>
<item>Dynamic allocation of port structures. This allow
for a much larger maximum amount of ports allowed as a
default. The previous default of 1024 has been raised to
65536. Maximum amount of ports can be set using the
<seealso marker="erts:erl#+Q">+Q</seealso> command line
flag of <seealso marker="erts:erl">erl(1)</seealso>. The
previously used environment variable <c>ERL_MAX_PORTS</c>
has been deprecated and scheduled for removal in
OTP-R17.</item> <item>Major rewrite of scheduling of port
tasks. Major benefits of the rewrite are reduced
contention on run queue locks, and reduced amount of
memory allocation operations needed. The rewrite was also
necessary in order to make it possible to schedule
signals from processes to ports.</item> <item>Improved
internal thread progress functionality for easy
management of unmanaged threads. This improvement was
necessary for the rewrite of the port task
scheduling.</item> <item>Rewrite of all process to port
signal implementations in order to make it possible to
schedule those operations. All port operations can now be
scheduled which allows for reduced lock contention on the
port lock as well as truly asynchronous communication
with ports.</item> <item>Optimized lookup of port handles
from drivers.</item> <item>Optimized driver lookup when
creating ports.</item> <item>Preemptable <seealso
marker="erts:erlang#ports-0">erlang:ports/0</seealso>
BIF.</item> <item>Improving responsiveness by bumping
reductions for a process calling a driver callback
directly.</item> </list>
<p>These changes imply changes of the characteristics of
the system. The most notable:</p> <taglist> <tag>Order of
signal delivery</tag> <item>The previous implementation
of the VM has delivered signals from processes to ports
in a synchronous stricter fashion than required by the
language. As of ERTS version 5.10, signals are truly
asynchronously delivered. The order of signal delivery
still adheres to the requirements of the language, but
only to the requirements. That is, some signal sequences
that previously always were delivered in one specific
order may now from time to time be delivered in different
orders. This may cause Erlang programs that have made
<em>false assumptions</em> about signal delivery order to
fail even though they previously succeeded. For more
information about signal ordering guarantees, see the
chapter on <seealso
marker="erts:communication">communication</seealso> in
the ERTS user's guide. The <seealso
marker="erts:erl#+n">+n</seealso> command line flag of
<seealso marker="erts:erl">erl(1)</seealso> can be
helpful when trying to find signaling order bugs in
Erlang code that have been exposed by these
changes.</item> <tag>Latency of signals sent from
processes to ports</tag> <item>Signals from processes to
ports where previously always delivered immediately. This
kept latency for such communication to a minimum, but it
could cause lock contention which was very expensive for
the system as a whole. In order to keep this latency low
also in the future, most signals from processes to ports
are by default still delivered immediately as long as no
conflicts occur. Such conflicts include not being able to
acquire the port lock, but also include other conflicts.
When a conflict occur, the signal will be scheduled for
delivery at a later time. A scheduled signal delivery may
cause a higher latency for this specific communication,
but improves the overall performance of the system since
it reduce lock contention between schedulers. The default
behavior of only scheduling delivery of these signals on
conflict can be changed by passing the <seealso
marker="erts:erl#+spp">+spp</seealso> command line flag
to <seealso marker="erts:erl">erl(1)</seealso>. The
behavior can also be changed on port basis using the
<seealso
marker="erts:erlang#open_port_parallelism">parallelism</seealso>
option of the <seealso
marker="erts:erlang#open_port-2">open_port/2</seealso>
BIF.</item> <tag>Execution time of the
<c>erlang:ports/0</c> BIF</tag> <item>Since <seealso
marker="erts:erlang#ports-0">erlang:ports/0</seealso> now
can be preempted, the responsiveness of the system as a
whole has been improved. A call to <c>erlang:ports/0</c>
may, however, take a much longer time to complete than
before. How much longer time heavily depends on the
system load.</item> <tag>Reduction cost of calling driver
callbacks</tag> <item>Calling a driver callback is quite
costly. This was previously not reflected in reduction
cost at all. Since the reduction cost now has increased,
a process performing lots of direct driver calls will be
scheduled out more frequently than before.</item>
</taglist>
<p><em>Potential incompatibilities</em>:</p> <list>
<item><c>driver_send_term()</c> has been deprecated and
has been scheduled for removal in OTP-R17. Replace usage
of <c>driver_send_term()</c> with usage of <seealso
marker="erts:erl_driver#erl_drv_send_term">erl_drv_send_term()</seealso>.</item>
<item><c>driver_output_term()</c> has been deprecated and
has been scheduled for removal in OTP-R17. Replace usage
of <c>driver_output_term()</c> with usage of <seealso
marker="erts:erl_driver#erl_drv_output_term">erl_drv_output_term()</seealso>.</item>
<item>The new function <seealso
marker="erts:erl_driver#erl_drv_busy_msgq_limits">erl_drv_busy_msgq_limits()</seealso>
has been added in order to able to control management of
port queues.</item> </list>
<p>The <seealso
marker="erts:erl_driver#version_management">driver API
version</seealso> has been bumped to 2.1 from 2.0 due to
the above changes in the driver API.</p>
<p>
*** POTENTIAL INCOMPATIBILITY ***</p>
<p>
Own Id: OTP-10336 Aux Id: OTP-9892 </p>
</item>
<item>
<p>
The experimental support for packages has been removed.</p>
<p>
Own Id: OTP-10348 Aux Id: kunagi-316 [227] </p>
</item>
<item>
<p>
Wrong parameters when setting seq_trace-tokens from
within a trace-pattern could crash the VM. This is now
corrected.</p>
<p>
Own Id: OTP-10522</p>
</item>
<item>
<p>
Erlang specification 4.7.3 defines max tuple size to
65535 elements It is now enforced to no more than
16777215 elements (arity 24 bits)</p>
<p>
Previous edge cases (28 bits) were not validated and
could cause undefined behaviour.</p>
<p>
*** POTENTIAL INCOMPATIBILITY ***</p>
<p>
Own Id: OTP-10633</p>
</item>
<item>
<p>
Add insert_element/3 and delete_element/2</p>
<p>
Own Id: OTP-10643</p>
</item>
<item>
<p>
The previous default of a maximum of 32768 simultaneous
processes has been raised to 262144. This value can be
changed using the the <seealso
marker="erl#+P">+P</seealso> command line flag of
<seealso marker="erl">erl(1)</seealso>. Note that the
value passed now is considered as a hint, and that actual
value chosen in most cases will be a power of two.</p>
<p>
*** POTENTIAL INCOMPATIBILITY ***</p>
<p>
Own Id: OTP-10647 Aux Id: OTP-9892, OTP-10336 </p>
</item>
<item>
<p>
The previously (in R15) proposed scheduler wakeup
strategy is now used by default. This strategy is not as
quick to forget about previous overload as the previous
strategy.</p>
<p>
This change imply changes of the characteristics the
system. Most notable: When a small overload comes and
then disappears repeatedly, the system will for a bit
longer time be willing to wake up schedulers than before.
Timing in the system will due to this also change.</p>
<p>
The previous strategy can still be enabled by passing the
<seealso marker="erl#+sws">+sws legacy</seealso> command
line flag to <c>erl</c>.</p>
<p>
Own Id: OTP-10661 Aux Id: OTP-10033 </p>
</item>
<item>
<p>
The <seealso marker="erl#+stbt">+stbt</seealso> command line
argument of <c>erl</c> was added. This argument can be
used for trying to set scheduler bind type. Upon failure
unbound schedulers will be used.</p>
<p>
Own Id: OTP-10668</p>
</item>
<item>
<p>
Support ANSI in console</p>
<p>
Unix platforms will no longer filter control sequences to
the ttsl driver thus enabling ANSI and colors in console.
(Thanks to Pedram Nimreezi)</p>
<p>
Own Id: OTP-10678</p>
</item>
<item>
<p>Add file:allocate/3 operation</p>
<p>This operation allows pre-allocation of space for
files. It succeeds only on systems that support such
operation. (Thanks to Filipe David Manana)</p>
<p>
Own Id: OTP-10680</p>
</item>
<item>
<p>Treat <c>-Wreturn-type</c> warnings as error when
using GCC (Thanks to Tuncer Ayaz)</p>
<p>
Own Id: OTP-10683</p>
</item>
<item>
<p>
Implement ./otp_build configure --enable-silent-rules</p>
<p>
With silent rules, the output of make is less verbose and
compilation warnings are easier to spot. Silent rules are
disabled by default and can be disabled or enabled at
will by make V=0 and make V=1. (Thanks to Anthony Ramine)</p>
<p>
Own Id: OTP-10726</p>
</item>
<item>
<p>
Use share flags for all file operations on Windows.
Thanks to Filipe David Borba Manana.</p>
<p>
Own Id: OTP-10727</p>
</item>
<item>
<p>
Make/fakefop adjustments. Thanks to Tuncer Ayaz and
Sebastian Rasmussen.</p>
<p>
Own Id: OTP-10733</p>
</item>
<item>
<p>
The runtime system will now by default use 10 async
threads if thread support has been enabled when building
the runtime system.</p>
<p>
This will prevent long blocking file-operations from
blocking scheduler threads for long periods of time,
which can be harmful. Apart from file-operations, it also
effects other operations scheduled on the async thread
pool by user implemented drivers.</p>
<p>
The amount of async threads can be controlled by using
the <seealso
marker="erl#async_thread_pool_size"><c>+A</c></seealso>
command line argument of <c>erl(1)</c>. When running some
offline tools you <em>might</em> want to disable async
threads, but you are advised <em>not</em> to in the
general case. Instead, you might want to increase the
amount of async threads used.</p>
<p>
This change imply changes of the characteristics the
system compared to the previous default. The
responsiveness of the system as a whole will be improved.
Operations scheduled on the async thread pool will get an
increased latency. The throughput of these operations may
increase, or decrease depending on the type of the
operations and how they get scheduled. In the case of
file operations, the throughput very much depends on how
the Erlang application access files. Multiple concurrent
accesses to different files have the potential of an
increased throughput.</p>
<p>
Own Id: OTP-10736</p>
</item>
<item>
<p>
The default reader group limit has been increased to 64
from 8. This limit can be set using the <c>+rg</c>
command line argument of <c>erl(1)</c>.</p>
<p>
This change of default value will reduce lock contention
on ETS tables using the <c>read_concurrency</c> option at
the expense of memory consumption when the amount of
schedulers and logical processors are beween 8 and 64.
For more information, see documentation of the <c>+rg</c>
command line argument of <c>erl(1)</c>.</p>
<p>
Own Id: OTP-10737</p>
</item>
<item>
<p>
New BIF float_to_list/2 which solves a problem of
float_to_list/1 that doesn't allow specifying the number
of digits after the decimal point when formatting floats
(Thanks to Serge Aleynikov).</p>
<p>
Own Id: OTP-10752</p>
</item>
<item>
<p>
Limited support for unicode atoms in the external format
and in the internal representation of the vm. This is a
preparative feature in order to support communication
with future releases of Erlang/OTP that may create
unicode atoms.</p>
<p>
Own Id: OTP-10753</p>
</item>
<item>
<p>
Increased potential concurrency in ETS for
<c>write_concurrency</c> option. The number of internal
table locks has increased from 16 to 64. This makes it
four times less likely that two concurrent processes
writing to the same table would collide and thereby
serialized. The cost is an increased constant memory
footprint for tables using write_concurrency. The memory
consumption per inserted record is not affected. The
increased footprint can be particularly large if
<c>write_concurrency</c> is combined with
<c>read_concurrency</c>.</p>
<p>
Own Id: OTP-10787</p>
</item>
</list>
</section>
</section>
<section><title>Erts 5.9.3.1</title>
<section><title>Known Bugs and Problems</title>
<list>
<item>
<p>
Create an erl_crash.dump if no heart exists and no
ERL_CRASH_DUMP_SECONDS is set (behaviour changed).</p>
<p>
Don't create an erl_crash.dump if heart do exists and no
ERL_CRASH_DUMP_SECONDS is set (behaviour not changed).</p>
<p>
This changes the behaviour back to the R15B02 default
considering if a beam was running with no heart.</p>
<p>
Own Id: OTP-10602</p>
</item>
</list>
</section>
</section>
<section><title>Erts 5.9.3</title>
<section><title>Fixed Bugs and Malfunctions</title>
<list>
<item>
<p>
Fix linking in OpenBSD. (Thanks to Matthew Dempsky)</p>
<p>
Own Id: OTP-10395</p>
</item>
<item>
<p>
Fix bug causing fallback atomics to be used even though
healthy gcc atomics or libatomic_ops was detected.</p>
<p>
Own Id: OTP-10418</p>
</item>
<item>
<p>
Ensure 'erl_crash.dump' when asked for it. This will
change erl_crash.dump behaviour.</p>
<p>
* Not setting ERL_CRASH_DUMP_SECONDS will now terminate
beam immediately on a crash without writing a crash dump
file.</p>
<p>
* Setting ERL_CRASH_DUMP_SECONDS to 0 will also terminate
beam immediately on a crash without writing a crash dump
file, i.e. same as not setting ERL_CRASH_DUMP_SECONDS
environment variable.</p>
<p>
* Setting ERL_CRASH_DUMP_SECONDS to a negative value will
let the beam wait indefinitely on the crash dump file
being written.</p>
<p>
* Setting ERL_CRASH_DUMP_SECONDS to a positive value will
let the beam wait that many seconds on the crash dump
file being written.</p>
<p>
A positive value will set an alarm/timeout for restart
both in beam and in heart if heart is running.</p>
<p>
*** POTENTIAL INCOMPATIBILITY ***</p>
<p>
Own Id: OTP-10422 Aux Id: kunagi-250 [161] </p>
</item>
<item>
<p>
Fix bug where MSVRT100.dll was not included in the
windows installer.</p>
<p>
Own Id: OTP-10481</p>
</item>
<item>
<p>In the expression
<c><<Bin/binary,...>></c>, if <c>Bin</c> was
a bitstring with a size not a multiple of 8, either no
exception was generated or an incorrect exception was
generated. (Thanks to Adam Rutkowski for reporting this
bug.)</p>
<p>
Own Id: OTP-10524</p>
</item>
<item>
<p>
The runtime system could crash while scheduling a port
task. The port task was scheduled either due to an
external I/O event being triggered, a driver timeout
being triggered, or data being sent over a distribution
channel.</p>
<p>
Own Id: OTP-10556</p>
</item>
<item>
<p>
<c>erlang:memory(ets)</c> erroneously included the size
of each ETS-table main structure twice.</p>
<p>
Own Id: OTP-10558</p>
</item>
<item>
<p>
Fix compile error in generated file hipe_amd64_bifs.S for
Solaris.</p>
<p>
Own Id: OTP-10577</p>
</item>
<item>
<p>
A faulty spec for process_info/2 could cause false
dialyzer warnings. The spec is corrected.</p>
<p>
Own Id: OTP-10584</p>
</item>
<item>
<p>
In very rare cases, the VM could crash if a garbage
collector was called while executing an appending bit
syntax instruction. The symptom was a core when
reallocating memory in the function erts_bs_append. The
garbage collector bug is now corrected.</p>
<p>
Own Id: OTP-10590</p>
</item>
</list>
</section>
<section><title>Improvements and New Features</title>
<list>
<item>
<p>
Improve support for building and testing in embedded ppc
environments.</p>
<p>
Own Id: OTP-10265 Aux Id: kunagi-159
[daf97f67-5724-4812-a5b6-7e86990133d2-1] </p>
</item>
<item>
<p>
Due to a race condition on Windows, sometimes when
printing to standard output and then immediately
terminating erlang all data would not be printed. The
emulator now waits for all data to be printed before
exiting.</p>
<p>
Own Id: OTP-10325 Aux Id: kunagi-166
[dd72d0e2-3e76-4a51-8b56-7564e24eecae] </p>
</item>
<item>
<p>
The frequency with which sleeping schedulers are woken
due to outstanding memory deallocation jobs has been
reduced.</p>
<p>
Own Id: OTP-10476 Aux Id: OTP-10162 </p>
</item>
<item>
<p>
Clearer warnings about the dangers of misuse of <seealso
marker="erl_nif#WARNING">native functions</seealso> and
<seealso marker="erl_driver#WARNING">drivers</seealso>
have been added to the documentation.</p>
<p>
Own Id: OTP-10557</p>
</item>
</list>
</section>
</section>
<section><title>Erts 5.9.2</title>
<section><title>Fixed Bugs and Malfunctions</title>
<list>
<item>
<p>
Fix erl_prim_loader errors in handling of primary
archive. The following errors have been corrected:</p>
<list> <item> If primary archive was named "xxx", then a
file in the same directory named "xxxyyy" would be
interpreted as a file named "yyy" inside the archive.
</item> <item> erl_prim_loader did not correctly create
and normalize absolute paths for primary archive and
files inside it, so unless given with exact same path
files inside the archive would not be found. E.g. if
escript was started as /full/path/to/xxx then
"./xxx/file" would not be found since erl_prim_loader
would try to match /full/path/to/xxx with
/full/path/to/./xxx. Same problem with ../. </item>
<item> Depending on how the primary archive was built,
erl_prim_loader:list_dir/1 would sometimes return an
empty string inside the file list. This was a virtual
element representing the top directory of the archive.
This has been removed. </item>
</list>
<p>
Thanks to Tuncer Ayaz and Shunichi Shinohara for
reporting and co-authoring corrections.</p>
<p>
Own Id: OTP-10071</p>
</item>
<item>
<p>Fix: Add port-I/O statistics for active once and true
and not only active false.</p>
<p>
Own Id: OTP-10073</p>
</item>
<item>
<p>
The 64-bit windows installer did not look in the right
directories for 64-bit version of Microsoft Visual C++
2010 Redistibutable Package and hence took the wrong
decision about having to install the redistributable
package if the 32-bit version was installed but not the
64-bit and vice versa. This bug has now been fixed
Furthermore the sub-installer for the redistributable
package is now run in silent mode if the erlang installer
is.</p>
<p>
Own Id: OTP-10096</p>
</item>
<item>
<p>
epmd would fail to start automatically when starting a
distributed erlang node installed in a location with a
whitespace in the path.</p>
<p>
Own Id: OTP-10106</p>
</item>
<item>
<p>
A more or less harmless bug that sometimes caused memory
deallocations to be delayed longer than intended has been
fixed.</p>
<p>
Own Id: OTP-10116</p>
</item>
<item>
<p>
Fix bug causing emulator crash when running HiPE on ARM.
Bug has existed since R15B.</p>
<p>
Own Id: OTP-10137</p>
</item>
<item>
<p> A bug regarding spaces in C function prototypes has
been fixed. (Thanks to Richard O'Keefe.) </p>
<p>
Own Id: OTP-10138</p>
</item>
<item>
<p>
Corrected dtrace pid length in message related probes.
(Thanks to Zheng Siyao)</p>
<p>
Own Id: OTP-10142</p>
</item>
<item>
<p>
Correct formating in exit error messages</p>
<p>
Ensure displayed sizes are not negative. (Thanks to
Michael Santos)</p>
<p>
Own Id: OTP-10148</p>
</item>
<item>
<p>
fix escript/primary archive reloading</p>
<p>
If the mtime of an escript/primary archive file changes
after being added to the code path, correctly reload the
archive and update the cache. (Thanks to Tuncer Ayaz)</p>
<p>
Own Id: OTP-10151</p>
</item>
<item>
<p>
Doc fix: link from erlang:now/0 to os:timestamp/0</p>
<p>
Sometimes os:timestamp/0 is more appropriate than
erlang:now/0. The documentation for the former has a link
to the latter; this patch adds a link in the other
direction to make os:timestamp/0 more visible. Thanks to
Magnus Henoch</p>
<p>
Own Id: OTP-10180</p>
</item>
<item>
<p>
The caret in the werl window (on Windows) could appear at
the wrong place after regaining focus. This is now
corrected.</p>
<p>
Own Id: OTP-10181</p>
</item>
<item>
<p>
Fix bug that in some cases could cause corrupted binaries
in ETS tables with <c>compressed</c> option.</p>
<p>
Own Id: OTP-10182</p>
</item>
<item>
<p>
Fix use of "clever" mktime</p>
<p>
Commit 1eef765 introduced regression (conditional
_always_ evaluates to true) in which
erlang:localtime_to_universaltime/2 stopped working on
systems configured with timezone without DST (i.e. UTC)
on *BSD platforms: 1>
erlang:localtime_to_universaltime({{2012,1,1},{0,0,0}},
true). ** exception error: bad argument Thanks to Piotr
Sikora</p>
<p>
Own Id: OTP-10187</p>
</item>
<item>
<p>
Relocate bodies of DTrace probes to the statically-linked
VM.</p>
<p>
Due to various operating systems (in both the DTrace and
SystemTap worlds) not fully supporting DTrace probes (or
SystemTap-compatibility mode probes) in shared libraries,
we relocate those probes to the statically-linked virtual
machine. This could be seen as pollution of the pristine
VM by a (yet) experimental feature. However:</p>
<p>
1. This code can be eliminated completely by the C
preprocessor. 2. Leaving the probes in the dyntrace NIF
shared library simply does not work correctly on too many
platforms. *Many* thanks to Macneil Shonle at Basho for
assisting when my RSI-injured fingers gave out. (note:
Solaris 10 and FreeBSD 9.0-RELEASE can take a long time
to compile)</p>
<p>
Own Id: OTP-10189</p>
</item>
<item>
<p>
Fix bug in <c>ets:test_ms/2</c> that could cause emulator
crash when using <c>'$_'</c> in match spec.</p>
<p>
Own Id: OTP-10190</p>
</item>
<item>
<p>
Supplying a filename longer than the operating system
MAX_PATH to file:read_link/1 would cause a crash
(Segemntation fault/Critical Error) on all platforms.
This is now corrected.</p>
<p>
Own Id: OTP-10200</p>
</item>
<item>
<p>
If Perl was configured to interpret files as being
encoded in UTF-8, the build would crash in
<c>make_preload</c>. (Thanks to Aaron Harnly for noticing
this issue.)</p>
<p>
Own Id: OTP-10201</p>
</item>
<item>
<p>
Fix the erlc -MP flag</p>
<p>
Because of a copy-and-paste error in erlc.c, the -MP flag
had the same effect as -MG. As a workaround, you had to
pass +makedep_phony to enable the MP option. This patch
makes -MP work as intended.</p>
<p>
Own Id: OTP-10211</p>
</item>
<item>
<p>
Allow mixed IPv4 and IPv6 addresses to sctp_bindx</p>
<p>
Also allow mixed address families to bind, since the
first address on a multihomed sctp socket must be bound
with bind, while the rest are to be bound using
sctp_bindx. At least Linux supports adding address of
mixing families. Make inet_set_faddress function
available also when HAVE_SCTP is not defined, since we
use it to find an address for bind to be able to mix ipv4
and ipv6 addresses. Thanks to Tomas Abrahamsson</p>
<p>
Own Id: OTP-10217</p>
</item>
<item>
<p>
Fix support for leap seconds-aware timezones</p>
<p>
erlang:universaltime_to_localtime is leap seconds-aware
(since 2008), however erlang:localtime_to_universaltime
is not, which gives surprising results on systems
configured with leap seconds-aware timezones: 1>
erlang:universaltime_to_localtime({{2012,1,1},{0,0,0}}).
{{2012,1,1},{0,0,0}} 2>
erlang:localtime_to_universaltime({{2012,1,1},{0,0,0}}).
{{2012,1,1},{0,0,24}} and completely breaks
calendar:local_time_to_universal_time_dst: 3>
calendar:local_time_to_universal_time_dst({{2011,1,1},{0,0,0}}).
[] Thanks to Piotr Sikora</p>
<p>
Own Id: OTP-10227</p>
</item>
<item>
<p>
erlsrv: gracefully stop emulator on Windows shutdown</p>
<p>
Windows will send the SERVICE_CONTROL_SHUTDOWN event to
the service control handler when shutting down the
system. Instead of ignoring the event, erlsrv will now
invoke the stop action. Likewise, the Erlang emulator
(and it's po drivers) must not quit upon reception of the
CTRL_SHUTDOWN_EVENT event in th console control handler.
Thanks to Jan Kloetzke</p>
<p>
Own Id: OTP-10228</p>
</item>
<item>
<p>
Fix dtrace bug in file rename operation.</p>
<p>
Own Id: OTP-10234</p>
</item>
<item>
<p>
Fix bug in memory management of driver port data locks
(PDL). In some cases PDLs could be deallocated before
<c>ready_async</c> or <c>async_free</c> callback was
called.</p>
<p>
Own Id: OTP-10249</p>
</item>
</list>
</section>
<section><title>Improvements and New Features</title>
<list>
<item>
<p>Add port and suspend options to lock-counter
profiling. (Thanks to Rick Reed)</p>
<p>
Own Id: OTP-10051</p>
</item>
<item>
<p>
Latency when using the active_once option in gen_tcp
communication is reduced.</p>
<p>
Own Id: OTP-10055 Aux Id: sto139 </p>
</item>
<item>
<p>Remove bit8 option support from inet</p>
<p>
Own Id: OTP-10056</p>
</item>
<item>
<p>
The OS Pid of a port program is now available by calling
erlang:port_info(Port,os_pid), Thanks to Matthias Lang
for the original patch.</p>
<p>
Own Id: OTP-10057</p>
</item>
<item>
<p>
Fix openpty usage in run_erl. </p>
<p>
Reopening a slave file descriptor which was closed
earlier could lead to a misbehaving connection. This has
now been remedied.</p>
<p>
Own Id: OTP-10076</p>
</item>
<item>
<p>
Remove all code, documentation, options and diagnostic
functions which were related to the experimental hybrid
heap implementation.</p>
<p>
Own Id: OTP-10105</p>
</item>
<item>
<p>
Optimizations of memory deallocations.</p>
<p>
Own Id: OTP-10162 Aux Id: OTP-7775 </p>
</item>
<item>
<p>
Optimization of process locking.</p>
<p>
Own Id: OTP-10163</p>
</item>
<item>
<p>
Added a xcomp example file for powerpc-dso-linux-gnu</p>
<p>
Own Id: OTP-10198</p>
</item>
<item>
<p>
Detect when middle endian doubles are used by a platform
and account for it when decoding floats. (Thanks to Mike
Sperber)</p>
<p>
Own Id: OTP-10209</p>
</item>
</list>
</section>
</section>
<section><title>Erts 5.9.1.2</title>
<section><title>Fixed Bugs and Malfunctions</title>
<list>
<item>
<p>
On Linux systems using heart (erl -heart) and a
HEAR_BEAT_TIMEOUT less than default, heart could fire
even though Erlang was running fine after approx 298 to
497 days (depending on kernel config). This was due to
the behaviour of the times(2) system call. Usage of
times(2) is now replaced with clock_gettime(2) and the
CLOCK_MONOTONIC clock, resulting in a more stable
solution. The Erlang VM itself has used clock_gettime(2)
on linux since before R12B, so this only affects the
heart program.</p>
<p>
Own Id: OTP-10111 Aux Id: seq12075 </p>
</item>
</list>
</section>
</section>
<section><title>Erts 5.9.1.1</title>
<section><title>Fixed Bugs and Malfunctions</title>
<list>
<item>
<p>
If threads support for the runtime system had been
disabled at compile time (<c>--disable-threads</c> had
been passed to <c>configure</c>), and the <c>+A</c>
command line argument of <c>erl</c> was passed when
starting the runtime system, <seealso
marker="erl_driver#driver_system_info">driver_system_info()</seealso>
erroneously claimed that the runtime system had async
threads even though it had not.</p>
<p>
Due to this bug the file driver did not split tasks into
smaller chunks, but instead completed the whole task at
once, i.e., the scheduler got occupied with I/O for a
longer time than intended.</p>
<p>
Own Id: OTP-10059</p>
</item>
</list>
</section>
<section><title>Improvements and New Features</title>
<list>
<item>
<p>
A proposal for a new scheduler wakeup strategy has been
implemented. For more information see the documentation
of the <seealso marker="erl#+sws">+sws</seealso> command
line argument of <c>erl</c>.</p>
<p>
Own Id: OTP-10033 Aux Id: Seq12025 </p>
</item>
<item>
<p>
A switch for configuration of busy wait length for
scheduler threads has been added. For more information
see the documentation of the <seealso
marker="erl#+sbwt">+sbwt</seealso> command line argument
of <c>erl</c>.</p>
<p>
Own Id: OTP-10044 Aux Id: Seq11976 </p>
</item>
<item>
<p>
The extra memory barriers introduced by bug-fix OTP-9281
were unnecessarily used also on tables without the
<c>write_concurrency</c> option enabled. This could
unnecessarily degrade performance of ETS tables without
<c>write_concurrency</c> on some hardware (e.g. PowerPC)
while not effecting performance at all on other hardware
(e.g. x86/x86_64).</p>
<p>
OTP-9281 (R14B03): ETS tables using the
<c>write_concurrency</c> option could potentially get
into an internally inconsistent state.</p>
<p>
Own Id: OTP-10048 Aux Id: OTP-9281 </p>
</item>
</list>
</section>
</section>
<section><title>Erts 5.9.1</title>
<section><title>Fixed Bugs and Malfunctions</title>
<list>
<item>
<p><c>erlang:system_profile</c> errorneous profiled the
profiler process when observing runnable processes. This
has been corrected. </p>
<p>
Own Id: OTP-9849</p>
</item>
<item>
<p>Calling trace_info/2 asking for information about a
function that had native could could crash the run-time
system.</p>
<p>
Own Id: OTP-9886</p>
</item>
<item>
<p>
reduce smp locking time range in erts_garbage_collect
(thanks to Jovi Zhang)</p>
<p>
Own Id: OTP-9912</p>
</item>
<item>
<p>
Fix typo in supervisor behaviour doc (Thanks to Ricardo
Catalinas Jiménez)</p>
<p>
Own Id: OTP-9924</p>
</item>
<item>
<p>
Correct spelling of registered (Thanks to Richard
Carlsson)</p>
<p>
Own Id: OTP-9925</p>
</item>
<item>
<p>
erts: Remove unused variable (Thanks to Jovi Zhang)</p>
<p>
Own Id: OTP-9926</p>
</item>
<item>
<p>
Fix bug in ETS with <c>compressed</c> option and
insertion of term containing large integers (>2G) on
64-bit machines. Seen to cause emulator crash. (Thanks to
Diego Llarrull for excellent bug report)</p>
<p>
Own Id: OTP-9932</p>
</item>
<item>
<p>
Handle Linux OS where /sys/devices/system/node is only
readable by root. Fallback to /sys/devices/system/cpu for
topology info.</p>
<p>
Own Id: OTP-9978</p>
</item>
<item>
<p> When an escript ends now all printout to standard
output and standard error gets out on the terminal. This
bug has been corrected by changing the behaviour of
erlang:halt/0,1, which should fix the same problem for
other escript-like applications, i.e that data stored in
the output port driver buffers got lost when printing on
a TTY and exiting through erlang:halt/0,1. </p>
<p> The BIF:s erlang:halt/0,1 has gotten improved
semantics and there is a new BIF erlang:halt/2 to
accomplish something like the old semantics. See the
documentation. </p>
<p> Now erlang:halt/0 and erlang:halt/1 with an integer
argument will close all ports and allow all pending async
threads operations to finish before exiting the emulator.
Previously erlang:halt/0 and erlang:halt(0) would just
wait for pending async threads operations but not close
ports. And erlang:halt/1 with a non-zero integer argument
would not even wait for pending async threads operations.
</p>
<p> To roughly the old behaviour, to not wait for ports
and async threads operations when you exit the emulator,
you use erlang:halt/2 with an integer first argument and
an option list containing {flush,false} as the second
argument. Note that now is flushing not dependant of the
exit code, and you can not only flush async threads
operations which we deemed as a strange behaviour anyway.
</p>
<p>Also, erlang:halt/1,2 has gotten a new feature: If the
first argument is the atom 'abort' the emulator is
aborted producing a core dump, if the operating system so
allows. </p>
<p>
Own Id: OTP-9985</p>
</item>
<item>
<p>
Added check to inet driver to avoid building on operating
systems that do not yet have IPv6 compatible socket API.
(Thanks to Peer Stritzinger)</p>
<p>
Own Id: OTP-9996</p>
</item>
<item>
<p>
Fix bug when the number of CPUs actually found is lower
than the configured value. (Thanks to Benjamin
Herrenschmidt)</p>
<p>
Own Id: OTP-10004</p>
</item>
<item>
<p>
The runtime system without SMP support and without thread
support erroneously busy waited when no work was present.
This bug first appeared in <c>erts-5.9</c>.</p>
<p>
Own Id: OTP-10019</p>
</item>
<item>
<p>
Various typographical errors corrected in documentation
for common_test, driver, erl_driver and windows
installation instructions. (Thanks to Tuncer Ayaz)</p>
<p>
Own Id: OTP-10037</p>
</item>
<item>
<p>
Fix memory leak caused by race on exiting process</p>
<p>
Own Id: OTP-10041</p>
</item>
</list>
</section>
<section><title>Improvements and New Features</title>
<list>
<item>
<p>Add <c>erlang:statistics(scheduler_wall_time)</c> to
ensure correct determination of scheduler utilization.
Measuring scheduler utilization is strongly preferred
over CPU utilization, since CPU utilization gives very
poor indications of actual scheduler/vm usage.</p>
<p>
Own Id: OTP-9858</p>
</item>
<item>
<p>
ERTS internal API improvements. In some cases the amount
of atomic read operations needed have been reduced due to
this.</p>
<p>
Own Id: OTP-9922</p>
</item>
<item>
<p>
The DTrace source patch from Scott Lystig Fritchie is
integrated in the source tree. Using an emulator with
dtrace probe is still not supported for production use,
but may be a valuable debugging tool. Configure with
--with-dynamic-trace=dtrace (or
--with-dynamic-trace=systemtap) to create a build with
dtrace probes enabled. See runtime_tools for
documentation and examples.</p>
<p>
Own Id: OTP-10017</p>
</item>
</list>
</section>
<section><title>Known Bugs and Problems</title>
<list>
<item>
<p>
enif_make_copy may invalidate enif_inspect_binary.</p>
<p>
Own Id: OTP-9828</p>
</item>
</list>
</section>
</section>
<section><title>Erts 5.9.0.1</title>
<section><title>Fixed Bugs and Malfunctions</title>
<list>
<item>
<p>
A feature test for the <c>lwsync</c> instruction
performed on PowerPC hardware at runtime system startup
got into an eternal loop if the instruction was not
supported. This bug was introduced in erts-5.9/OTP-R15B.</p>
<p>
Own Id: OTP-9843</p>
</item>
<item>
<p>
I/O events could potentially be delayed for ever when
enabling kernel-poll on a non-SMP runtime system
executing on Solaris. When also combined with
async-threads the runtime system hung before completing
the boot phase. This bug was introduced in
erts-5.9/OTP-R15B.</p>
<p>
Own Id: OTP-9844</p>
</item>
</list>
</section>
</section>
<section><title>Erts 5.9</title>
<section><title>Fixed Bugs and Malfunctions</title>
<list>
<item>
<p>
Honor option <c>packet_size</c> for http packet parsing
by both TCP socket and <c>erlang:decode_packet</c>. This
gives the ability to accept HTTP headers larger than the
default setting, but also avoid DoS attacks by accepting
lines only up to whatever length you wish to allow. For
consistency, packet type <c>line</c> also honor option
<c>packet_size</c>. (Thanks to Steve Vinoski)</p>
<p>
Own Id: OTP-9389</p>
</item>
<item>
<p> A few contracts in the <c>lists</c> module have been
corrected. </p>
<p>
Own Id: OTP-9616</p>
</item>
<item>
<p>
The Unicode noncharacter code points 16#FFFE and 16#FFFE
were not allowed to be encoded or decoded using the
<c>unicode</c> module or bit syntax. That was
inconsistent with the other noncharacters 16#FDD0 to
16#FDEF that could be encoded/decoded. To resolve the
inconsistency, 16#FFFE and 16#FFFE can now be encoded and
decoded. (Thanks to Alisdair Sullivan.)</p>
<p>
*** POTENTIAL INCOMPATIBILITY ***</p>
<p>
Own Id: OTP-9624</p>
</item>
<item>
<p>
Make epp search directory of current file first when
including another file This completes a partial fix in
R11 that only worked for include_lib(). (Thanks to
Richard Carlsson)</p>
<p>
Own Id: OTP-9645</p>
</item>
<item>
<p>
Fixed memory leak in
<c>enif_inspect_io_list_as_binary</c> when applied on a
process independent environment.</p>
<p>
Own Id: OTP-9668</p>
</item>
<item>
<p>The number of beam catches allowed in code are no
longer statically defined and will grow according to its
need.</p>
<p>
Own Id: OTP-9692</p>
</item>
<item>
<p>
Add missing parenthesis in heart doc.</p>
<p>
Add missing spaces in the Reference Manual distributed
section.</p>
<p>
In the HTML version of the doc those spaces are necessary
to separate those words.</p>
<p>
Own Id: OTP-9693</p>
</item>
<item>
<p>
Fixes module erlang doc style: option description (Thanks
to Ricardo Catalinas Jiménez)</p>
<p>
Own Id: OTP-9697</p>
</item>
<item>
<p>
Specifying a scope to binary:match/3 when using multiple
searchstrings resulted in faulty return values. This is
now corrected.</p>
<p>
Own Id: OTP-9701</p>
</item>
<item>
<p>
The runtime system crashed if more than one thread tried
to exit the runtime system at the same time.</p>
<p>
Own Id: OTP-9705</p>
</item>
<item>
<p>
Fix documentation for erlang:process_flag/2</p>
<p>
For the subsection about process_flag(save_calls, N)
there's an unrelated paragraph about process priorities
which was copied from the preceeding subsection regarding
process_flag(priority, Level). (Thanks to Filipe David
Manana)</p>
<p>
Own Id: OTP-9714</p>
</item>
<item>
<p>
Calls to <c>erlang:system_flag(schedulers_online, N)</c>
and/or <c>erlang:system_flag(multi_scheduling,
block|unblock)</c> could cause internal data used by this
functionality to get into an inconsistent state. When
this happened various problems occurred. This bug was
quite hard to trigger, so hopefully no-one has been
effected by it.</p>
<p>
A spinlock used by the run-queue management sometimes got
heavily contended. This code has now been rewritten, and
the spinlock has been removed.</p>
<p>
Own Id: OTP-9727</p>
</item>
<item>
<p>
Use libdlpi to get physical address (Thanks to Trond
Norbye)</p>
<p>
Own Id: OTP-9818</p>
</item>
</list>
</section>
<section><title>Improvements and New Features</title>
<list>
<item>
<p> An option list argument can now be passed to
<c>file:read_file_info/2, file:read_link_info/2</c> and
<c>file:write_file_info/3</c> and set time type
information in the call. Valid options are <c>{time,
local}, {time, universal}</c> and <c>{time, posix}</c>.
In the case of <c>posix</c> time no conversions are made
which makes the operation a bit faster. </p>
<p>
Own Id: OTP-7687</p>
</item>
<item>
<p>A number of memory allocation optimizations have been
implemented. Most optimizations reduce contention caused
by synchronization between threads during allocation and
deallocation of memory. Most notably:</p> <list> <item>
Synchronization of memory management in scheduler
specific allocator instances has been rewritten to use
lock-free synchronization. </item> <item> Synchronization
of memory management in scheduler specific pre-allocators
has been rewritten to use lock-free synchronization.
</item> <item> The 'mseg_alloc' memory segment allocator
now use scheduler specific instances instead of one
instance. Apart from reducing contention this also
ensures that memory allocators always create memory
segments on the local NUMA node on a NUMA system. </item>
</list>
<p>
Own Id: OTP-7775</p>
</item>
<item>
<p>
The ethread atomic memory operations API used by the
runtime system has been extended and improved.</p>
<p>
The ethread library now also performs runtime tests for
presence of hardware features, such as for example SSE2
instructions, instead of requiring this to be determined
at compile time.</p>
<p>
All uses of the old deprecated atomic API in the runtime
system have been replaced with the use of the new atomic
API. In a lot of places this change imply a relaxation of
memory barriers used.</p>
<p>
Own Id: OTP-9014</p>
</item>
<item>
<p>gen_sctp:open/0-2 may now return
{error,eprotonosupport} if SCTP is not supported</p>
<p>gen_sctp:peeloff/1 has been implemented and creates a
one-to-one socket which also are supported now</p>
<p>
*** POTENTIAL INCOMPATIBILITY ***</p>
<p>
Own Id: OTP-9239</p>
</item>
<item>
<p>
Sendfile has been added to the file module's API.
sendfile/2 is used to read data from a file and send it
to a tcp socket using a zero copying mechanism if
available on that OS.</p>
<p>
Thanks to Tuncer Ayaz and Steve Vinovski for original
implementation</p>
<p>
Own Id: OTP-9240</p>
</item>
<item>
<p>
enif_get_reverse_list function added to nif API. This
function should be used to reverse small lists which are
deep within other structures making it impractical to do
the reverse in Erlang.</p>
<p>
Own Id: OTP-9392</p>
</item>
<item>
<p>
The deprecated concat_binary/1 BIF has been removed. Use
<c>list_to_binary</c> or <c>iolist_to_binary/1</c>
instead.</p>
<p>
*** POTENTIAL INCOMPATIBILITY ***</p>
<p>
Own Id: OTP-9421</p>
</item>
<item>
<p>Erlang/OTP can now be built using parallel make if you
limit the number of jobs, for instance using '<c>make
-j6</c>' or '<c>make -j10</c>'. '<c>make -j</c>' does not
work at the moment because of some missing
dependencies.</p>
<p>
Own Id: OTP-9451</p>
</item>
<item>
<p>Line number and filename information are now included
in exception backtraces as a fourth element in the MFA
tuple. The information will be pretty-printed by the
shell and used by <c>common_test</c> to provide better
indication of where a test case.</p>
<p>
*** POTENTIAL INCOMPATIBILITY ***</p>
<p>
Own Id: OTP-9468</p>
</item>
<item>
<p>All binary constants used to be handled as heap
binaries (i.e. the entire binary would be copied when
sent to another process). Binary constants larger than 64
bytes are now refc binaries (i.e. the actual data in the
binary will not be copied when sent to another
process).</p>
<p>
Own Id: OTP-9486</p>
</item>
<item>
<p>
If a float and an integer is compared, the integer is
only converted to a float if the float datatype can
contain it. Otherwise the float is converted to an
integer.</p>
<p>
*** POTENTIAL INCOMPATIBILITY ***</p>
<p>
Own Id: OTP-9497</p>
</item>
<item>
<p>
Add NIF function enif_is_number</p>
<p>
This function allows for easily determining if a term
represents or not a number (integer, float, small or
big).(Thanks to Filipe David Manana)</p>
<p>
Own Id: OTP-9629</p>
</item>
<item>
<p>
The ERTS internal system block functionality has been
replaced by new functionality for blocking the system.
The old system block functionality had contention issues
and complexity issues. The new functionality piggy-backs
on thread progress tracking functionality needed by newly
introduced lock-free synchronization in the runtime
system. When the functionality for blocking the system
isn't used, there is more or less no overhead at all.
This since the functionality for tracking thread progress
is there and needed anyway.</p>
<p>
Own Id: OTP-9631</p>
</item>
<item>
<p>
An ERTS internal, generic, many to one, lock-free queue
for communication between threads has been introduced.
The many to one scenario is very common in ERTS, so it
can be used in a lot of places in the future. Currently
it is used by scheduling of certain jobs, and the async
thread pool, but more uses are planned for the future.</p>
<p>
Drivers using the driver_async functionality are not
automatically locked to the system anymore, and can be
unloaded as any dynamically linked in driver.</p>
<p>
Scheduling of ready async jobs is now also interleaved in
between other jobs. Previously all ready async jobs were
performed at once.</p>
<p>
Own Id: OTP-9632</p>
</item>
<item>
<p>
Tuple funs (a two-element tuple with a module name and a
function) are now officially deprecated and will be
removed in R16. Use '<c>fun M:F/A</c>' instead. To make
you aware that your system uses tuple funs, the very
first time a tuple fun is applied, a warning will be sent
to the error logger.</p>
<p>
Own Id: OTP-9649</p>
</item>
<item>
<p>
Changed the internal BIF calling convention. Will make
simpler faster code and allow BIFs with an arbitrary
arity.</p>
<p>
Own Id: OTP-9662</p>
</item>
<item>
<p>
Windows native critical sections are now used internally
in the runtime system on Windows as mutex implementation.
This since they perform better under extreme contention
than our own implementation.</p>
<p>
Own Id: OTP-9671</p>
</item>
<item>
<p>
Convert some erl_nif macros into inline functions. Allow
for better compile time type checking. (Thanks to Tuncer
Ayaz)</p>
<p>
Own Id: OTP-9675</p>
</item>
<item>
<p>
The <c>+scl</c> command line flag has been added. It can
be used for disabling compaction of scheduler load. For
more information see the <c>erl(1)</c> documentation.</p>
<p>
Own Id: OTP-9695</p>
</item>
<item>
<p>The build system has been updated so that Erlang/OTP
can be built on Mac OS X Lion systems without a GCC
compiler. The INSTALL guide has been updated with
instructions on how to install a GCC compiler and build
Erlang/OTP with it, in order to get a run-time system
with better performance.</p>
<p>
Own Id: OTP-9712</p>
</item>
<item>
<p>
When loading a module, the system use to run on a single
scheduler during the entire loading process. This has
been changed to only take down the system just before
inserting the loaded code into the system tables,
resulting in a much shorter disruption if a module is
loaded in a busy system. (Suggested by Bob Ippolito.)</p>
<p>
Own Id: OTP-9720</p>
</item>
<item>
<p>
Possible to run HiPE without floating point exceptions
(FPE). Useful on platforms that lack reliable FPE. Slower
float operations compared to HiPE with FPE.</p>
<p>
Own Id: OTP-9724</p>
</item>
<item>
<p>
As of ERTS version 5.9 (OTP-R15B) the runtime system will
by default <em>not</em> bind schedulers to logical
processors.</p>
<p>
If the Erlang runtime system is the only operating system
process that binds threads to logical processors, this
improves the performance of the runtime system. However,
if other operating system processes (as for example
another Erlang runtime system) also bind threads to
logical processors, there might be a performance penalty
instead. In some cases this performance penalty might be
severe. Due to this, we change the default so that the
user must make an active decision in order to bind
schedulers.</p>
<p>
*** POTENTIAL INCOMPATIBILITY ***</p>
<p>
Own Id: OTP-9726</p>
</item>
<item>
<p>
The use of <c>erlang:system_flag(scheduler_bind_type,
_)</c> and <c>erlang:system_flag(cpu_topology, _)</c>
have been deprecated and scheduled for removal in
erts-5.10/OTP-R16. For more information see the
documentation of <c>erlang:system_flag/2</c>.</p>
<p>
*** POTENTIAL INCOMPATIBILITY ***</p>
<p>
Own Id: OTP-9749</p>
</item>
<item>
<p>
An ancient workaround for a Windows bug was removed from
the open_port code, open_port({spawn,...}...) is now
faster. Thanks to Daniel Goertzen.</p>
<p>
Own Id: OTP-9766</p>
</item>
<item>
<p>
The use of deprecated 32bit time_t on 32bit Windows is
removed.</p>
<p>
Own Id: OTP-9767</p>
</item>
<item>
<p>
The NIF <c>reload</c> mechanism is deprecated. Do not use
it as an upgrade method for live production systems. It
might be removed in future releases. It can still serve
as a development feature but a warning message will be
logged each time it is used.</p>
<p>
*** POTENTIAL INCOMPATIBILITY ***</p>
<p>
Own Id: OTP-9771</p>
</item>
<item>
<p>The driver interface has been changed to enable 64-bit
aware drivers. Most importantly the return types for
ErlDrvEntry callbacks 'call' and 'control' has ben
enlarged which require drivers to be changed or they will
cause emulator crashes. See <seealso
marker="erl_driver#rewrites_for_64_bits"> Rewrites for
64-bit driver interface </seealso> in the driver manual.
</p>
<p>Due to this driver <seealso
marker="erl_driver#version_management">version
management</seealso> is now mandatory. A driver that is
not written with version management or a driver that was
compiled with the wrong major version will be not be
loaded by the emulator.</p>
<p>
*** POTENTIAL INCOMPATIBILITY ***</p>
<p>
Own Id: OTP-9795</p>
</item>
<item>
<p>
Eliminate use of deprecated regexp module</p>
<p>
Own Id: OTP-9810</p>
</item>
</list>
</section>
</section>
<section><title>Erts 5.8.5</title>
<section><title>Fixed Bugs and Malfunctions</title>
<list>
<item>
<p>
Several bugs concerning constant binary constructions
such as <<0:4294967295>> have been corrected.
Depending on the actual size of the binary and the type
of run-time system (32-bit, halfword, 64-bit), such
expression could either crash the run-time system or make
the loader refuse loading of the module.</p>
<p>
Own Id: OTP-9284</p>
</item>
<item>
<p>
The Erlsrv utility failed to stop the erlang machine if
no StopAction was defined when the service was stopped.
This is now corrected.</p>
<p>
Own Id: OTP-9344</p>
</item>
<item>
<p>
Due to a bug in glibc the runtime system could abort
while trying to destroy a mutex. A fix for this was
introduced in R14B02. This fix did, however, not solve
the problem. The runtime system will now issue a warning
instead of aborting.</p>
<p>
Own Id: OTP-9373 Aux Id: OTP-9009 </p>
</item>
<item>
<p>
Replace atom in DRV macro in prim_file with string</p>
<p>
An experimental version of Dialyzer discovered that the
atom that replaced the DRV macro in prim_file ends up in
calls to erlang:open_port({spawn, Driver}, Portopts) as
the Driver argument. The documentation states that this
call requires a string there.</p>
<p>
This change is also consistent with the one introduced in
commit 0f03b1e9d2bef3bc830c31a369261af4c5234727 by Kostis
Sagonas.</p>
<p>
Own Id: OTP-9377</p>
</item>
<item>
<p>
Fix typos in the epmd documentation (Thanks to Holger
Weiß )</p>
<p>
Own Id: OTP-9387</p>
</item>
<item>
<p>
Fix faulty integer terms created by NIF API from 64-bit
integers on halfword emulator. (Thanks to Paolo Negri and
Paul Davis)</p>
<p>
Own Id: OTP-9394</p>
</item>
<item>
<p>
Fix <c>epmd</c> crash on vxworks caused by faulty
argument to select() system call.</p>
<p>
Own Id: OTP-9427 Aux Id: seq11855 </p>
</item>
<item>
<p>
The ets:test_ms function could in rare cases truncate the
error messages. This is now corrected.</p>
<p>
Own Id: OTP-9435</p>
</item>
<item>
<p>
Fix bug related to hibernate and HiPE (clear
F_HIBERNATE_SCHED flag)</p>
<p>
F_HIBERNATE_SCHED flag that was introduced in
b7ecdcd1ae9e11b8f75e must be cleared in hipe_mode_switch
as well. Otherwise, processes running HiPE code that
hibernate, wake up and then trap into a BIF will not be
rescheduled.(Thanks to Paul Guyot)</p>
<p>
Own Id: OTP-9452</p>
</item>
<item>
<p>
Fix bug in FreeBSD topology detection code (Thanks to
Paul Guyot)</p>
<p>
Own Id: OTP-9453</p>
</item>
<item>
<p>
Fix use of logical operator && with constant
operand instead of bitwise & (Thanks to Cristian
Greco)</p>
<p>
Own Id: OTP-9454</p>
</item>
<item>
<p>
inet: error if fd does not match socket domain</p>
<p>
If an IPv4 fd is opened as an IPv6 socket, unexpected
behaviour can occur. For example, if an IPv4 UDP socket
is opened and passed into Erlang as an IPv6 socket, the
first 3 bytes (corresponding to 1 byte representing the
protocol family, 2 bytes set to the port) are stripped
from the payload. The cause of the UDP payload truncation
happens in inet_drv.c:packet_inet_input when a call to
inet_get_address fails silently because the family is set
to PF_INET6 but the buffer len is the size of an IPv4
struct sockaddr_in.</p>
<p>
(Thanks to Andrew Tunnell-Jones for finding the bug and
the test case!)</p>
<p>
Own Id: OTP-9455</p>
</item>
<item>
<p>
erts: use a union to avoid strict aliasing issues</p>
<p>
Use a union for pointer type conversion to avoid compiler
warnings about strict-aliasing violations with gcc-4.1.
gcc >= 4.2 does not emit the warning. erts: adapt
matrix_nif to R14 erl_nif API changes (Thanks To Tuncer
Ayaz)</p>
<p>
Own Id: OTP-9487</p>
</item>
<item>
<p>
fix 64-bit issues in the garbage collection (Thanks to
Richard Carlsson)</p>
<p>
Own Id: OTP-9488</p>
</item>
<item>
<p>
epmd: fix compiler warnings</p>
<p>
Suppress compiler warnings about ignored return values.
(Thanks to Michael Santos )</p>
<p>
Own Id: OTP-9500</p>
</item>
<item>
<p>
Fix non-existing function (erlang:disconnect/1) in
distributed reference manual (Thanks to Fabian Król)</p>
<p>
Own Id: OTP-9504</p>
</item>
<item>
<p>
Document fdatasync -lrt requirement (SunOS <= 5.10)
(Thanks to Tuncer Ayaz)</p>
<p>
Own Id: OTP-9512</p>
</item>
<item>
<p>
Let epmd ignore empty ERL_EPMD_ADDRESS</p>
<p>
If the environment variable ERL_EPMD_ADDRESS is set to
the empty string, empd now behaves like it does by
default when ERL_EPMD_ADDRESS is unset. That is, in this
case, epmd now listens on all available interfaces
instead of using only the loopback interface, which
happened because epmd added the loopback address to the
(in this case empty) list of addresses specified via
ERL_EPMD_ADDRESS.</p>
<p>
Also, epmd now ignores ERL_EPMD_ADDRESS if it contains
only separator characters (comma and space).</p>
<p>
The same applies to epmd's -address option.(Thanks to
Holger Weiß)</p>
<p>
Own Id: OTP-9525</p>
</item>
<item>
<p>
Remove dead code in erl_compile (Thanks to Tuncer Ayaz)</p>
<p>
Own Id: OTP-9527</p>
</item>
<item>
<p>
Add erlang:external_size/2 BIF</p>
<p>
This BIF's second parameter is a list of options.
Currently the only allowed option is {minor_version,
Version} where version is either 0 (default) or 1.
(Thanks to Filipe David Manana )</p>
<p>
Own Id: OTP-9528</p>
</item>
<item>
<p>
Fix enif_compare on 64bits machines</p>
<p>
In 64bits machines the Sint type has a size of 8 bytes,
while on 32bits machines it has a 4 bytes size.
enif_compare was ignoring this and therefore returning
incorrect values when the result of the CMP function
(which returns a Sint value) doesn't fit in 4 bytes.
(Thanks to Filipe David Manana)</p>
<p>
Own Id: OTP-9533</p>
</item>
<item>
<p>
Implement or fix -Werror option</p>
<p>
If -Werror is enabled and there are warnings no output
file is written. Also make sure that error/warning
reporting is consistent. (Thanks to Tuncer Ayaz)</p>
<p>
Own Id: OTP-9536</p>
</item>
<item>
<p>In some rare cases we did not have a run queue when
scheduling misc ops. This is now fixed.</p>
<p>
Own Id: OTP-9537</p>
</item>
<item>
<p>Remove misc. compiler warnings</p>
<p>
Own Id: OTP-9542</p>
</item>
<item>
<p>
Two bugs in gen_sctp has been corrected: getopts/setopts
hence also send could only be called from socket owner,
and options 'linger', 'rcvbuf' and 'sndbuf' was read from
wrong protocol layer hence read wrong values by getopts.</p>
<p>
Own Id: OTP-9544</p>
</item>
<item>
<p>
Erlang/OTP can now be built on MacOS X Lion.</p>
<p>
Own Id: OTP-9547</p>
</item>
<item>
<p> XML files have been corrected. </p>
<p>
Own Id: OTP-9550 Aux Id: OTP-9541 </p>
</item>
<item>
<p>
Fix potential errors inspired by running cppcheck(1)
(Thanks to Christian von Roques)</p>
<p>
Own Id: OTP-9557</p>
</item>
<item>
<p>When auxiliary work was enqueued on a scheduler, the
wakeup of the scheduler in order to handle this work
could be lost. Wakeups in order to handle ordinary work
were not effected by this bug. The bug only effected
runtime systems with SMP support as follows:</p> <list>
<item>Deallocation of some ETS data structures could be
delayed.</item> <item>On Linux systems not using the NPTL
thread library (typically ancient systems with kernel
versions prior to 2.6) and Windows systems, the <c>{Port,
{exit_status, Status}}</c> message from a terminating
port program could be delayed. That is, it only effected
port programs which had been started by passing
<c>exit_status</c> as an option to
<c>open_port/2</c>.</item> </list>
<p>
Own Id: OTP-9567</p>
</item>
<item>
<p>
Handle rare race in the crypto key server functionality</p>
<p>
Own Id: OTP-9586</p>
</item>
</list>
</section>
<section><title>Improvements and New Features</title>
<list>
<item>
<p> Types and specifications have been added. </p>
<p>
Own Id: OTP-9356</p>
</item>
<item>
<p>
New allocator strategy "address order first fit". May
ease the emptying of memory carriers and thereby real
release of memory back to the OS.</p>
<p>
Own Id: OTP-9424</p>
</item>
<item>
<p>
The new <c>erlang:check_old_code/1</c> BIF checks whether
a module has old code.</p>
<p>
Own Id: OTP-9495</p>
</item>
<item>
<p> Update documentation and specifications of some of
the zlib functions. </p>
<p>
Own Id: OTP-9506</p>
</item>
<item>
<p>
Detect the available CPUs on IRIX</p>
<p>
Add support for querying the number of configured and
online processors on SGI systems running IRIX.(Thanks to
Holger Weiß)</p>
<p>
Own Id: OTP-9531</p>
</item>
</list>
</section>
</section>
<section><title>Erts 5.8.4</title>
<section><title>Fixed Bugs and Malfunctions</title>
<list>
<item>
<p>Fix binary and iolist overflow problems. Typically
problems arose in length calculation where the result
would exceed (1 bsl 32 - 1).</p>
<p>
Own Id: OTP-9118</p>
</item>
<item>
<p>
Using the old erlang shell (i.e. erl instead on werl) on
windows and doing several init:restart's would eventually
hang the VM. That is no longer the case.</p>
<p>
Own Id: OTP-9139</p>
</item>
<item>
<p>
Removed recursive C code when printing Erlang terms to
buffers, avoiding stack overflows that could cause VM to
crash.</p>
<p>
Own Id: OTP-9140</p>
</item>
<item>
<p>
The send_timeout option in gen_tcp did not work properly
in active mode or with {active,once} options. This is now
corrected.</p>
<p>
Own Id: OTP-9145</p>
</item>
<item>
<p>
Fixed various typos across the documentation (Thanks to
Tuncer Ayaz)</p>
<p>
Own Id: OTP-9154</p>
</item>
<item>
<p>Remove duplicate stack entries which could occur after
calling certain BIFs.</p>
<p>
Own Id: OTP-9163</p>
</item>
<item>
<p>
A race when starting two nodes simultaneously using
run_erl has been removed.</p>
<p>
Own Id: OTP-9164</p>
</item>
<item>
<p>
Add documentation on .erlang processing back again
(Thanks to Gabor Liptak)</p>
<p>
Own Id: OTP-9189</p>
</item>
<item>
<p>
Remove gratuitous paren in driver_entry(Thanks to Tuncer
Ayaz)</p>
<p>
Own Id: OTP-9192</p>
</item>
<item>
<p>
Fix some wrong pointer dereferences (Thanks to Cristian
Greco)</p>
<p>
Own Id: OTP-9194</p>
</item>
<item>
<p>
erts: Remove unused variables (Thanks to Tuncer Ayaz)</p>
<p>
Own Id: OTP-9205</p>
</item>
<item>
<p>
The documentation for <c>init:get_args/0</c> has been
removed. <c>init:get_args/0</c> itself was deprecated in
R9C and removed in R12B. (Thanks to Eric Pailleau.)</p>
<p>
Own Id: OTP-9209</p>
</item>
<item>
<p>
Allow user to specify the IP address epmd binds to</p>
<p>
The IP address(es) epmd binds to can now be specified by
the user, either via epmd's new "-address" option or (if
that's not used) by setting the environment variable
ERL_EPMD_ADDRESS. Multiple addresses may be specified
using a comma-separated list. If the loopback address is
not in this list, it will be added implicitly, so that
the daemon can be queried by an interactive epmd
process.(Thanks to Holger Weiß)</p>
<p>
Own Id: OTP-9213</p>
</item>
<item>
<p>
epmd: include host address in local access check</p>
<p>
In FreeBSD jails, the source and destination address of
connections to localhost are changed to be the IP address
of the jail. Consider connections from the host's IP
address to itself (e.g., the source and destination
address match) to be local for the access control checks.
(Thanks to Michal Santos and Tom at diogunix.com)</p>
<p>
Own Id: OTP-9214</p>
</item>
<item>
<p>
Fix list returned by net_kernel:epmd_module</p>
<p>
Function epmd_module of net_kernel returns a list instead
of an atom, when the epmd_module-flag is used. (Thanks to
Markus Knofe)</p>
<p>
Own Id: OTP-9215</p>
</item>
<item>
<p>
Fix epmd's dbg_perror() output</p>
<p>
The dbg_perror() function now hands the current errno
value over to dbg_gen_printf(). This fixes the problem
that errno had been reset to zero by the time it was used
(to print the corresponding error message) in the
dbg_gen_printf() function. (Thanks to Holger Weiß)</p>
<p>
Own Id: OTP-9223</p>
</item>
<item>
<p>
heart: remove garbage appended to heart command</p>
<p>
heart:get_cmd/0 is documented to return an empty string
if the command is cleared. get_cmd/0 returns 2 extra
bytes: 1 byte for the trailing null, 1 byte from the op
(the op is an unsigned char and 2 bytes are allocated for
it in the returned buffer). (Thanks to Michael Santos)</p>
<p>
Own Id: OTP-9224</p>
</item>
<item>
<p>
file: fix hang reading compressed files</p>
<p>
The gzio driver goes into an infinite loop when reading
past the end of a compressed file. Reported-By: Alex
Morarash (Thanks to Michael Santos)</p>
<p>
Own Id: OTP-9245</p>
</item>
<item>
<p>Eliminate alias warnings from gcc 4.5.2</p>
<p>
Own Id: OTP-9250</p>
</item>
<item>
<p>
Unsigned integer may overflow in error message (Thanks to
Michael Santos)</p>
<p>
Own Id: OTP-9251</p>
</item>
<item>
<p>
Driver names should be strings, not atoms</p>
<p>
Own Id: OTP-9253</p>
</item>
<item>
<p>
driver_entry: Remove gratuitous paren and fix typo
(Thanks to Tuncer Ayaz)</p>
<p>
Own Id: OTP-9254</p>
</item>
<item>
<p>
Fix format specifiers in erl_exit messages</p>
<p>
Fix an error message by using an unsigned integer
specifier as seen in a tweet by @metabrew: #erlang VM
crashed with "no next heap size found: -2090496108,
offset 0", suddenly allocated all available RAM</p>
<p>
Also correct mis-typed string formats in bif.c.(Thanks to
Michael Santos)</p>
<p>
Own Id: OTP-9262</p>
</item>
<item>
<p>
net_drv: remove unused tcp request id inet_drv: remove
gratuitous assignment (Thanks to Tuncer Ayaz)</p>
<p>
Own Id: OTP-9263</p>
</item>
<item>
<p>
Teach run_erl RUN_ERL_DISABLE_FLOWCNTRL for disabling
flow control</p>
<p>
Flow control can cause unwanted behaviour of the beam
process, if accidentally hit Ctrl-S (instead of Ctrl-D to
detach) the entire beam may be blocked.</p>
<p>
Fix this problem by making it possible to turn off flow
control by setting the environment variable
RUN_ERL_DISABLE_FLOWCNTRL. (Thanks to Jonas Faklkevik)</p>
<p>
Own Id: OTP-9270</p>
</item>
<item>
<p>The following bugs due to missing memory barriers have
been fixed:</p> <list> <item><p> ETS tables using the
<c>write_concurrency</c> option could potentially get
into an internally inconsistent state. </p></item>
<item><p> ETS tables using the <c>ordered_set</c> option
could potentially get into an internally inconsistent
state. </p></item> <item><p> A number of memory barriers
have been added when building with the
<c>libatomic_ops</c> API (i.e. when passing
<c>--with-libatomic_ops=PATH</c> to <c>configure</c>) and
the tilera atomics API (i.e. when building for the tilera
chip). Note that these bugs were due to erroneous usage
of the APIs, and not in the implementations of the APIs.
When using these APIs the following bugs where
present:</p> <list> <item><p> The BIF
<c>erlang:ports/0</c> could return an erroneous result.
</p></item> <item><p> A thread blocking other threads
during code loading, or setup of tracing could
potentially read invalid data. </p></item> <item><p>
Fixation of ETS tables could potentially get into an
internally inconsistent state. </p></item> </list>
</item> </list>
<p>
Own Id: OTP-9281</p>
</item>
<item>
<p>
Fix halfword bug for ETS ordered_set when doing
<c>select/match</c> with partly bound key.</p>
<p>
Own Id: OTP-9292</p>
</item>
<item>
<p>
Fix bug in <c>code:is_module_native</c> that caused crash
for deleted modules.</p>
<p>
Own Id: OTP-9298</p>
</item>
<item>
<p>
Calling <c>driver_async_cancel()</c> could cause a
scheduler thread to enter an eternal loop doing no useful
work. (Thanks to Anders Ramsell)</p>
<p>
Own Id: OTP-9302</p>
</item>
</list>
</section>
<section><title>Improvements and New Features</title>
<list>
<item>
<p>
New <c>enif_is_exception function</c> to allow NIFs to
determine whether an <c>ERL_NIF_TERM</c> represents an
exception. (Thanks to Steve Vinoski)</p>
<p>
Own Id: OTP-9150</p>
</item>
<item>
<p>
A process being garbage collected by another process
could be scheduled on another scheduler. This prevented
this scheduler from doing any useful work until the
garbage collection was done. This either occurred due to
a explicit call to the <c>garbage_collect/1</c> BIF, or
due to a garbage collection part of code loading. A
process being garbage collected like this will now not be
scheduled until the garbage collection has completed.</p>
<p>
Own Id: OTP-9211</p>
</item>
<item>
<p>
Remove unnecessary validation copy in
prim_file:drv_command/3 (Thanks to Tony Rogvall)</p>
<p>
Own Id: OTP-9276</p>
</item>
<item>
<p>
Symbolic link handling on windows have been slightly
updated to map error conditions more consequently and
correctly read directory links created outside of the
Erlang environment.</p>
<p>
Own Id: OTP-9279</p>
</item>
<item>
<p>
Due to standard library DLL mismatches between versions
of OpenSSL and Erlang/OTP, OpenSSL is now linked
statically to the crypto driver on Windows. This fixes
problems starting crypto when running Erlang as a service
on all Windows versions.</p>
<p>
Own Id: OTP-9280</p>
</item>
<item>
<p>Halfword emulator memory handling improvements:</p>
<list> <item><p>Much more of internal memory structures
have been made able to use "high" memory and are no
longer restricted to the 4Gb limit that still applies for
all process heap data.</p> </item> <item><p>Fixed faulty
values from <c>erlang:memory()</c> caused by 32-bit
counter overflow.</p> </item> <item><p>New counter
<c>low</c> in <c>erlang:memory()</c> that sums up all
memory restricted by 4Gb limit.</p> </item> </list>
<p>
Own Id: OTP-9291 Aux Id: seq11841 </p>
</item>
<item>
<p>
The value set in the undocumented and unsupported
ERL_version_FLAGS (e.g. ERL_R14B03_FLAGS) environment
variable can now be overridden by the command line
(similar to ERL_AFLAGS).</p>
<p>
Own Id: OTP-9297</p>
</item>
</list>
</section>
</section>
<section><title>Erts 5.8.3.2</title>
<section><title>Known Bugs and Problems</title>
<list>
<item>
<p>
Fix halfword emulator bug in <c>ets:select_delete</c> for
<c>ordered_set</c> that caused emulator to crash.</p>
<p>
Own Id: OTP-9258 Aux Id: seq11836 </p>
</item>
</list>
</section>
</section>
<section><title>Erts 5.8.3.1</title>
<section><title>Fixed Bugs and Malfunctions</title>
<list>
<item>
<p>
Ets table type ordered_set could order large integer keys
wrongly on pure 64bit platforms. This is now corrected.</p>
<p>
Own Id: OTP-9181</p>
</item>
<item>
<p>
The status of a process was unnecessarily set to waiting
before a process was enqueued in a run queue. This bug
was harmless up until OTP-R14B01. In OTP-R14B02
<c>erlang:hibernate/3</c> was fixed (OTP-9125). After the
introduction of OTP-9125, the previously harmless process
status bug sometimes caused erroneous badarg exceptions
from <c>process_info()</c>.</p>
<p>
OTP-9125 also introduced a thread unsafe access to the
status field of a process which now also have been fixed.</p>
<p>
*** INCOMPATIBILITY with noxs ***</p>
<p>
Own Id: OTP-9197</p>
</item>
</list>
</section>
</section>
<section><title>Erts 5.8.3</title>
<section><title>Fixed Bugs and Malfunctions</title>
<list>
<item>
<p>
The scroll wheel now scrolls the werl window on Windows.</p>
<p>
Own Id: OTP-8985</p>
</item>
<item>
<p>
Some malformed distribution messages could cause VM to
crash, this is now corrected.</p>
<p>
Own Id: OTP-8993</p>
</item>
<item>
<p>
The OS function getifaddrs() can return NULL in some
address fields for e.g PPP and tunnel devices which
caused the emulator to segfault. This bug has now been
corrected.</p>
<p>
Own Id: OTP-8996</p>
</item>
<item>
<p>
The expression <<A:0>> would always produce
an empty binary, even if <c>A</c> was not an integer.
Corrected to cause a <c>badarg</c> exception if the type
of <c>A</c> is invalid. (Thanks to Zvi.)</p>
<p>
Own Id: OTP-8997</p>
</item>
<item>
<p>
A bug that potentially could cause an emulator crash when
deleting an ETS-table has been fixed. A resource leak
when hitting the maximum amount of ETS-tables allowed has
also been fixed.</p>
<p>
Own Id: OTP-8999</p>
</item>
<item>
<p>
A bug in the <c>exit/2</c> BIF could potentially cause an
emulator crash.</p>
<p>
Own Id: OTP-9005</p>
</item>
<item>
<p>
Due to a bug in glibc the runtime system could abort
while trying to destroy a mutex. The runtime system will
now issue a warning instead of aborting.</p>
<p>
Own Id: OTP-9009</p>
</item>
<item>
<p>
A bug in epmd could create strange behaviour when
listen() calls failed. This is now corrected thanks to
Steve Vinoski.</p>
<p>
Own Id: OTP-9024</p>
</item>
<item>
<p>When setting file_info the win32_driver will now
correctly set access and modified time. Previously these
entities were swapped.</p>
<p>
Own Id: OTP-9046</p>
</item>
<item>
<p>
Setting scheduler bind type to <c>unbound</c> failed if
binding of schedulers wasn't supported, or if CPU
topology wasn't present. This even though the
documentation stated that it is possible to set the bind
type to <c>unbound</c>.</p>
<p>
Own Id: OTP-9056 Aux Id: Seq11779 </p>
</item>
<item>
<p>Two problems were fixed in crash dump: The time left
for timers are now shown as unsigned integers and the
contents of ordered_set ETS tables is no longer
included.</p>
<p>
Own Id: OTP-9057</p>
</item>
<item>
<p>
The VM could fail to set IP_TOS and SO_PRIORITY in
certain situations, either because sockets were supplied
as open file descriptors, or because SO_PRIORITY by
default was set higher than the user can explicitly set
it to. Those situations are now handled.</p>
<p>
Own Id: OTP-9069</p>
</item>
<item>
<p>
Wx on MacOS X generated complains on stderr about certain
cocoa functions not beeing called from the "Main thread".
This is now corrected.</p>
<p>
Own Id: OTP-9081</p>
</item>
<item>
<p>
Fix a couple typos in driver_entry(3) (thanks to Tuncer
Ayaz).</p>
<p>
Own Id: OTP-9085</p>
</item>
<item>
<p>
Mention that "-detached" implies "-noinput"</p>
<p>
Clarify that specifying "-noinput" is unnecessary if the
"-detached" flag is given. (thanks to Holger Weiß)</p>
<p>
Own Id: OTP-9086</p>
</item>
<item>
<p>
A potential problem (found by code inspection) when
calling a fun whose code was not loaded has been fixed.</p>
<p>
Own Id: OTP-9095</p>
</item>
<item>
<p>
The emulator could get into a state where it didn't check
for I/O.</p>
<p>
Own Id: OTP-9105 Aux Id: Seq11798 </p>
</item>
<item>
<p>
Attempting to create binaries exceeding 2Gb (using for
example <c>term_to_binary/1</c>) would crash the emulator
with an attempt to allocate huge amounts of memory.
(Thanks to Jon Meredith.)</p>
<p>
Own Id: OTP-9117</p>
</item>
<item>
<p>
Fix erlang:hibernate/3 on HiPE enabled emulator (Thanks
to Paul Guyot)</p>
<p>
Own Id: OTP-9125</p>
</item>
</list>
</section>
<section><title>Improvements and New Features</title>
<list>
<item>
<p>From this release, the previously experimental
halfword emulator is now official. It can be enabled by
giving the <c>--enable-halfword-emulator</c> option to
the <c>configure</c> script.</p>
<p>The halfword emulator is a 64-bit application, but
uses halfwords (32-bit words) for all data in Erlang
processes, therefore using less memory and being faster
than the standard 64-bit emulator. The total size of all
BEAM code and all process data for all processes is
limited to 4Gb, but ETS tables and off-heap binaries are
only limited by the amount of available memory.</p>
<p>
Own Id: OTP-8941</p>
</item>
<item>
<p>
32-bit atomic memory operations have been introduced
internally in the run time system, and are now used where
appropriate. There were previously only atomic memory
operations of word size available. The 32-bit atomic
memory operations slightly reduce memory consumption, and
slightly improve performance on 64-bit runtime systems.</p>
<p>
Own Id: OTP-8974</p>
</item>
<item>
<p>
Performance enhancements for looking up timer-entries and
removing timers from the wheel.</p>
<p>
Own Id: OTP-8990</p>
</item>
<item>
<p>
Write accesses to ETS tables have been optimized by
reducing the amount of atomic memory operations needed
during a write access.</p>
<p>
Own Id: OTP-9000</p>
</item>
<item>
<p>
Strange C coding in the VM made the -D_FORTIFY_SOURCE
option to gcc-4.5 react badly. The code is now cleaned up
so that it's accepted by gcc-4.5.</p>
<p>
Own Id: OTP-9025</p>
</item>
<item>
<p>
The memory footprint for loaded code has been somewhat
reduced (especially in the 64-bit BEAM machine).</p>
<p>
Own Id: OTP-9030</p>
</item>
<item>
<p>
The maximum number of allowed arguments for an Erlang
function has been lowered from 256 to 255, so that the
number of arguments can now fit in a byte.</p>
<p>
Own Id: OTP-9049</p>
</item>
<item>
<p>
Dependency generation for Makefiles has been added to the
compiler and erlc. See the manual pages for
<c>compile</c> and <c>erlc</c>. (Thanks to Jean-Sebastien
Pedron.)</p>
<p>
Own Id: OTP-9065</p>
</item>
</list>
</section>
</section>
<section><title>Erts 5.8.2</title>
<section><title>Fixed Bugs and Malfunctions</title>
<list>
<item>
<p> Fix format_man_pages so it handles all man sections
and remove warnings/errors in various man pages. </p>
<p>
Own Id: OTP-8600</p>
</item>
<item>
<p>
The <c>configure</c> command line argument <seealso
marker="doc/installation_guide:INSTALL#Advanced-configuration-and-build-of-ErlangOTP">--enable-ethread-pre-pentium4-compatibility</seealso>
had no effect. This option is now also automatically
enabled if required on the build machine.</p>
<p>
Own Id: OTP-8847</p>
</item>
<item>
<p>
Windows 2003 and Windows XP pre SP3 would sometimes not
start the Erlang R14B VM at all due to a bug in the cpu
topology detection. The bug affects Windows only, no
other platform is even remotely affected. The bug is now
corrected.</p>
<p>
Own Id: OTP-8876</p>
</item>
<item>
<p>
The HiPE run-time in the 64-bit emulator could do a
64-bit write to a 32-bit struct field. It happened to be
harmless on Intel/AMD processors. Corrected. (Thanks to
Mikael Pettersson.)</p>
<p>
Own Id: OTP-8877</p>
</item>
<item>
<p>
A bug in <seealso
marker="erl_driver#erl_drv_tsd_get">erl_drv_tsd_get()</seealso>
and <seealso
marker="erl_nif#enif_tsd_get">enif_tsd_get()</seealso>
could cause an emulator crash. These functions are
currently not used in OTP. That is, the crash only occur
on systems with user implemented NIF libraries, or
drivers that use one of these functions.</p>
<p>
Own Id: OTP-8889</p>
</item>
<item>
<p>
Calling <c>erlang:system_info({cpu_topology,
CpuTopologyType})</c> with another <c>CpuTopologyType</c>
element than one of the documented atoms <c>defined</c>,
<c>detected</c>, or <c>used</c> caused an emulator crash.
(Thanks to Paul Guyot)</p>
<p>
Own Id: OTP-8914</p>
</item>
<item>
<p>
The ERTS internal rwlock implementation could get into an
inconsistent state. This bug was very seldom triggered,
but could be during heavy contention. The bug was
introduced in R14B (erts-5.8.1).</p>
<p>
The bug was most likely to be triggered when using the
<c>read_concurrency</c> option on an ETS table that was
frequently accessed from multiple processes doing lots of
writes and reads. That is, in a situation where you
typically don't want to use the <c>read_concurrency</c>
option in the first place.</p>
<p>
Own Id: OTP-8925 Aux Id: OTP-8544 </p>
</item>
<item>
<p>
Tracing to port could cause an emulator crash when
unloading the trace driver.</p>
<p>
Own Id: OTP-8932</p>
</item>
<item>
<p>
Removed use of CancelIoEx on Windows that had been shown
to cause problems with some drivers.</p>
<p>
Own Id: OTP-8937</p>
</item>
<item>
<p>
The fallback implementation used when no native atomic
implementation was found did not compile. (Thanks to
Patrick Baggett, and Tuncer Ayaz)</p>
<p>
Own Id: OTP-8944</p>
</item>
<item>
<p>
Some integer values used during load balancing could
under rare circumstances wrap causing a load unbalance
between schedulers.</p>
<p>
Own Id: OTP-8950</p>
</item>
<item>
<p>
The windows VM now correctly handles appending to large
files (> 4GB).</p>
<p>
Own Id: OTP-8958</p>
</item>
<item>
<p>
Name resolving of IPv6 addresses has been implemented for
Windows versions that support it. The use of ancient
resolver flags (AI_V4MAPPED | AI_ADDRCONFIG) to the
getaddrinfo() function has been removed since e.g FreeBSD
regard mapped IPv4 addresses to be a security problem and
the semantics of the address configured flag is
uncertain.</p>
<p>
Own Id: OTP-8969</p>
</item>
</list>
</section>
<section><title>Improvements and New Features</title>
<list>
<item>
<p>
The help texts produced by the <c>configure</c> scripts
in the top directory and in the erts directory have been
aligned and cleaned up.</p>
<p>
Own Id: OTP-8859</p>
</item>
<item>
<p>
When the runtime system had fewer schedulers than logical
processors, the system could get an unnecessarily large
amount reader groups.</p>
<p>
Own Id: OTP-8861</p>
</item>
<item>
<p>
<c>run_rel</c> has been updated to support Solaris's
/dev/ptmx device and to load the necessary STREAMS
modules so that <c>to_erl</c> can provide terminal echo
of keyboard input. (Thanks to Ryan Tilder.)</p>
<p>
Own Id: OTP-8878</p>
</item>
<item>
<p>
The Erlang VM now supports Unicode filenames. The feature
is turned on by default on systems where Unicode
filenames are mandatory (Windows and MacOSX), but can be
enabled on other systems with the '+fnu' emulator option.
Enabling the Unicode filename feature on systems where it
is not default is however considered experimental and not
to be used for production. Together with the Unicode file
name support, the concept of "raw filenames" is
introduced, which means filenames provided without
implicit unicode encoding translation. Raw filenames are
provided as binaries, not lists. For further information,
see stdlib users guide and the chapter about using
Unicode in Erlang. Also see the file module manual page.</p>
<p>
*** POTENTIAL INCOMPATIBILITY ***</p>
<p>
Own Id: OTP-8887</p>
</item>
<item>
<p>Buffer overflows have been prevented in <c>erlc</c>,
<c>dialyzer</c>, <c>typer</c>, <c>run_test</c>,
<c>heart</c>, <c>escript</c>, and <c>erlexec</c>.</p>
(Thanks to Michael Santos.)
<p>
Own Id: OTP-8892</p>
</item>
<item>
<p>
The runtime system is now less eager to suspend processes
sending messages over the distribution. The default value
of the distribution buffer busy limit has also been
increased from 128 KB to 1 MB. This in order to improve
throughput.</p>
<p>
Own Id: OTP-8901</p>
</item>
<item>
<p>
The distribution buffer busy limit can now be configured
at system startup. For more information see the
documentation of the <c>erl</c> <seealso
marker="erl#+zdbbl">+zdbbl</seealso> command line flag.
(Thanks to Scott Lystig Fritchie)</p>
<p>
Own Id: OTP-8912</p>
</item>
<item>
<p>
The inet driver internal buffer stack implementation has
been rewritten in order to reduce lock contention.</p>
<p>
Own Id: OTP-8916</p>
</item>
<item>
<p>
New ETS option <c>compressed</c>, to enable a more
compact storage format at the expence of heavier table
operations. For test and evaluation, <c>erl +ec</c> can
be used to force compression on all ETS tables.</p>
<p>
Own Id: OTP-8922 Aux Id: seq11658 </p>
</item>
<item>
<p>
There is now a new function inet:getifaddrs/0 modeled
after C library function getifaddrs() on BSD and LInux
that reports existing interfaces and their addresses on
the host. This replaces the undocumented and unsupported
inet:getiflist/0 and inet:ifget/2.</p>
<p>
Own Id: OTP-8926</p>
</item>
<item>
<p>
Support for detection of CPU topology and binding of
schedulers on FreeBSD 8 have been added. (Thanks to Paul
Guyot)</p>
<p>
Own Id: OTP-8939</p>
</item>
<item>
<p>
Several bugs related to hibernate/3 and HiPE have been
corrected. (Thanks to Paul Guyot.)</p>
<p>
Own Id: OTP-8952</p>
</item>
<item>
<p>
Support for soft and hard links on Windows versions and
filesystems that support them is added.</p>
<p>
Own Id: OTP-8955</p>
</item>
<item>
<p>
The win32 virtual machine is now linked large address
aware. his allows the Erlang VM to use up to 3 gigs of
address space on Windows instead of the default of 2
gigs.</p>
<p>
Own Id: OTP-8956</p>
</item>
</list>
</section>
</section>
<section><title>Erts 5.8.1.2</title>
<section><title>Fixed Bugs and Malfunctions</title>
<list>
<item>
<p> Fix that the documentation top index generator can
handle an Ericsson internal application group. </p>
<p>
Own Id: OTP-8875</p>
</item>
<item>
<p>In embedded mode, on_load handlers that called
<c>code:priv_dir/1</c> or other functions in <c>code</c>
would hang the system. Since the <c>crypto</c>
application now contains an on_loader handler that calls
<c>code:priv_dir/1</c>, including the <c>crypto</c>
application in the boot file would prevent the system
from starting.</p>
<p>Also extended the <c>-init_debug</c> option to print
information about on_load handlers being run to
facilitate debugging.</p>
<p>
Own Id: OTP-8902 Aux Id: seq11703 </p>
</item>
</list>
</section>
</section>
<section><title>Erts 5.8.1.1</title>
<section><title>Fixed Bugs and Malfunctions</title>
<list>
<item>
<p>
Windows 2003 and Windows XP pre SP3 would sometimes not
start the Erlang R14B VM at all due to a bug in the cpu
topology detection. The bug affects Windows only, no
other platform is even remotely affected. The bug is now
corrected.</p>
<p>
Own Id: OTP-8876</p>
</item>
</list>
</section>
</section>
<section><title>Erts 5.8.1</title>
<section><title>Fixed Bugs and Malfunctions</title>
<list>
<item>
<p>
Very small floating point numbers generated errors when
converting from list to float in some versions of the VM,
this is now corrected so that i.e.
list_to_float("1.0e-324"). returns 0.0 in all versions of
Erlang.</p>
<p>
Own Id: OTP-7178</p>
</item>
<item>
<p>
Windows Vista and Windows 7 file system virtualization,
which makes "old style" windows programs execute in a
file system sandbox, was previously unintentionally
turned on for the Erlang VM. This is now corrected so
that i.e. writes to C:\Program Files\. without
administrator privileges will fail.</p>
<p>
Own Id: OTP-7405</p>
</item>
<item>
<p>
Fix faulty 64-bit integer term output from drivers. Large
64-bits integers did not generate correct bignums and
could even cause emulator crash. Only affects drivers
using ERL_DRV_INT64 or ERL_DRV_UINT64, introduced in
R13B03.</p>
<p>
Own Id: OTP-8716</p>
</item>
<item>
<p>
Fixed: inet:setopts(S, [{linger,{true,2}}]) returned
{error,einval} for SCTP sockets. The inet_drv had a bug
when checking the option size.</p>
<p>
Own Id: OTP-8726 Aux Id: seq11617 </p>
</item>
<item>
<p>Fix libm linking with --as-needed flag</p>
<p>
When building with "--as-needed" linker flags on Linux
the build will fail. This has now been fixed.</p>
<p>
(Thanks to Christian Faulhammer)</p>
<p>
Own Id: OTP-8728</p>
</item>
<item>
<p>
gen_udp:connect/3 was broken for SCTP enabled builds. It
did not detect remote end errors as it should.</p>
<p>
Own Id: OTP-8729</p>
</item>
<item>
<p>Reduce the risk of integer wrapping in bin vheap size
counting.</p> <p>The vheap size series will now use the
golden ratio instead of doubling and fibonacci
sequences.</p>
<p>
Own Id: OTP-8730</p>
</item>
<item>
<p>
ETS ordered_set containing <c>[]</c> as key could cause
strange thing to happen, like an infinite hanging
<c>ets:select</c>.</p>
<p>
Own Id: OTP-8732</p>
</item>
<item>
<p>reference() has been substituted for ref() in the
documentation.</p>
<p>
Own Id: OTP-8733</p>
</item>
<item>
<p>
When a native compiled module called a not loaded
non-native compiled module that had an on_load function,
the export entries were trashed after code loading so on
the next call from the native compiled module to the
non-native compiled the emulator crashed. This bug has
now been fixed.</p>
<p>
Own Id: OTP-8736</p>
</item>
<item>
<p>
HiPE-enabled Erlang VMs running on BSD systems sometimes
generated messages like "Yikes! erts_alloc() returned
misaligned address 0x8016a512c". Fixed. (Thanks to Mikael
Pettersson.)</p>
<p>
Own Id: OTP-8769</p>
</item>
<item>
<p>
A race condition in <c>erts_poll()</c> could cause delay
of poll for I/O.</p>
<p>
Own Id: OTP-8773</p>
</item>
<item>
<p>
Removed some potential vulnerabilities from the Erlang
Port Mapper Daemon (epmd) and straightened up access
control. Also removed hazardous interfaces allowing
anyone on a machine to forcefully unregister other nodes.
This means that the ei_unregister/erl_unregister
interfaces in erl_interface is rendered not only error
prone and mystifying as before, but totally ineffective.
The old behaviour of unchecked node unregistering can be
restored if needed, see epmd documentation for details.</p>
<p>
*** POTENTIAL INCOMPATIBILITY ***</p>
<p>
Own Id: OTP-8780</p>
</item>
<item>
<p> Building in a source tree without prebuilt platform
independent build results failed on the SSL examples
when: </p> <list><item> cross building. This has been
solved by not building the SSL examples during a cross
build. </item><item> building on Windows. </item></list>
<p>
Own Id: OTP-8791</p>
</item>
<item>
<p>
inet:getsockopt for SCTP sctp_default_send_param had a
bug to not initialize required feilds causing random
answers. It is now corrected.</p>
<p>
Own Id: OTP-8795 Aux Id: seq11655 </p>
</item>
<item>
<p>
The hipe_bifs:get_hrvtime/0 BIF now always returns a real
value even if the "perfctr" Linux kernel extension is not
available. It used to return a dummy value. (Thanks to
Mikael Pettersson.)</p>
<p>
Own Id: OTP-8798</p>
</item>
<item>
<p>
Calling a native-code compiled module with an
<c>on_load</c> function could cause a crash. (Thanks to
Mikael Pettersson.)</p>
<p>
Own Id: OTP-8799</p>
</item>
<item>
<p>The emulator could crash while writing a crash dump if
native-compiled modules had been loaded. (Thanks to Paul
Guyot.)</p>
<p>
Own Id: OTP-8801</p>
</item>
<item>
<p>
The garbage collector could crash if invoked from
native-compiled code after a call to a BIF. (Thanks to
Paul Guyot.)</p>
<p>
Own Id: OTP-8821</p>
</item>
<item>
<p>
A rare memory leak in binary:matches is removed</p>
<p>
Own Id: OTP-8823</p>
</item>
<item>
<p>For a socket in the HTTP packet mode, the return value
from <c>gen_tcp:recv/2,3</c> if there is an error in the
header will be <c>{ok,{http_error,String}}</c> instead of
<c>{error,{http_error,String}}</c> to be consistent with
<c>ssl:recv/2,3</c>.</p>
<p>
*** POTENTIAL INCOMPATIBILITY ***</p>
<p>
Own Id: OTP-8831</p>
</item>
</list>
</section>
<section><title>Improvements and New Features</title>
<list>
<item>
<p>
ets:select_reverse/{1,2,3} are now documented.</p>
<p>
Own Id: OTP-7863</p>
</item>
<item>
<p>
External format of integers changed to make full use of
all 32 bits of INTEGER_EXT. This is a compatible change
as old code can read full 32-bit integers but only
produce 28-bit integers as INTEGER_EXT.</p>
<p>
Own Id: OTP-8540 Aux Id: seq11534 </p>
</item>
<item>
<p>
Large parts of the <c>ethread</c> library have been
rewritten. The <c>ethread</c> library is an Erlang
runtime system internal, portable thread library used by
the runtime system itself.</p>
<p>
Most notable improvement is a reader optimized rwlock
implementation which dramatically improve the performance
of read-lock/read-unlock operations on multi processor
systems by avoiding ping-ponging of the rwlock cache
lines. The reader optimized rwlock implementation is used
by miscellaneous rwlocks in the runtime system that are
known to be read-locked frequently, and can be enabled on
ETS tables by passing the <seealso
marker="stdlib:ets#new_2_read_concurrency">{read_concurrency,
true}</seealso> option upon table creation. See the
documentation of <seealso
marker="stdlib:ets#new/2">ets:new/2</seealso> for more
information. The reader optimized rwlock implementation
can be fine tuned when starting the runtime system. For
more information, see the documentation of the <seealso
marker="erts:erl#+rg">+rg</seealso> command line argument
of <c>erl</c>.</p>
<p>
There is also a new implementation of rwlocks that is not
optimized for readers. Both implementations interleaves
readers and writers during contention as opposed to,
e.g., the NPTL (Linux) pthread rwlock implementation
which use either a reader or writer preferred strategy.
The reader/writer preferred strategy is problematic since
it starves threads doing the non-preferred operation.</p>
<p>
The new rwlock implementations in general performs better
in ERTS than common pthread implementations. However, in
some extremely heavily contended cases this is not the
case. Such heavy contention can more or less only appear
on ETS tables. This when multiple processes do very large
amounts of write locked operations simultaneously on the
same table. Such use of ETS is bad regardless of rwlock
implementation, will never scale, and is something we
strongly advise against.</p>
<p>
The new rwlock implementations depend on atomic
operations. If no native atomic implementation is found,
a fallback solution will be used. Using the fallback
implies a performance degradation. That is, it is more
important now than before to build OTP with a native
atomic implementation.</p>
<p>
The <c>ethread</c> library contains native atomic
implementations for, x86 (32 and 64 bit), powerpc (32
bit), sparc V9 (32 and 64 bit), and tilera (32 bit). On
other hardware gcc's builtin support for atomic memory
access will be used if such exists. If no such support is
found, <c>configure</c> will warn about no atomic
implementation available.</p>
<p>
The <c>ethread</c> library can now also use the
<c>libatomic_ops</c> library for atomic memory accesses.
This makes it possible for the Erlang runtime system to
utilize optimized native atomic operations on more
platforms than before. If <c>configure</c> warns about no
atomic implementation available, try using the
<c>libatomic_ops</c> library. Use the <seealso
marker="doc/installation_guide:INSTALL#Advanced-configuration-and-build-of-ErlangOTP">--with-libatomic_ops=PATH</seealso>
<c>configure</c> command line argument when specifying
where the <c>libatomic_ops</c> installation is located.
The <c>libatomic_ops</c> library can be downloaded from:
<url
href="http://www.hpl.hp.com/research/linux/atomic_ops/">http://www.hpl.hp.com/research/linux/atomic_ops/</url></p>
<p>
The changed API of the <c>ethread</c> library has also
caused modifications in the Erlang runtime system.
Preparations for the to come "delayed deallocation"
feature has also been done since it depends on the
<c>ethread</c> library.</p>
<p>
<em>Note</em>: When building for x86, the <c>ethread</c>
library will now use instructions that first appeared on
the pentium 4 processor. If you want the runtime system
to be compatible with older processors (back to 486) you
need to pass the <seealso
marker="doc/installation_guide:INSTALL#Advanced-configuration-and-build-of-ErlangOTP">--enable-ethread-pre-pentium4-compatibility</seealso>
<c>configure</c> command line argument when configuring
the system.</p>
<p>
Own Id: OTP-8544</p>
</item>
<item>
<p>
erlang:localtime_to_universaltime({{2008, 8, 1}, {0, 0,
0}},true) when TZ=UTC now behaves consistently on all
Unix platforms.</p>
<p>
The problem fixed was originally reported by Paul Guyot
on erlang-bugs mailing list:</p>
<p>
http://www.erlang.org/pipermail/erlang-bugs/2008-November/001077.html</p>
<p>
Own Id: OTP-8580</p>
</item>
<item>
<p>
Optimization reducing memory consumption by two words per
ETS object.</p>
<p>
Own Id: OTP-8737</p>
</item>
<item>
<p>
Fixes for unsupported halfword-emulator</p>
<p>
Own Id: OTP-8745</p>
</item>
<item>
<p>
NIF 64-bit integer support; <c>enif_get_int64</c>,
<c>enif_get_uint64</c>, <c>enif_make_int64</c>,
<c>enif_make_uint64</c>.</p>
<p>
Own Id: OTP-8746</p>
</item>
<item>
<p>
Alignment of trailing data in messages has been adjusted.
This in order to be able to pass data of any type as
trailing data in the future.</p>
<p>
Own Id: OTP-8754</p>
</item>
<item>
<p>
The obsolete/driver.h header file has been removed. It
has been obsolete and deprecated since R8B. Drivers that
still include obsolete/driver.h must be updated to
include erl_driver.h.</p>
<p>
*** POTENTIAL INCOMPATIBILITY ***</p>
<p>
Own Id: OTP-8758</p>
</item>
<item>
<p>
Added erlang:system_info(build_type) which makes it
easier to chose drivers, NIF libraries, etc based on
build type of the runtime system.</p>
<p>
The NIF library for crypto can now be built for valgrind
and/or debug as separate NIF libraries that will be
automatically loaded if the runtime system has been built
with a matching build type.</p>
<p>
Own Id: OTP-8760</p>
</item>
<item>
<p>
Further lessened the memory requirements of ETS objects.</p>
<p>
Own Id: OTP-8762</p>
</item>
<item>
<p>The broken elib_malloc alternate memory allocator has
been removed. <c>erlang:system_info(elib_malloc)</c> will
always return <c>false</c>, and in R15,
<c>erlang:system_info(elib_malloc)</c> will fail with a
<c>badarg</c> exception.</p>
<p>
Own Id: OTP-8764</p>
</item>
<item>
<p>
Calling <c>erlang:system_info/1</c> with the new argument
<c>update_cpu_info</c> will make the runtime system
reread and update the internally stored CPU information.
For more information see the documentation of <seealso
marker="erlang#update_cpu_info">erlang:system_info(update_cpu_info)</seealso>.</p>
<p>
The CPU topology is now automatically detected on Windows
systems with less than 33 logical processors. The runtime
system will now, also on Windows, by default bind
schedulers to logical processors using the
<c>default_bind</c> bind type if the amount of schedulers
is at least equal to the amount of logical processors
configured, binding of schedulers is supported, and a CPU
topology is available at startup.</p>
<p>
Own Id: OTP-8765</p>
</item>
<item>
<p>
The SMP ERTS internal child waiter thread used on Linux
system with NPTL was unintentionally disabled during
cross compilation rewrites (OTP-8323 in R13B03). It has
now been re-enabled. Enabling it again gives a slight
performance improvement.</p>
<p>
Own Id: OTP-8774</p>
</item>
<item>
<p>
<c>epmd</c> used to generate a message to the syslog when
it started up, which could be annoying. This has been
changed to only generate the message if the debug swith
is given. (Thanks to Michael Santos.)</p>
<p>
Own Id: OTP-8775</p>
</item>
<item>
<p>
The scheduler wakeup threshold is now possible to adjust
at system boot. For more information see the <seealso
marker="erl#+swt">+swt</seealso> command line argument of
<c>erl</c>.</p>
<p>
Own Id: OTP-8811</p>
</item>
<item>
<p>
The undocumented function inet:ifget/2 has been improved
to return interface hardware address (MAC) on platforms
supporting getaddrinfo() (such as BSD unixes). Note it
still does not work on all platforms for example not
Windows nor Solaris, so the function is still
undocumented.</p>
<p>
Buffer overflow and field init bugs for inet:ifget/2 and
inet:getservbyname/2 has also been fixed.</p>
<p>
Thanks to Michael Santos.</p>
<p>
Own Id: OTP-8816</p>
</item>
<item>
<p>
Optimizations for MIPS when using gcc atomics. (Thanks to
Steve Vinoski)</p>
<p>
Own Id: OTP-8834</p>
</item>
<item>
<p>
Lock optimization in timer functionality.</p>
<p>
Own Id: OTP-8835</p>
</item>
</list>
</section>
<section><title>Known Bugs and Problems</title>
<list>
<item>
<p>Fix epmd and build environment to build on VxWorks</p>
<p>
Own Id: OTP-8838</p>
</item>
</list>
</section>
</section>
<section><title>Erts 5.8</title>
<section><title>Fixed Bugs and Malfunctions</title>
<list>
<item>
<p>
Driver threads, such as async threads, using <seealso
marker="erl_driver#ErlDrvPDL">port data locks</seealso>
peeked at the port status field without proper locking
when looking up the driver queue.</p>
<p>
Own Id: OTP-8475</p>
</item>
<item>
<p>
The use of <c>mmap()</c> was unnecessarily disabled when
cross compiling.</p>
<p>
The <c>configure</c> arguments <c>--with-ssl</c>, and
<c>--with-odbc</c> refused to accept libraries outside of
<c>$erl_xcomp_sysroot</c> when cross compiling for no
good reason.</p>
<p>
The <c>configure</c> argument <c>--with-odbc</c> didn't
handle the value <c>yes</c> correct.</p>
<p>
The <c>configure</c> arguments <c>--with-odbc</c>, and
<c>--without-odbc</c> have also been added to the
configure help.</p>
<p>
(Thanks to Steve Vinoski for reporting these issues)</p>
<p>
Own Id: OTP-8484</p>
</item>
<item>
<p>
A call to the BIF <c>unregister(RegName)</c> when a port
had the name <c>RegName</c> registered in the runtime
system without SMP support caused a runtime system crash.
(Thanks to Per Hedeland for the bugfix and test case.)</p>
<p>
Own Id: OTP-8487</p>
</item>
<item>
<p>
The runtime system crashed if fewer logical processors
were found than reported by <c>sysconf(
SC_NPROCESSORS_CONF)</c>.</p>
<p>
Own Id: OTP-8549</p>
</item>
<item>
<p>
Fix memory management bug causing crash of non-SMP
emulator with async threads enabled. The bug did first
appear in R13B03.</p>
<p>
Own Id: OTP-8591 Aux Id: seq11554 </p>
</item>
<item>
<p>
Port locks could be prematurely destroyed.</p>
<p>
Own Id: OTP-8612</p>
</item>
<item>
<p>The <c>empd</c> program could loop and consume 100%
CPU time if an unexpected error ocurred in
<c>listen()</c> or <c>accept()</c>. Now <c>epmd</c> will
terminate if a non-recoverable error occurs. (Thanks to
Michael Santos.)</p>
<p>
Own Id: OTP-8618</p>
</item>
<item>
<p>
When kernel poll has been enabled, a livelock could in
rare circumstances occur. Problem reported by Chetan
Ahuja, fix by Mikael Pettersson.</p>
<p>
Own Id: OTP-8632</p>
</item>
<item>
<p>
Windows: Closing port of program that stalled without
reading all data could deadlock scheduler thread.</p>
<p>
Own Id: OTP-8641</p>
</item>
<item>
<p>
On some combination of Montavista Linux on Cavium Octeon
processors, some socket-related system calls returned
other numbers than -1 for errors. This caused a core dump
in inet_drv.c. Now the code works around this problem.</p>
<p>
Own Id: OTP-8654</p>
</item>
<item>
<p>
Missing memory barriers in <c>erts_poll()</c> could cause
the runtime system to hang indefinitely.</p>
<p>
Own Id: OTP-8658</p>
</item>
<item>
<p>
<c>ethr_rwmutex_tryrlock()</c> acquired and refused to
acquire a lock with inverted logic. The lock was however
never acquired in a thread unsafe manner. (Thanks to JR
Zhang for noting this issue)</p>
<p>
Own Id: OTP-8662</p>
</item>
<item>
<p>
Extreme combinations of register/unregister in a highly
parallell SMP application could crash the VM. The error
is corrected.</p>
<p>
Own Id: OTP-8663</p>
</item>
<item>
<p>
On Windows, files are now opened with FILE_SHARE_DELETE
to get closer to Unix semantics.</p>
<p>
Own Id: OTP-8667</p>
</item>
<item>
<p>
<c>erlang:system_info(multi_scheduling)</c> sometimes
erroneously returned <c>enabled</c> when it should have
returned <c>blocked</c>.</p>
<p>
Own Id: OTP-8675</p>
</item>
<item>
<p>
Fix bug causing <c>erlang:decode_packet</c> and
<c>enif_make_string</c> to generate faulty strings with
negative character values for ascii values larger than
127. (Thanks to Paul Guyot)</p>
<p>
Own Id: OTP-8685</p>
</item>
<item>
<p>
<c>open_port/2</c> with the <c>spawn</c> and
<c>spawn_executable</c> options can include an
<c>{env,Env}</c> option. In some cases unsetting
variables would not work on Unix (typically if more
variables were unset than were actually present in the
environment).</p>
<p>
Own Id: OTP-8701</p>
</item>
<item>
<p>
A user defined CPU topology set via a call to <seealso
marker="erlang#system_flag_cpu_topology">erlang:system_flag(cpu_topology,
CPUTopology)</seealso> was not properly verified, and
could in worst case cause an emulator crash. The emulator
crash could only occur when a user defined CPU topology
already existed and was redefined.</p>
<p>
Own Id: OTP-8710</p>
</item>
</list>
</section>
<section><title>Improvements and New Features</title>
<list>
<item>
<p>
The grammar for match specifications in ERTS users guide
only described the tracing dialect of match
specifications. An informal grammar for the ETS dialect
is added.</p>
<p>
Own Id: OTP-8086 Aux Id: seq11333 </p>
</item>
<item>
<p>
The module binary from EEP31 (and EEP9) is implemented.</p>
<p>
Own Id: OTP-8217</p>
</item>
<item>
<p>
New NIF API function <c>enif_make_new_binary</c></p>
<p>
Own Id: OTP-8474</p>
</item>
<item>
<p>
The guard BIF <c>is_boolean/1</c> (introduced in R10B)
has now been included in the lists of BIFs allowed in
guards in the Reference Manual.</p>
<p>
Own Id: OTP-8477</p>
</item>
<item>
<p>
Added function <c>zip:foldl/3</c> to iterate over zip
archives.</p>
<p>
Added functions to create and extract escripts. See
<c>escript:create/2</c> and <c>escript:extract/2</c>.</p>
<p>
The undocumented function <c>escript:foldl/3</c> has been
removed. The same functionality can be achieved with the
more flexible functions <c>escript:extract/2</c> and
<c>zip:foldl/3</c>.</p>
<p>
Record fields has been annotated with type info. Source
files as been adapted to fit within 80 chars and trailing
whitespace has been removed.</p>
<p>
Own Id: OTP-8521</p>
</item>
<item>
<p>A regular expression with many levels of parenthesis
could cause a buffer overflow. That has been corrected.
(Thanks to Michael Santos.)</p>
<p>
Own Id: OTP-8539</p>
</item>
<item>
<p>
<c>erlang:decode_packet(httph_bin,..)</c> could return
corrupt header strings or even crash the VM. This has
been fixed. It only happened on 32-bit VM if the header
name was unknown and between 16 and 20 characters long.
Sockets with simular <c>packet</c> option did not suffer
from this bug.</p>
<p>
Own Id: OTP-8548</p>
</item>
<item>
<p>New NIF features:</p> <list><item> Send messages from
a NIF, or from thread created by NIF, to any local
process (<c>enif_send</c>) </item><item> Store terms
between NIF calls (<c>enif_alloc_env</c>,
<c>enif_make_copy</c>) </item><item> Create binary terms
with user defined memory management
(<c>enif_make_resource_binary</c>) </item></list> <p>And
some incompatible changes made to the API. For more
information see the warning text in <seealso
marker="erl_nif">erl_nif(3)</seealso>.</p>
<p>
*** POTENTIAL INCOMPATIBILITY ***</p>
<p>
Own Id: OTP-8555</p>
</item>
<item>
<p>If the '<c>fop</c>' program (needed for building PDF
files) cannot not be found, it is now possible to build
the HTML and man pages anyway (there will also be dummy
PDF files with no real content created). (Thanks to
Tuncer Ayaz.)</p>
<p>
Own Id: OTP-8559</p>
</item>
<item>
<p>When defining macros the closing right parenthesis
before the dot is now mandatory.</p>
<p>
*** POTENTIAL INCOMPATIBILITY ***</p>
<p>
Own Id: OTP-8562</p>
</item>
<item>
<p>Local and imported functions now override the
auto-imported BIFs when the names clash. The pre R14
behaviour was that auto-imported BIFs would override
local functions. To avoid that old programs change
behaviour, the following will generate an error:</p>
<list><item><p>Doing a call without explicit module name
to a local function having a name clashing with the name
of an auto-imported BIF that was present (and
auto-imported) before OTP R14A</p></item>
<item><p>Explicitly importing a function having a name
clashing with the name of an autoimported BIF that was
present (and autoimported) before OTP R14A</p></item>
<item><p>Using any form of the old compiler directive
<c>nowarn_bif_clash</c></p></item> </list> <p>If the BIF
was added or auto-imported in OTP R14A or later,
overriding it with an import or a local function will
only result in a warning,</p> <p>To resolve clashes, you
can either use the explicit module name <c>erlang</c> to
call the BIF, or you can remove the auto-import of that
specific BIF by using the new compiler directive
<c>-compile({no_auto_import,[F/A]}).</c>, which makes all
calls to the local or imported function without explicit
module name pass without warnings or errors.</p> <p>The
change makes it possible to add auto-imported BIFs
without breaking or silently changing old code in the
future. However some current code ingeniously utilizing
the old behaviour or the <c>nowarn_bif_clash</c> compiler
directive, might need changing to be accepted by the
compiler.</p>
<p>
*** POTENTIAL INCOMPATIBILITY ***</p>
<p>
Own Id: OTP-8579</p>
</item>
<item>
<p>
A bug in re that could cause certain regular expression
matches never to terminate is corrected. (Thanks to
Michael Santos and Gordon Guthrie.)</p>
<p>
Own Id: OTP-8589</p>
</item>
<item>
<p>The erlang:open_port spawn and spawn_executable
directives can include an <c>{env,Env}</c> directive to
set up environment variables for the spawned process. A
bug prevented applications from using <c>{env,Env}</c> to
set an environment variable whose value ended with a
'<c>=</c>' (equal sign) character; the trailing equal
sign was mistaken as an indication that an environment
variable was to be cleared from the environment of the
spawned process. (Thanks to Steve Vinoski.)</p>
<p>
Own Id: OTP-8614</p>
</item>
<item>
<p><c>receive</c> statements that can only read out a
newly created reference are now specially optimized so
that it will execute in constant time regardless of the
number of messages in the receive queue for the process.
That optimization will benefit calls to
<c>gen_server:call()</c>. (See <c>gen:do_call/4</c> for
an example of a receive statement that will be
optimized.)</p>
<p>
Own Id: OTP-8623</p>
</item>
<item>
<p>
The functions file:advise/4 and file:datasync/1 have been
added. (Thanks to Filipe David Manana.)</p>
<p>
Own Id: OTP-8637</p>
</item>
<item>
<p>
New NIF API functions: <c>enif_make_atom_len</c>,
<c>enif_make_existing_atom_len</c>,
<c>enif_make_string_len</c>, <c>enif_get_atom_length</c>,
<c>enif_get_list_length</c>, <c>enif_is_list</c>,
<c>enif_is_tuple</c> (by Tuncer Ayaz)</p>
<p>
Own Id: OTP-8640</p>
</item>
<item>
<p>
Support for using gcc's built-in functions for atomic
memory access has been added. This functionallity will be
used if available and no other native atomic
implementation in ERTS is available.</p>
<p>
Own Id: OTP-8659</p>
</item>
<item>
<p>
The number of spinlocks used when implementing atomic
fall-backs when no native atomic implementation is
available has been increased from 16 to 1024.</p>
<p>
Own Id: OTP-8660</p>
</item>
<item>
<p>
Writer preferred pthread read/write locks has been
enabled on Linux.</p>
<p>
Own Id: OTP-8661</p>
</item>
<item>
<p>
The runtime system will by default bind schedulers to
logical processors using the <c>default_bind</c> bind
type if the amount of schedulers are at least equal to
the amount of logical processors configured, binding of
schedulers is supported, and a CPU topology is available
at startup. </p>
<p>
<em>NOTE:</em> If the Erlang runtime system is the only
operating system process that binds threads to logical
processors, this improves the performance of the runtime
system. However, if other operating system processes (as
for example another Erlang runtime system) also bind
threads to logical processors, there might be a
performance penalty instead. If this is the case you, are
are advised to unbind the schedulers using the <seealso
marker="erl#+sbt">+sbtu</seealso> command line argument,
or by invoking <seealso
marker="erlang#system_flag_scheduler_bind_type">erlang:system_flag(scheduler_bind_type,
unbound)</seealso>.</p>
<p>
Own Id: OTP-8666</p>
</item>
<item>
<p>
The recently added BIFs erlang:min/2, erlang:max/2 and
erlang:port_command/3 are now auto-imported (as they were
originally intended to be). Due to the recent compiler
change (OTP-8579), the only impact on old code defining
it's own min/2, max/2 or port_command/3 functions will be
a warning, the local functions will still be used. The
warning can be removed by using
-compile({no_auto_import,[min/2,max/2,port_command/3]}).
in the source file.</p>
<p>
*** POTENTIAL INCOMPATIBILITY ***</p>
<p>
Own Id: OTP-8669 Aux Id: OTP-8579 </p>
</item>
<item>
<p>
There is a new option 'exclusive' to file:open/2 that
uses the OS O_EXCL flag where supported to open the file
in exclusive mode.</p>
<p>
Own Id: OTP-8670</p>
</item>
<item>
<p>
Now, binary_to_term/2 is auto-imported. This will cause a
compile warning if and only if a module has got a local
function with that name.</p>
<p>
*** POTENTIAL INCOMPATIBILITY ***</p>
<p>
Own Id: OTP-8671</p>
</item>
<item>
<p>
Alignment of scheduler data and run queues were adjusted.</p>
<p>
Own Id: OTP-8673</p>
</item>
<item>
<p>Call time breakpoint tracing</p> <list><item>Introduce
a <c>call_time</c> option to
<c>erlang:trace_pattern/3</c>.This option enables call
time breakpoint tracing on code that is executed by
processes with call tracing enabled. Call time tracing
stores the number of calls and the time spent of each
function with this trace pattern enabled. The information
can be retrieved with <c>erlang:trace_info/2</c></item>
<item>Add a scheduler array for BpData. To solve the
issue of multiple schedulers constantly updating the head
pointer to the bp data wheel, each scheduler now has its
own entrypoint to the wheel. This head pointer can be
updated without a locking being taken.</item></list>
<p>Teach call count tracing to use atomics</p>
<list><item>Call count previously used a global lock for
accessing and writing its counter in the breakpoint. This
is now changed to atomics instead.</item> <item>The
change will let call count tracing and cprof to scale
better when increasing the number of
schedulers.</item></list>
<p>
Own Id: OTP-8677</p>
</item>
<item>
<p><c>eprof</c> has been reimplemented with support in
the Erlang virtual machine and is now both faster (i.e.
slows down the code being measured less) and scales much
better. In measurements we saw speed-ups compared to the
old eprof ranging from 6 times (for sequential code that
only uses one scheduler/core) up to 84 times (for
parallel code that uses 8 cores).</p>
<p>Note: The API for the <c>eprof</c> has been cleaned up
and extended. See the documentation.</p>
<p>
*** POTENTIAL INCOMPATIBILITY ***</p>
<p>
Own Id: OTP-8706</p>
</item>
</list>
</section>
</section>
<section><title>Erts 5.7.5.2</title>
<section><title>Known Bugs and Problems</title>
<list>
<item>
<p>
Two bugs in gen_sctp has been corrected: getopts/setopts
hence also send could only be called from socket owner,
and options 'linger', 'rcvbuf' and 'sndbuf' was read from
wrong protocol layer hence read wrong values by getopts.</p>
<p>
Own Id: OTP-9544</p>
</item>
</list>
</section>
</section>
<section><title>Erts 5.7.5.1</title>
<section><title>Fixed Bugs and Malfunctions</title>
<list>
<item>
<p>
Driver threads, such as async threads, using <seealso
marker="erl_driver#ErlDrvPDL">port data locks</seealso>
peeked at the port status field without proper locking
when looking up the driver queue.</p>
<p>
Own Id: OTP-8475</p>
</item>
<item>
<p>
A call to the BIF <c>unregister(RegName)</c> when a port
had the name <c>RegName</c> registered in the runtime
system without SMP support caused a runtime system crash.
(Thanks to Per Hedeland for the bugfix and test case.)</p>
<p>
Own Id: OTP-8487</p>
</item>
<item>
<p>
Fix memory management bug causing crash of non-SMP
emulator with async threads enabled. The bug did first
appear in R13B03.</p>
<p>
Own Id: OTP-8591 Aux Id: seq11554 </p>
</item>
</list>
</section>
</section>
<section><title>Erts 5.7.5</title>
<section><title>Fixed Bugs and Malfunctions</title>
<list>
<item>
<p>
Fix <c>binary_to_term</c> crash on compressed term with
corrupt size field.</p>
<p>
Own Id: OTP-8336</p>
</item>
<item>
<p>
Processes and/or ports could get stuck on a blocked
scheduler when <c>erlang:system_flag(multi_scheduling,
block)</c> was used.</p>
<p>
Processes and/or ports could get stuck on an offline
scheduler when schedulers online were reduced using
<c>erlang:system_flag(schedulers_online,
SchedulersOnline)</c>.</p>
<p>
Own Id: OTP-8342</p>
</item>
<item>
<p>Building on Windows will now work if the paths to
<c>mc.exe</c> and <c>rc.exe</c> contain spaces. The
README.win32 file has been updated with some information
about building using Visual Studio 2008. (Thanks to
Andrew Thompson.)</p>
<p>
Own Id: OTP-8345</p>
</item>
<item>
<p>
EPMD now correctly handles the extra data field which can
be given in the ALIVE2_REQ request and retrieved in the
PORT2_RESP response. (Thanks to Klas Johansson.)</p>
<p>
Own Id: OTP-8361</p>
</item>
<item>
<p>
The configure test for reliable floating point exceptions
has been update to work on modern versions of Mac OS X.
(Thanks to Trannie Carter.)</p>
<p>
Own Id: OTP-8368</p>
</item>
<item>
<p>
ERTS makefiles used to detect the use of a gcc C compiler
by checking if CC equaled gcc. That is, the makefiles
failed to detect gcc C compilers with other command line
names than gcc. `configure' now substitute GCC into the
makefiles. If CC is a gcc C compiler, GCC will have the
value yes. (Thanks to Jean-Sébastien Pédron)</p>
<p>
Own Id: OTP-8373</p>
</item>
<item>
<p>
ETS bug causing the <c>memory</c> counter from
<c>ets:info</c> for ordered_set's to sometimes get out of
synch and absurdly high.</p>
<p>
Own Id: OTP-8377 Aux Id: seq11442 </p>
</item>
<item>
<p>
Immediately repeated multi-scheduling block/unblock
cycles using <c>erlang:system_flag(multi_scheduling,
block | unblock)</c> could deadlock the runtime system.</p>
<p>
Own Id: OTP-8386</p>
</item>
<item>
<p>A number of bugs concerning re and unicode are
corrected:</p>
<p>re:compile no longer loses unicode option, which also
fixes bug in re:split.</p>
<p>re:replace now handles unicode charlist replacement
argument</p>
<p>re:replace now handles unicode RE charlist argument
correctly</p>
<p>re:replace now handles binary unicode output correctly
when nothing is replaced.</p>
<p>Most code, testcases and error isolation done by Rory
Byrne.</p>
<p>
Own Id: OTP-8394</p>
</item>
<item>
<p>
The loading of native code was not properly atomic in the
SMP emulator, which could cause crashes. Also a per-MFA
information table for the native code has now been
protected with a lock since it turns that it could be
accessed concurrently in the SMP emulator. (Thanks to
Mikael Pettersson.)</p>
<p>
Own Id: OTP-8397</p>
</item>
<item>
<p>
Fix processes in exiting status that are about to be
scheduled, to not be allowed to garbage collect.</p>
<p>
Own Id: OTP-8420</p>
</item>
<item>
<p>Removed bogus "\011" character sequence in
documentation.</p>
<p>
Own Id: OTP-8422</p>
</item>
<item>
<p>
The resolver routines failed to look up the own node name
as hostname, if the OS native resolver was erroneously
configured, bug reported by Yogish Baliga, now fixed.</p>
<p>
The resolver routines now tries to parse the hostname as
an IP string as most OS resolvers do, unless the native
resolver is used.</p>
<p>
The DNS resolver inet_res and file resolver inet_hosts
now do not read OS configuration files until they are
needed. Since the native resolver is default, in most
cases they are never needed.</p>
<p>
The DNS resolver's automatic updating of OS configuration
file data (/etc/resolv.conf) now uses the 'domain'
keyword as default search domain if there is no 'search'
keyword.</p>
<p>
Own Id: OTP-8426 Aux Id: OTP-8381 </p>
</item>
<item>
<p>
The re module: A regular expression with an option change
at the start of a pattern that had top-level alternatives
could cause overwriting and/or a crash. (Thanks to
Michael Santos.)</p>
<p>
Own Id: OTP-8438</p>
</item>
<item>
<p>
Harmless buffer overflow by one byte in asn1 and
ram_file_drv.</p>
<p>
Own Id: OTP-8451</p>
</item>
</list>
</section>
<section><title>Improvements and New Features</title>
<list>
<item>
<p>
Improved GC performance after BIF/NIF call when a lot of
heap fragments was created. This will mainly benefit NIFs
that return large compound terms.</p>
<p>
Own Id: OTP-8240</p>
</item>
<item>
<p>Incompatible changes in the experimental NIF
feature.</p> <list><item> Changed the NIF function
prototypes in order to allow more than 3 function
arguments. </item><item> <c>enif_get_data</c> renamed as
<c>enif_priv_data</c>. </item><item>
<c>enif_make_string</c> got a third argument for
character encoding. </item><item> The return value of
<c>erlang:load_nif/2</c> on error changed. </item></list>
<p>Read more in the documentation of <c>erl_nif</c> and
<c>erlang:load_nif/2</c></p>.
<p>
*** POTENTIAL INCOMPATIBILITY ***</p>
<p>
Own Id: OTP-8304</p>
</item>
<item>
<p>Cross compilation improvements and other build system
improvements.</p>
<p>Most notable:</p> <list><item> Lots of cross
compilation improvements. The old cross compilation
support was more or less non-existing as well as broken.
Please, note that the cross compilation support should
still be considered as experimental. Also note that old
cross compilation configurations cannot be used without
modifications. For more information on cross compiling
Erlang/OTP see the <c>$ERL_TOP/INSTALL-CROSS.md</c> file.
</item><item> Support for staged install using <url
href="http://www.gnu.org/prep/standards/html_node/DESTDIR.html">DESTDIR</url>.
The old broken <c>INSTALL_PREFIX</c> has also been fixed.
For more information see the <c>$ERL_TOP/INSTALL.md</c>
file. </item><item> Documentation of the <c>release</c>
target of the top <c>Makefile</c>. For more information
see the <c>$ERL_TOP/INSTALL.md</c> file. </item><item>
<c>make install</c> now by default creates relative
symbolic links instead of absolute ones. For more
information see the <c>$ERL_TOP/INSTALL.md</c> file.
</item><item> <c>$ERL_TOP/configure --help=recursive</c>
now works and prints help for all applications with
<c>configure</c> scripts. </item><item> Doing <c>make
install</c>, or <c>make release</c> directly after
<c>make all</c> no longer triggers miscellaneous
rebuilds. </item><item> Existing bootstrap system is now
used when doing <c>make install</c>, or <c>make
release</c> without a preceding <c>make all</c>.
</item><item> The <c>crypto</c> and <c>ssl</c>
applications use the same runtime library path when
dynamically linking against <c>libssl.so</c> and
<c>libcrypto.so</c>. The runtime library search path has
also been extended. </item><item> The <c>configure</c>
scripts of <c>erl_interface</c> and <c>odbc</c> now
search for thread libraries and thread library quirks the
same way as <c>erts</c> do. </item><item> The
<c>configure</c> script of the <c>odbc</c> application
now also looks for odbc libraries in <c>lib64</c> and
<c>lib/64</c> directories when building on a 64-bit
system. </item><item> The <c>config.h.in</c> file in the
<c>erl_interface</c> application is now automatically
generated in instead of statically updated which reduces
the risk of <c>configure</c> tests without any effect.
</item></list>
<p>(Thanks to Henrik Riomar for suggestions and
testing)</p>
<p>(Thanks to Winston Smith for the AVR32-Linux cross
configuration and testing)</p>
<p>
*** POTENTIAL INCOMPATIBILITY ***</p>
<p>
Own Id: OTP-8323</p>
</item>
<item>
<p>NIF improvements:</p> <list><item> Driver API for
multi-threading made available for NIFs. </item><item>
Support for mempory managed (garbage collected) resource
objects. A way to pass "pointers" to native data
structures between C and Erlang in a safe way.
</item><item> Support for references, floats and term
comparison. </item><item> Various new functions, like
<c>enif_inspect_iolist_as_binary</c>,
<c>enif_make_sub_binary</c>, <c>enif_get_string</c>,
<c>enif_get_atom</c>, <c>enif_make_tuple_from_array</c>,
<c>enif_make_list_from_array</c>,
<c>enif_make_existing_atom</c>. </item></list>
<p>
Own Id: OTP-8335</p>
</item>
<item>
<p>
Minor alignment adjustments of scheduler specific data.</p>
<p>
Own Id: OTP-8341</p>
</item>
<item>
<p>The documentation is now possible to build in an open
source environment after a number of bugs are fixed and
some features are added in the documentation build
process. </p>
<p>- The arity calculation is updated.</p>
<p>- The module prefix used in the function names for
bif's are removed in the generated links so the links
will look like
"http://www.erlang.org/doc/man/erlang.html#append_element-2"
instead of
"http://www.erlang.org/doc/man/erlang.html#erlang:append_element-2".</p>
<p>- Enhanced the menu positioning in the html
documentation when a new page is loaded.</p>
<p>- A number of corrections in the generation of man
pages (thanks to Sergei Golovan)</p>
<p>- The legal notice is taken from the xml book file so
OTP's build process can be used for non OTP
applications.</p>
<p>
Own Id: OTP-8343</p>
</item>
<item>
<p>
There is new <c>erlang:binary_to_term/2</c> BIF that
takes an option list. The option <c>safe</c> can be used
to prevent creation of resources that are not garbage
collected (such as atoms). (Thanks to Jayson Vantuyl.)</p>
<p>
Own Id: OTP-8367</p>
</item>
<item>
<p>
The default settings for garbage collection of binaries
has been adjusted to be less aggressive than in R13B03.
It is now also possible configure the settings for binary
GC. See the documentation for <c>spawn_opt/2-5</c>,
<c>erlang:system_info/1</c>, <c>erlang:system_flag/2</c>,
<c>process_flag/2-3</c>, <c>erlang:trace/3</c>, and the
documenation for <c>erl</c> for the new command line
options <c>+hms</c> and <c>+hmbs</c>.</p>
<p>
Own Id: OTP-8370</p>
</item>
<item>
<p>
A bug causing memory corruption in re:run() has been
corrected. (Thanks to Yamashina Hio.)</p>
<p>
Own Id: OTP-8375</p>
</item>
<item>
<p>
The <c>-Werror</c> option for <c>erlc</c> and the
compiler option <c>warnings_as_errors</c> will cause
warnings to be treated as errors. (Thanks to Christopher
Faulet.)</p>
<p>
Own Id: OTP-8382</p>
</item>
<item>
<p>
Explicit top directories in archive files are now
optional.</p>
<p>
For example, if an archive (app-vsn.ez) just contains an
app-vsn/ebin/mod.beam file, the file info for the app-vsn
and app-vsn/ebin directories are faked using the file
info from the archive file as origin. The virtual
direcories can also be listed. For short, the top
directories are virtual if they does not exist.</p>
<p>
Own Id: OTP-8387</p>
</item>
<item>
<p>
An unnecessary lock operation in os:timestamp/0 has been
eliminated, making it slightly more efficient. (Thanks to
Jonas Falkevik and Tuncer Ayaz.)</p>
<p>
Own Id: OTP-8390</p>
</item>
<item>
<p>
There is a new <c>+t</c> emulator option for changing the
maximum number of atoms. (Thanks to Julien Barbot.)</p>
<p>
Own Id: OTP-8405</p>
</item>
<item>
<p>
Fixed numerous compiler warnings generated by gcc 4.4.1
and tile-cc 2.0.1.78377 when compiling the runtime
system.</p>
<p>
Own Id: OTP-8412</p>
</item>
<item>
<p>
<c>configure</c> learned the option
<c>--enable-m32-build</c> to force the building of a
32-bit run-time on systems where the default C compiler
generates 64-bit executables by default.</p>
<p>
Own Id: OTP-8415</p>
</item>
<item>
<p>
HiPE now works in the 64-bit emulator on Mac OS X.
(Thanks to Geoff Cant.)</p>
<p>
Own Id: OTP-8416</p>
</item>
<item>
<p>
Improved handling of symbolic links to escripts</p>
<p>
Own Id: OTP-8417</p>
</item>
<item>
<p>
Add lock profiling tool.</p>
<p>
The Lock profiling tool, lcnt, can make use of the
internal lock statistics when the runtime system is built
with this feature enabled.</p>
<p>
This provides a mechanism to examine potential lock
bottlenecks within the runtime itself.</p>
<p>
- Add erts_debug:lock_counters({copy_save, bool()}). This
option enables or disables statistics saving for
destroyed processes and ets-tables. Enabling this might
consume a lot of memory.</p>
<p>
- Add id-numbering for lock classes which is otherwise
undefined.</p>
<p>
Own Id: OTP-8424</p>
</item>
<item>
<p>
Removed spurious options to the emulator from escript.</p>
<p>
Own Id: OTP-8427</p>
</item>
<item>
<p>
Minor documentation fixes. Mainly anchor adjustments.</p>
<p>
Own Id: OTP-8457</p>
</item>
</list>
</section>
</section>
<section><title>Erts 5.7.4</title>
<section><title>Fixed Bugs and Malfunctions</title>
<list>
<item>
<p>
An insufficient stack allocation was made when reading
CPU information on BSD operating systems. (Thanks Michael
Turner and Akira Kitada)</p>
<p>
Own Id: OTP-8207</p>
</item>
<item>
<p>
A bug when supplying an argument without a dash directly
after the program name when starting erlang could prevent
distribution to start. This is now corrected.</p>
<p>
Own Id: OTP-8209</p>
</item>
<item>
<p>
A ticker process could potentially be blocked
indefinitely trying to send a tick to a node not
responding. If this happened, the connection would not be
brought down as it should.</p>
<p>
Own Id: OTP-8218</p>
</item>
<item>
<p>
Using certain firewalls (i.e. MS IAS Client and certain
versions of COMODO) could expose an undocumented
behaviour in the Win32 socket interface causing the name
resolution calls to hang infinitely. This is now worked
around by adding possibilities for port programs under
Windows to use overlapped I/O on their standard
input/output file handles.</p>
<p>
Own Id: OTP-8230</p>
</item>
<item>
<p>
Fixed bug on ETS tables with <c>write_concurrency</c>.
The emulator could crash when doing a <c>select</c> or
<c>match</c> with a bound key without finding any object.</p>
<p>
Own Id: OTP-8242</p>
</item>
<item>
<p>The <c>information-request</c> /
<c>information-response</c>, and
<c>group-leader-change-request</c> /
<c>group-leader-changed-response</c> signal pairs
described below did not always adhere to the signal order
guarantees of Erlang's signal model in the runtime system
with SMP support. These signals could for example
sometimes pass exit signals.</p>
<p>The following BIFs behaviors can be modeled as if an
asynchronous <c>information-request</c> signal is sent to
<c>Pid</c>. If <c>Pid</c> is alive, it responds with an
asynchronous <c>information-response</c> signal;
otherwise, the runtime system responds with a
<c>no-such-process</c> signal. When the response is
received, the caller transforms it into the result of the
BIF.</p> <list> <item><c>is_process_alive(Pid)</c></item>
<item><c>erlang:process_display(Pid, Type)</c></item>
<item><c>process_info(Pid)</c></item>
<item><c>process_info(Pid, ItemSpec)</c></item> </list>
<p>When <c>Pid</c> resides on the same node as the caller
of <c>group_leader(GroupLeader, Pid)</c>, the
<c>group_leader/2</c> BIFs behavior can be modeled as if
an asynchronous <c>group-leader-change-request</c> signal
is sent to <c>Pid</c>. If <c>Pid</c> is alive, it
responds with an asynchronous
<c>group-leader-changed-response</c> signal; otherwise,
the runtime system responds with a <c>no-such-process</c>
signal. When the response is received, the caller
transforms it into the result of the BIF. The distributed
case which only consists of an asynchronous
<c>group-leader-change-request</c> signal and no response
is not effected.</p>
<p>
Own Id: OTP-8245</p>
</item>
<item>
<p>
Errors in the <c>system_profile</c> documentation has
been corrected.</p>
<p>
Own Id: OTP-8257</p>
</item>
<item>
<p>
Low watermark socket option modified high watermark
instead of low watermark in the inet_driver. (Thanks to
Feng Yu and Tuncer Ayaz)</p>
<p>
Own Id: OTP-8279</p>
</item>
<item>
<p>
A race condition could cause the runtime system with SMP
support to end up in a completely unresponsive state.</p>
<p>
Own Id: OTP-8297</p>
</item>
</list>
</section>
<section><title>Improvements and New Features</title>
<list>
<item>
<p>
The use of <c>pthread_cond_timedwait()</c> have been
completely removed from the runtime system. This since
its behavior is unpredictable when the system clock is
suddenly changed. The previous use of it was harmless.</p>
<p>
Own Id: OTP-8193</p>
</item>
<item>
<p>
The documentation is now built with open source tools
(xsltproc and fop) that exists on most platforms. One
visible change is that the frames are removed.</p>
<p>
Own Id: OTP-8201</p>
</item>
<item>
<p>
A new garbage collecting strategy for binaries which is
more aggressive than the previous implementation.
Binaries now has a virtual binary heap tied to each
process. When binaries are created or received to a
process it will check if the heap limit has been reached
and if a reclaim should be done. This imitates the
behavior of ordinary Erlang terms. The virtual heaps are
grown and shrunk like ordinary heaps. This will lessen
the memory footprint of binaries in a system.</p>
<p>
Own Id: OTP-8202</p>
</item>
<item>
<p>
The <c>ErlDrvTermData</c> term types used by
<c>driver_output_term()</c> and <c>driver_send_term()</c>
have been extended with the term types
<c>ERL_DRV_INT64</c>, and <c>ERL_DRV_UINT64</c> for
passing 64-bit integers. Also the 64-bit integer data
types <c>ErlDrvSInt64</c> and <c>ErlDrvUInt64</c> have
been introduced.</p>
<p>
For more information see the <seealso
marker="erl_driver">erl_driver(3)</seealso>
documentation.</p>
<p>
Own Id: OTP-8205</p>
</item>
<item>
<p>
[escript] The restriction that the first line in escripts
must begin with <c>#!</c> has been removed.</p>
<p>
[escript] Some command line options to the escript
executable has now been documented. For example you can
run an escript in the debugger by just adding a command
line option.</p>
<p>
[escript] The documentation of the escript header syntax
has been clarified. For example the header is optional.
This means that it is possible to directly "execute"
<c>.erl</c>, <c>.beam</c> and<c>.zip</c> files.</p>
<p>
Own Id: OTP-8215</p>
</item>
<item>
<p>
The instruction for building OTP on Windows was outdated
and incomplete, the document is updated.</p>
<p>
Also the otp_build script required windows drives to show
up in Cygwin using the /cygdrive prefix. That requirement
is now removed.</p>
<p>
Own Id: OTP-8219</p>
</item>
<item>
<p>
A module can have native implemented functions (NIFs)
that are dynamically loaded by calling
<c>erlang:load_nif/2</c>. This is an experimental feature
that is not yet intended for production systems. It is
released with intention to get some early feedback on the
interfaces before they are carved in stone.</p>
<p>
Own Id: OTP-8220</p>
</item>
<item>
<p>
The <c>float/1</c> BIF would always force a garbage
collection. The BIFs <c>size/1</c>, <c>byte_size/1</c>,
<c>bit_size/1</c>, <c>abs/1</c>, and <c>round/1</c> would
force a garbage-collection if the result was not a
sufficiently small integer.</p>
<p>
Own Id: OTP-8221</p>
</item>
<item>
<p>
The <seealso
marker="erlang#port_command/3">erlang:port_command/3</seealso>
BIF has been added. <c>erlang:port_command/3</c> is
currently not auto imported, but it is planned to be auto
imported in OTP R14. For more information see the
<seealso marker="erlang">erlang(3)</seealso>
documentation.</p>
<p>
Own Id: OTP-8225</p>
</item>
<item>
<p>
'<c>configure --enable-darwin-64bit</c>' would fail if
Snow Leopard had been booted with the 64-bit kernel.
(Thanks to Ryan M. Graham.)</p>
<p>
Own Id: OTP-8236</p>
</item>
</list>
</section>
</section>
<section><title>Erts 5.7.3</title>
<section><title>Fixed Bugs and Malfunctions</title>
<list>
<item>
<p>
On Windows, open_port({spawn,Command},Opts) could not run
executables with spaces in the path or filename,
regardless of quoting efforts. While
open_port({spawn_executable,Exec},Opts) can run any
executable, it was still impossible to use 'spawn' to do
the same thing. This is now corrected.</p>
<p>
Own Id: OTP-8055</p>
</item>
<item>
<p>
The scheduler bind type <c>processor_spread</c> spread
schedulers too much on large NUMA systems.</p>
<p>
The new scheduler bind type <c>spread</c> spreads
schedulers as much as possible, and behaves as
<c>processor_spread</c> previously did. For more
information see the documentation of the <c>+sbt</c>
command line argument in the <c>erl(1)</c> documentation,
and the documentation of
<c>erlang:system_flag(scheduler_bind_type,
SchedulerBindType)</c>.</p>
<p>
Own Id: OTP-8063</p>
</item>
<item>
<p>
Automatically detected CPU topology on Linux system could
erroneously contain logical processors with <c>-1</c> as
identifiers. This happened when
<c>sysconf(_SC_NPROCESSORS_CONF)</c> returned a value
larger than the amount of logical processors found.</p>
<p>
Own Id: OTP-8064</p>
</item>
<item>
<p>
When the minimal term [] (end of list) was sent as the
complete message to a process on another node, and
received there, it could not be decoded. This bug is now
corrected. Fortunately [] is uncommon as the complete
message in real applications but it is a serious bug
anyway.</p>
<p>
Own Id: OTP-8092</p>
</item>
<item>
<p>A bug when the floating point exception pointer was
not initialized has been corrected. It manifested itself
on CentOS 5.1 sometimes when a floating point value was
sent to a remote node. Bug reported and patch suggested
by David Reiss, confirmed by Mikael Pettersson.</p>
<p>Some build problems on IRIX was also corrected.
Problem reported by Patrick Baggett, patch by Mikael
Pettersson.</p>
<p>
Own Id: OTP-8095</p>
</item>
<item>
<p>
A terminating process could erroneously unregister a name
for another process. This could occur under the following
conditions: The name of the terminating process was
unregistered and then registered for another process
simultaneously as the process that first had the name was
terminating.</p>
<p>
Own Id: OTP-8099 Aux Id: seq11344 </p>
</item>
<item>
<p>
Running erlc in a very deep directory (with a path length
of more 256 or more characters) would cause the emulator
to crash in a call to <c>list_to_atom/1</c>. (Thanks to
Chris Newcombe.)</p>
<p>
Own Id: OTP-8124</p>
</item>
<item>
<p>
A deadlock of the runtime system could occur when
unregistering the name of a port.</p>
<p>
Own Id: OTP-8145</p>
</item>
<item>
<p>
<c>Makefile.in</c> has been updated to use the LDFLAGS
environment variable (if set). (Thanks to Davide
Pesavento.)</p>
<p>
Own Id: OTP-8157</p>
</item>
<item>
<p>
The pthread rwlock implemention on Linux could cause
starvation of writers. We, therefore, now use our own
rwlock implementation on Linux.</p>
<p>
Own Id: OTP-8158</p>
</item>
<item>
<p>
Open source Erlang builds are updated to work well on
Snow Leopard (MacOS X 10.6)</p>
<p>
Own Id: OTP-8168</p>
</item>
<item>
<p>
A call to <c>erlang:system_info(schedulers_online)</c>
could end up in an infinite loop. This happened if the
amount of schedulers was larger than one, the amount of
schedulers online was one, and someone was blocking
multi-scheduling.</p>
<p>
Own Id: OTP-8169</p>
</item>
<item>
<p>
An error in erlang:system_profile/2 could cause
timestamped messages to arrive out of order in the SMP
case. This has now been fixed.</p>
<p>
Own Id: OTP-8171</p>
</item>
<item>
<p>
<c>binary_to_atom/2</c> and
<c>binary_to_existing_atom/2</c> would leak memory if the
binary contained unaligned data.</p>
<p>
Own Id: OTP-8192</p>
</item>
<item>
<p>
The async thread pool in the runtime system without SMP
support used a memory allocator that was not thread safe
for async jobs.</p>
<p>
Own Id: OTP-8194</p>
</item>
</list>
</section>
<section><title>Improvements and New Features</title>
<list>
<item>
<p>
Processor internal NUMA nodes are now supported in the
ERTS internal CPU topology representation. For more
information see the documentation of the <c>+sct</c>
command line argument in the <c>erl(1)</c> documentation,
and the documentation of
<c>erlang:system_info(cpu_topology)</c>.</p>
<p>
Own Id: OTP-8041</p>
</item>
<item>
<p>
Documentation for ets improved about concurrency.</p>
<p>
Own Id: OTP-8050</p>
</item>
<item>
<p>
Emulator flags in an escript were earlier inherited to
emulators started from from the emulator running the
escript. For example when an escript invoked
<c>os:cmd("erl")</c>, the new emulator were given
erroneous emulator flags. This bug has now been fixed</p>
<p>
Escript filenames may now contain dots.</p>
<p>
Own Id: OTP-8060</p>
</item>
<item>
<p>
Made some BIFs non-recursive (relational operators,hash
and phash) to limit internal stack usage.</p>
<p>
Own Id: OTP-8065</p>
</item>
<item>
<p>
Fixed Windows specific bug in erl_prim_loader. Now it
handles the root directory (e.g. c:/) better. This bug
affected the directory listing in the debugger.</p>
<p>
Own Id: OTP-8080</p>
</item>
<item>
<p>
A TCP socket with option <c>{packet,4}</c> could crash
the emulator if it received a packet header with a very
large size value (>2Gb). The same bug caused
<c>erlang:decode_packet/3</c> to return faulty values.
(Thanks to Georgos Seganos.)</p>
<p>
Own Id: OTP-8102</p>
</item>
<item>
<p>
The maximum size of the export table has been raised from
65536 to 524288 entries.</p>
<p>
Own Id: OTP-8104 Aux Id: seq11345 </p>
</item>
<item>
<p>
The file module has now a read_line/1 function similar to
the io:get_line/2, but with byte oriented semantics. The
function file:read_line/1 works for raw files as well,
but for good performance it is recommended to use it
together with the 'read_ahead' option for raw file
access.</p>
<p>
Own Id: OTP-8108</p>
</item>
<item>
<p>
Fixed bug causing emulator crash when reading a term in
external format containing a corrupt list with a negative
length.</p>
<p>
Own Id: OTP-8117</p>
</item>
<item>
<p>
New emulator flag <c>+sss</c>, to set stack size of
scheduler threads.</p>
<p>
Own Id: OTP-8119</p>
</item>
<item>
<p>
The Windows utility Erlsrv, run in interactive mode now
accepts options for registering internal service name and
description field of Windows registry database.</p>
<p>
Own Id: OTP-8132</p>
</item>
<item>
<p>
<c>erlang:demonitor(Mon, [flush])</c> has been optimized.
Previously it always searched the message queue of the
caller for a <c>'DOWN'</c> message. Current
implementation only search the message queue when
necessary. It is quite common that the search is not
necessary.</p>
<p>
A new option <c>info</c> has been added to
<c>erlang:demonitor/2</c>. For more information see the
<c>erlang(3)</c> documentation.</p>
<p>
Own Id: OTP-8143</p>
</item>
<item>
<p>
I/O tasks could unnecessarily be rescheduled. This was
harmless, but not useful work.</p>
<p>
Own Id: OTP-8148</p>
</item>
<item>
<p>
Minor improvements of <c>erlang:memory/[1,2]</c>.</p>
<p>
Own Id: OTP-8152</p>
</item>
<item>
<p>
New configuration option to enable use of shared zlib.</p>
<p>
Own Id: OTP-8155</p>
</item>
<item>
<p>
Fixed smp bug in ETS that could cause emulator crash when
table with more than 1000 objects accessed by several
processes, including calls to variants of <c>select</c>
or <c>match</c> combined with concurrent object deletion.</p>
<p>
Own Id: OTP-8166 Aux Id: seq11392 </p>
</item>
<item>
<p>
The code path interpretation is now more relaxed. The
flag -code_path_choice now defaults to relaxed instead of
strict. See the documentation of code and init for more
info.</p>
<p>
Own Id: OTP-8170</p>
</item>
<item>
<p>
Load balancing of run queues and check for I/O are
triggered more often than before in situations where
processes are scheduled often but are doing very little
work each time they execute.</p>
<p>
Own Id: OTP-8172</p>
</item>
<item>
<p>
Call tracing binary comprehensions would cause an
emulator crash. (Thanks to Paul Mineiro.)</p>
<p>
Own Id: OTP-8179</p>
</item>
<item>
<p>
<c>binary_to_term/1</c> would crash the emulator instead
of generating a <c>badarg</c> exception when given
certain invalid terms. (Thanks to Scott Lystig Fritchie.)</p>
<p>
Own Id: OTP-8180</p>
</item>
</list>
</section>
</section>
<section><title>Erts 5.7.2</title>
<section><title>Fixed Bugs and Malfunctions</title>
<list>
<item>
<p>
Crash dumps should now cause less problems for the
crashdump_viewer application. (For processes where arity
was non-zero, the arguments are now longer printed - they
used to be printed in a format that was not parseable.)</p>
<p>
Own Id: OTP-7472 Aux Id: seq11019, 11292 </p>
</item>
<item>
<p>
Processes could potentially get stuck on an offline
scheduler.</p>
<p>
Own Id: OTP-7990</p>
</item>
<item>
<p>
<c>binary_to_atom/2</c> and
<c>binary_to_existing_atom/2</c> could leak memory if
they caused a <c>badarg</c> exception.</p>
<p>
Own Id: OTP-7997</p>
</item>
<item>
<p>
A process could under very rare circumstances erroneously
be resumed.</p>
<p>
Own Id: OTP-8000</p>
</item>
<item>
<p>
Load balancing between schedulers could under rare
circumstances cause an emulator crash.</p>
<p>
Own Id: OTP-8008</p>
</item>
<item>
<p>
<c>erlang:memory(processes_used)</c> always returned
<c>0</c> instead of the correct value. (Thanks to Geoff
Cant)</p>
<p>
Own Id: OTP-8022</p>
</item>
</list>
</section>
<section><title>Improvements and New Features</title>
<list>
<item>
<p>
Major improvements of the Erlang distribution for Erlang
runtime systems with SMP support. Previously distribution
port locks were heavily contended, and all encoding and
decoding for a specific distribution channel had to be
done in sequence. Lock contention due to the distribution
is now negligible and both encoding and decoding of
Erlang messages can be done in parallel.</p>
<p>
The old atom cache protocol used by the Erlang
distribution has been dropped since it effectively
prevented all parallel encoding and decoding of messages
passed over the same distribution channel.</p>
<p>
A new atom cache protocol has been introduced which
isolates atom cache accesses and makes parallel encoding
and decoding of messages passed over the same
distribution channel possible. The new atom cache
protocol also use an atom cache size 8 times larger than
before. The new atom cache protocol is documented in the
ERTS users guide.</p>
<p>
Erlang messages received via the distribution are now
decoded by the receiving Erlang processes without holding
any distribution channel specific locks. Erlang messages
and signals sent over the distribution are as before
encoded by the sending Erlang process, but now without
holding any distribution channel specific locks during
the encoding. That is, both encoding and decoding can be
and are done in parallel regardless of distribution
channel used.</p>
<p>
The part that cannot be parallelized is the atom cache
updates. Atom cache updates are therefore now scheduled
on the distribution port. Since it is only one entity per
distribution channel doing this work there is no lock
contention due to the atom cache updates.</p>
<p>
The new runtime system does not understand the old atom
cache protocol. New and old runtime systems can however
still communicate, but no atom cache will be used.</p>
<p>
Own Id: OTP-7774</p>
</item>
<item>
<p>
Fixed a bug that caused error logging from
<c>driver_select</c> sometimes with additional symptoms
such as failing IP communications or even an emulator
crash.</p>
<p>
Own Id: OTP-7898 Aux Id: seq11304 </p>
</item>
<item>
<p>
Improved SMP concurrency for ETS tables. Several mutating
operations can now be performed truly concurrent on
different records of the same table. To support this, the
table has to be created with option
<c>write_concurrency</c>, as it is achieved at the
expense of some execution and memory overhead.
<c>ets:select</c> and <c>select_count</c> has also been
improved for all tables to not acquire exclusive table
lock during the iteration.</p>
<p>
Own Id: OTP-7922</p>
</item>
<item>
<p>
erl (that is erl.exe and dyn_erl) and erlexec has been
made more dynamic so no hard coded paths needs to added
at installation time to erl (that is erl.ini and erl).
Reltool will make use of this in a future release.</p>
<p>
Own Id: OTP-7952</p>
</item>
<item>
<p>
Added functionality to get higher resolution timestamp
from system. The erlang:now function returns a timestamp
that's not always consistent with the actual operating
system time (due to resilience against large time changes
in the operating system). The function os:timestamp/0 is
added to get a similar timestamp as the one being
returned by erlang:now, but untouched by Erlangs time
correcting and smoothing algorithms. The timestamp
returned by os:timestamp is always consistent with the
operating systems view of time, like the calendar
functions for getting wall clock time, but with higher
resolution. Example of usage can be found in the os
manual page.</p>
<p>
Own Id: OTP-7971</p>
</item>
<item>
<p>
Two new options are added to open_port - spawn_executable
which runs external executables in a controlled way, and
spawn_driver which only opens port to loaded Erlang
drivers. See the erlang manual page for details.</p>
<p>
Own Id: OTP-7995</p>
</item>
<item>
<p>
New functionality in ETS to transfer the ownership of a
table. A table can either change owner be declaring an
"heir", another process that will inherit the table if
the owner terminates. A table can also change owner by
calling a new function <c>ets:give_away</c>.</p>
<p>
Own Id: OTP-8006</p>
</item>
<item>
<p>
Updates to Tilera build environment.</p>
<p>
Own Id: OTP-8009</p>
</item>
<item>
<p>
A stack trace was unnecessarily saved during process
termination.</p>
<p>
Own Id: OTP-8014</p>
</item>
<item>
<p>
User defined CPU topology and scheduler bind type can now
be set from the command line when starting an emulator.
For more information see the documentation of the
<c>+sct</c>, and the <c>+sbt</c> emulator flags in the
<c>erl(1)</c> documentation.</p>
<p>
The CPU topologies returned from
<c>erlang:system_info/1</c> and
<c>erlang:system_flag/2</c> now always contain the
<c>processor</c> level, also when not strictly necessary.</p>
<p>
Own Id: OTP-8030</p>
</item>
<item>
<p>
Various fixes in ETS: <c>ets:first</c> could return a
deleted key in a fixated table. <c>ets:lookup</c> could
return objects out of order if a deleted object was
re-inserted into a fixed bag. <c>ets:delete_object</c>
could fail to delete duplicate objects in a
duplicate_bag.</p>
<p>
Own Id: OTP-8040</p>
</item>
</list>
</section>
</section>
<section><title>Erts 5.7.1</title>
<section><title>Fixed Bugs and Malfunctions</title>
<list>
<item>
<p>
Fixed a bug on Windows that could make
<c>gen_tcp:send</c> hang trying to send an iolist of more
than 16 binaries.</p>
<p>
Own Id: OTP-7816</p>
</item>
<item>
<p>
The runtime system could under rare circumstances crash
during load balancing.</p>
<p>
Own Id: OTP-7908 Aux Id: otp-7500 </p>
</item>
<item>
<p>
<c>run_erl</c> uses fallback if Unix98 pseudo-terminal is
not present on host.</p>
<p>
Own Id: OTP-7916 Aux Id: seq11249 </p>
</item>
<item>
<p>
A message buffer memory leak in the runtime system
without smp support has been fixed.</p>
<p>
Own Id: OTP-7941</p>
</item>
<item>
<p>Attempting to append a binary of 16Mb or greater to
another binary using the bit syntax would cause a
<c>system_limit</c> exception. There was also several
cases when constructing binaries when a <c>badarg</c>
exception was generated when it should have been
<c>system_limit</c>.</p>
<p>
Own Id: OTP-7942</p>
</item>
<item>
<p>
The runtime system with SMP support failed to terminate
the caller of <c>link(RemotePid)</c> properly, if
<c>RemotePid</c> was the pid of a process on an
unreachable node. The calling process was in this case
marked as exiting, but never terminated.</p>
<p>
Own Id: OTP-7946</p>
</item>
</list>
</section>
<section><title>Improvements and New Features</title>
<list>
<item>
<p>
Rudimentary support for cross compiling is added to the
source release. The support is still in its infancy and
has only been used to cross compile on Linux for a
different cpu architecture and a different Linux version,
but should be extendible to support other platforms as
well. The cross configuration files with examples are
placed in $ERL_TOP/xcomp/. View README.xcomp and run
$ERL_TOP/otp_build -help for further information.</p>
<p>
Own Id: OTP-7854</p>
</item>
<item>
<p>The escape sequence <c>\{</c> which was given a new
interpretation in R13A has retained its old meaning (the
ASCII code for <c>{</c>), which means that codes greater
than 255 have to be stated using hexadecimal characters
(for example, <c>\x{AAA}</c>). The escape sequence
<c>\xH</c> where H is a hexadecimal character followed by
something else but a hexadecimal character is no longer
valid (incompatibility with R13A). Character codes less
than 256 can be stated using two hexadecimal characters
(for example, <c>\x0D</c>).</p>
<p>
Own Id: OTP-7891 Aux Id: OTP-7855 </p>
</item>
<item>
<p>The <c>term_to_binary/1</c> BIF used to be implemented
with recursive C code, which could cause the Erlang
emulator to terminate because of a stack overflow.</p>
<p>Also fixed some minor issues in
<c>term_to_binary/1</c> and <c>binary_to_term/1</c>
pointed out by Matthew Dempsky.</p>
<p>
Own Id: OTP-7894</p>
</item>
<item>
<p>
Several glitches and performance issues in the Unicode
and I/O-system implementation of R13A have been
corrected.</p>
<p>
Own Id: OTP-7896 Aux Id: OTP-7648 OTP-7887 </p>
</item>
<item>
<p>
Minor documentation improvements of the
<c>scheduler_bind_type</c> argument of
<c>erlang:system_flag/2</c>, and the
<c>scheduler_bind_type</c>, and the
<c>scheduler_bindings</c> arguments of
<c>erlang:system_info/1</c>.</p>
<p>
Own Id: OTP-7901 Aux Id: OTP-7777 </p>
</item>
<item>
<p>
There is a new BIF <c>erlang:make_tuple/3</c>.</p>
<p>
Own Id: OTP-7913</p>
</item>
</list>
</section>
</section>
<section><title>Erts 5.7</title>
<section><title>Fixed Bugs and Malfunctions</title>
<list>
<item>
<p><em>OpenSource:</em></p>
<p>FreeBSD leap-seconds are handled according to patch
submitted by OpenSource user Kenji Rikitake. No test case
covers this functionality (unsupported platform).</p>
<p>
Own Id: OTP-7609</p>
</item>
<item>
<p>
A corrected bug in <c>ets</c> for <c>bag</c> and
<c>duplicate_bag</c>. A <c>delete/2</c> or
<c>lookup_element/3</c> could miss objects in a fixed
table if one or more objects with the same key had
already been deleted.</p>
<p>
Own Id: OTP-7665</p>
</item>
<item>
<p>
A new driver call-back <c>stop_select</c> is introduced
to allow drivers to de-select and then close a file
descriptor in a safe way in a SMP emulator. The old way
was not strictly according to posix standard and could in
some rare cases lead to unexpected behavior. A new flag
<c>ERL_DRV_USE</c> can be passed to
<c>driver_select()</c> to tell it that the descriptor
should be closed. <c>stop_select</c> is then called when
it is safe to do so. Old drivers will however still work
as before.</p>
<p>
Own Id: OTP-7670</p>
</item>
<item>
<p>
A bug fixed for TCP sockets with option
<c>{packet,http}</c>. An HTTP request with an absolute
URI was returned with a corrupt path string. This bug did
only exist in R12B-4 and R12B-5.</p>
<p>
Own Id: OTP-7682 Aux Id: OTP-7647 </p>
</item>
<item>
<p>
run_erl did in some cases fail to extract control
sequences from to_erl (like: winsize=X,Y) and did instead
send them to be interpreted by the erlang shell.</p>
<p>
Own Id: OTP-7688</p>
</item>
<item>
<p>
A bug in the installer on Windows not updating file
associations properly is now corrected.</p>
<p>
Own Id: OTP-7746</p>
</item>
<item>
<p>More space than necessary could be allocated in
binaries when appending to a binary (also in a binary
comprehension) and the data appended did not consist of
wholes bytes (e.g. 13 bits).</p>
<p>
Own Id: OTP-7747</p>
</item>
<item>
<p>
The gen_sctp option sctp_peer_addr_params,
#sctp_paddrparams{address={IP,Port} was erroneously
decoded in the inet driver. This bug has now been
corrected.</p>
<p>
Own Id: OTP-7755</p>
</item>
<item>
<p>
Outstanding async driver jobs leaked memory if the
issuing port died before the async jobs completed.</p>
<p>
Own Id: OTP-7784</p>
</item>
<item>
<p>A bug in the dynamic library loading affecting, among
others, OpenSolaris is now corrected. (Thanks to Paul
Fisher.)</p>
<p>
Own Id: OTP-7796</p>
</item>
<item>
<p>
run_erl compile errors fixed for FreeBSD</p>
<p>
Own Id: OTP-7817</p>
</item>
<item>
<p>
A bug in the inet driver for SCTP on Solaris showing for
e.g gen_sctp:abort/1 and gen_sctp:eof/1 has been
corrected. Patch suggestion by Simon Cornish.</p>
<p>
Own Id: OTP-7866</p>
</item>
</list>
</section>
<section><title>Improvements and New Features</title>
<list>
<item>
<p>
The order of objects visited in select for ordered_set is
now documented.</p>
<p>
Own Id: OTP-7339</p>
</item>
<item>
<p>
The runtime system with SMP support now uses multiple,
scheduler specific run queues, instead of one globally
shared run queue.</p>
<p>
The lock protecting the shared run queue was heavily
contended, and the shared run queue also caused Erlang
processes to randomly migrate between schedulers with
negative cache effects as a result.</p>
<p>
With the current scheduler specific run queue solution,
lock contention due to run queue protection has been
reduced, and Erlang processes are only migrated when
needed to balance the load between the schedulers. The
reduced amount of migration also reduce lock contention
on locks protecting the scheduler specific instances of
the erts internal memory allocators.</p>
<p>
The scheduler specific run queues are also a necessity
for a lot of future planned NUMA (Non-Uniform Memory
Access) specific optimizations.</p>
<p>
Own Id: OTP-7500</p>
</item>
<item>
<p>Support for Unicode is implemented as described in
EEP10. Formatting and reading of unicode data both from
terminals and files is supported by the io and io_lib
modules. Files can be opened in modes with automatic
translation to and from different unicode formats. The
module 'unicode' contains functions for conversion
between external and internal unicode formats and the re
module has support for unicode data. There is also
language syntax for specifying string and character data
beyond the ISO-latin-1 range.</p>
<p>The interactive shell will support input and output of
unicode characters when the terminal and operating system
supports it.</p>
<p>Please see the EEP and the io/io_lib manual pages as
well as the stdlib users guide for details.</p>
<p><em>I/O-protocol incompatibilities:</em></p>
<p>The io_protocol between io_Server and client is
updated to handle protocol data in unicode formats. The
updated protocol is now documented. The specification
resides in the stdlib <em>users manual</em>, which is a
new part of the manual.</p>
<p><em>io module incompatibilities:</em></p>
<p>The io:put_chars, io:get_chars and io:get_line all
handle and return unicode data. In the case where
binaries can be provided (as to io:put_chars), they shall
be encoded in UTF-8. When binaries are returned (as by
io:get_line/get_chars when the io_server is set in
<em>binary mode</em>) the returned data is also
<em>always</em> encoded as UTF-8. The file module however
still returns byte-oriented data, why file:read can be
used instead of io:get_chars to read binary data in
ISO-latin-1.</p>
<p><em>io_lib module incompatibilities:</em></p>
<p>io_lib:format can, given new format directives (i.e
"~ts" and "~tc"), return lists containing integers larger
than 255. </p>
<p>
*** POTENTIAL INCOMPATIBILITY ***</p>
<p>
Own Id: OTP-7648 Aux Id: OTP-7580 OTP-7514 OTP-7494
OTP-7443 OTP-7181 EEP10 EEP11 </p>
</item>
<item>
<p>
The format of the string returned by
<c>erlang:system_info(system_version)</c> (as well as the
first message when Erlang is started) has changed. The
string now contains the both the OTP version number as
well as the erts version number.</p>
<p>
Own Id: OTP-7649</p>
</item>
<item>
<p>
Message passing has been further optimized for parallel
execution. Serial message passing is slightly more
expensive than before, but parallel send to a common
receiver is much cheaper.</p>
<p>
Own Id: OTP-7659</p>
</item>
<item>
<p>
Lock contention on the atom table lock when decoding
Erlang terms on the external format has been drastically
reduced.</p>
<p>
Own Id: OTP-7660</p>
</item>
<item>
<p>
The undocumented, unsupported, and deprecated guard BIF
<c>is_constant/1</c> has been removed.</p>
<p>
*** INCOMPATIBILITY with R12B ***</p>
<p>
Own Id: OTP-7673</p>
</item>
<item>
<p>
The Erlang process lock implementation has been improved
by Mat Hostetter at Tilera Corporation.</p>
<p>
Own Id: OTP-7692</p>
</item>
<item>
<p>
A <c>{nodedown, Node}</c> message passed by the
<c>net_kernel:monitor_nodes/X</c> functionality is now
guaranteed to be sent after <c>Node</c> has been removed
from the result returned by <c>erlang:nodes/Y</c>.</p>
<p>
Own Id: OTP-7725</p>
</item>
<item>
<p>The short-circuit operators <c>andalso</c> and
<c>orelse</c> no longer guarantees that their second
argument is either <c>true</c> or <c>false</c>. As a
consequence, <c>andalso</c>/<c>orelse</c> are now
tail-recursive.</p>
<p>
*** POTENTIAL INCOMPATIBILITY ***</p>
<p>
Own Id: OTP-7748</p>
</item>
<item>
<p>
A new BIF, <c>lists:keyfind/3</c>, has been added. It
works like <c>lists:keysearch/3</c> except that it does
not wrap the returned tuple in a <c>value</c> tuple in
case of success. (Thanks to James Hague for suggesting
this function.)</p>
<p>
Own Id: OTP-7752</p>
</item>
<item>
<p>
Optimization for drivers by creating small binaries
direct on process heap.</p>
<p>
Own Id: OTP-7762</p>
</item>
<item>
<p><c>I bsl N</c> could cause the Erlang virtual machine
to run of memory instead generating a <c>system_limit</c>
if N was absurdly huge. (Thanks to Daniel Hedlund.)</p>
<p>There would always be a garbage collection when
evaluating <c>I bsl N</c> or <c>I bsr N</c> if <c>I</c>
was a bignum.</p>
<p>If <c>I</c> is an integer and <c>N</c> a bignum, <c>I
bsl N</c> will now cause the correct <c>system_limit</c>
exception instead of <c>bad_arith</c> as in earlier
releases.</p>
<p>If <c>I</c> is an integer and <c>N</c> a bignum, <c>I
bsr N</c> will return either 0 or -1 depending on the
sign of <c>I</c> instead of causing a <c>bad_arith</c>
exception as in earlier releases.</p>
<p>
Own Id: OTP-7764</p>
</item>
<item>
<p>
Scheduler threads can now be bound to logical processors
on newer Linux and Solaris systems. More systems will be
supported in the future.</p>
<p>
In some cases performance has increased drastically when
binding schedulers. Schedulers are not bound by default,
though. This since it might cause a performance
degradation if multiple programs have bound to
processors, e.g. multiple Erlang runtime systems. For
more information see the documentation of
<c>erlang:system_flag/2</c>.</p>
<p>
In order to bind scheduler threads the CPU topology need
to be known. On some newer Linux and Solaris systems the
runtime system automatically detects the CPU topology. If
the emulator isn't able to automatically detect the CPU
topology, the CPU topology can be defined. For more
information see the documentation of
<c>erlang:system_flag/2</c>.</p>
<p>
Own Id: OTP-7777</p>
</item>
<item>
<p>The compiler will refuse to a compile file where the
module name in the file differs from the output file
name.</p>
<p>When compiling using <c>erlc</c>, the current working
directory will no be included in the code path (unless
explicitly added using "-pa .").</p>
<p>
*** POTENTIAL INCOMPATIBILITY ***</p>
<p>
Own Id: OTP-7793</p>
</item>
<item>
<p>
The BIFs <c>atom_to_binary/2</c>,
<c>binary_to_atom/2</c>, and
<c>binary_to_existing_atom/2</c> have been added.</p>
<p>
Own Id: OTP-7804</p>
</item>
<item>
<p>
The amount of schedulers online can now be changed during
operation. The amount of schedulers online defaults to
the same amount as available logical processors. For more
information see the documentation of
<c>erlang:system_flag/2</c> and <c>erl</c>.</p>
<p>
Own Id: OTP-7811</p>
</item>
<item>
<p>The deprecated functions <c>erlang:fault/1</c>,
<c>erlang:fault/2</c>, and <c>file:rawopen/2</c> have
been removed.</p>
<p>
*** POTENTIAL INCOMPATIBILITY ***</p>
<p>
Own Id: OTP-7812</p>
</item>
<item>
<p>
Erts internal dynamically allocated process and port
specific data for rarely used data. This is used to
reduce memory usage of processes and ports that do not
use specific functionality. More functionality will be
moved to process and port specific data in future
releases.</p>
<p>
Own Id: OTP-7818</p>
</item>
<item>
<p>
New packet type <c>http_bin</c> for gen_tcp sockets and
<c>erlang:decode_packet</c>. It works like <c>http</c>
except that strings are returned as binaries instead of
lists.</p>
<p>
Own Id: OTP-7821</p>
</item>
<item>
<p>
The obsolete wd_keeper program for embedded Solaris
systems has been removed.</p>
<p>
Own Id: OTP-7822</p>
</item>
<item>
<p>
Nodes belonging to different independent clusters can now
co-exist on the same host with the help of a new
environment variable setting ERL_EPMD_PORT.</p>
<p>
Own Id: OTP-7826</p>
</item>
<item>
<p>There are new functions <c>erlang:min/2</c> and
<c>erlang:max/2</c> to calculate the minimum and maximum
of two terms, respectively. Note that the functions are
not auto-imported, so they need to be imported explicitly
or the <c>erlang</c> prefix must be used when calling
them.</p>
<p>
Own Id: OTP-7841</p>
</item>
<item>
<p>The copyright notices have been updated.</p>
<p>
Own Id: OTP-7851</p>
</item>
<item>
<p>Enhanced build environment for cross compilation to
Tilera Tile architecture.</p>
<p>Support for native ethread atomics on Tilera
Tile64/TilePro (Thanks to Tilera Corporation).</p>
<p>
Own Id: OTP-7852</p>
</item>
<item>
<p>The escape sequences <c>\x</c> and <c>\{</c> have been
assigned new interpretations (they used to return the
ASCII code for <c>x</c> and <c>{</c> respectively). One
or more octal characters inside curly brackets after a
leading backslash is from now on an alternative to the
existing syntax <c>\NNN</c>, but can also be used for
codes greater than 255. In a similar fashion, one or more
hexadecimal characters can be put inside curly brackets
after a leading <c>\x</c>. Furthermore, the escape
sequences <c>\xH</c> and <c>\xHH</c>, where N is a
hexadecimal character, can be used for codes less than
256.</p>
<p>NOTE: These new escape sequences are still considered
experimental and may be changed in the R13B release.</p>
<p>
*** POTENTIAL INCOMPATIBILITY ***</p>
<p>
Own Id: OTP-7855</p>
</item>
<item>
<p>
The PCRE library's exported function names are now
prefixed with erts_ in the erlang emulator to avoid
clashes with dynamically loaded drivers.</p>
<p>
Own Id: OTP-7861</p>
</item>
<item>
<p>
A runtime system with SMP support will now be built by
default on most platforms if a usable posix thread
library or native windows threads are found.</p>
<p>
For more information see the top README file.</p>
<p>
Own Id: OTP-7872</p>
</item>
</list>
</section>
</section>
<section><title>Erts 5.6.5.1</title>
<section><title>Fixed Bugs and Malfunctions</title>
<list>
<item>
<p>
A corrected bug in <c>ets</c> for <c>bag</c> and
<c>duplicate_bag</c>. A <c>delete/2</c> or
<c>lookup_element/3</c> could miss objects in a fixed
table if one or more objects with the same key had
already been deleted.</p>
<p>
Own Id: OTP-7665</p>
</item>
<item>
<p>
A bug fixed for TCP sockets with option
<c>{packet,http}</c>. An HTTP request with an absolute
URI was returned with a corrupt path string. This bug did
only exist in R12B-4 and R12B-5.</p>
<p>
Own Id: OTP-7682 Aux Id: OTP-7647 </p>
</item>
<item>
<p>
Calling <c>gen_tcp:send()</c> from several processes on
socket with option <c>send_timeout</c> could lead to much
longer timeout than specified. The solution is a new
socket option <c>{send_timeout_close,true}</c> that will
do automatic close on timeout. Subsequent calls to send
will then immediately fail due to the closed connection.</p>
<p>
Own Id: OTP-7731 Aux Id: seq11161 </p>
</item>
<item>
<p>
A process being garbage collected via the
<c>garbage_collect/1</c> BIF or the
<c>check_process_code/2</c> BIF didn't handle message
receive and resume correctly during the garbage collect.
When this occurred, the process returned to the state it
had before the garbage collect instead of entering the
new state.</p>
<p>
Own Id: OTP-7738</p>
</item>
</list>
</section>
</section>
<section><title>Erts 5.6.5</title>
<section><title>Fixed Bugs and Malfunctions</title>
<list>
<item>
<p>
A bug in inet_drv concerning gen_tcp:connect has been
corrected. A connect towards a non-open port through open
firewalls could sometimes erroneously be successful. Any
subsequent operation would fail, though.</p>
<p>
Own Id: OTP-6542</p>
</item>
<item>
<p>
Floating point arithmetics in drivers could cause a
runtime system crash and/or unexpected results on runtime
systems with floating point exceptions enabled. Floating
point exceptions are disabled unless explicitly enabled
or if hipe is enabled.</p>
<p>
Own Id: OTP-7237</p>
</item>
<item>
<p>
A bug when many sockets got signalled simultaneously
causing the emulator to panic with the message
"Inconsistent, why isnt io reported?" is now corrected.</p>
<p>
Own Id: OTP-7420</p>
</item>
<item>
<p>
Starting erl with option "-detached" now disconnects
correctly from terminal session on Unix.</p>
<p>
Own Id: OTP-7461</p>
</item>
<item>
<p>
Mended gdb etp-commands for ETS access.</p>
<p>
Own Id: OTP-7538</p>
</item>
<item>
<p>
<c>erlang:decode_packet/3</c> allows white space between
HTTP header tag and colon according to RFC2616.</p>
<p>
Own Id: OTP-7543</p>
</item>
<item>
<p>
An emulator compiled for SCTP now starts even if the
dynamic libraries are not present. The SCTP driver is
then of course not loaded.</p>
<p>
Own Id: OTP-7551</p>
</item>
<item>
<p>To build on Mac OS X, 10.3.0 or later is now required
because of fixes for two problems:</p>
<p>There would be a resource leak when <c>erl_ddl</c>
attempted to unload a driver. This problem has been
corrected by using <c>dlopen()</c> (which works on all
modern Unix-like platforms) instead of the Mac OS X
specific API calls.</p>
<p>Signal handling in the run-time system for HiPE has
been updated to work on later versions of Mac OS X than
10.2.x. Therefore, <c>--enable-hipe</c> now works on Mac
OS X with Intel CPUs.</p>
<p>Thanks to Geoff Cant for the patches.</p>
<p>
Own Id: OTP-7562</p>
</item>
<item>
<p>Corrected some information about the protocol between
EPMD and Erlang nodes. (Thanks to Michael Regen.)</p>
<p>
Own Id: OTP-7594</p>
</item>
<item>
<p>
When using
<c>erlang:system_monitor(Pid,{long_gc,Time})</c>, and the
GC time exceeded 1 second, it sometimes erroneously
showed up as about 4300 seconds. (This bug was corrected
in R9C, but re-introduced in R12B.) (Thanks to Chris
Newcombe.)</p>
<p>
Own Id: OTP-7622 Aux Id: OTP-4903, seq8379 </p>
</item>
</list>
</section>
<section><title>Improvements and New Features</title>
<list>
<item>
<p>
The driver entry of a dynamically loaded driver is now
copied when loaded which enables some internal
optimizations. Note that drivers that modify the driver
entry during execution will not work anymore. Such a
miss-use of the driver interface is however not
supported.</p>
<p>
Own Id: OTP-6900</p>
</item>
<item>
<p>
The split function is now added to the re library.
Exceptions and errors from both run, replace and split
are made more consistent.</p>
<p>
Own Id: OTP-7514 Aux Id: OTP-7494 </p>
</item>
<item>
<p>
Fixed harmless compiler warnings when building the
emulator and minor build changes in order to avoid
unnecessary rebuilds.</p>
<p>
Own Id: OTP-7530</p>
</item>
<item>
<p>
There is now experimental support for loading of code
from archive files. See the documentation of <c>code</c>,
<c>init</c>, <c>erl_prim_loader </c> and <c>escript</c>
for more info.</p>
<p>
The error handling of <c>escripts</c> has been improved.</p>
<p>
An <c>escript</c> may now set explicit arguments to the
emulator, such as <c>-smp enabled</c>.</p>
<p>
An <c>escript</c> may now contain a precompiled beam
file.</p>
<p>
An <c>escript</c> may now contain an archive file
containing one or more applications (experimental).</p>
<p>
The internal module <c>code_aux</c> has been removed.</p>
<p>
Own Id: OTP-7548 Aux Id: otp-6622 </p>
</item>
<item>
<p>
The reallocation functionality part of the ERTS internal
memory allocators, now consider current block in
combination with surrounding free blocks as an
alternative location for a reallocation.</p>
<p>
Own Id: OTP-7555</p>
</item>
<item>
<p>There could remain false references from a process to
a module that has been called earlier, so that the
process would be killed if the module was reloaded.
(Thanks to Richard Carlsson.)</p>
<p>Also, the fix for this bug also made it possible to
make stack backtraces (as returned from
<c>erlang:get_stacktrace/0</c> and other functions) more
correct in that the immediate caller is always included
in the stack backtrace (it could sometimes be
missing).</p>
<p>
Own Id: OTP-7559</p>
</item>
<item>
<p>
Improved locking in IO-handling for better smp
performance.</p>
<p>
Own Id: OTP-7560</p>
</item>
<item>
<p>
Improved BIF rescheduling functionality.</p>
<p>
Own Id: OTP-7587</p>
</item>
<item>
<p>
Loading a module compiled with Erlang/OTP R9C and calling
<c>module_info/0</c> in the module would crash the
emulator. The emulator now refuses to load any module
compiled with R9C or earlier. (Note: only trivial modules
compiled with R10B or earlier could be loaded anyway.)
(Thanks to Martin Kjellin.)</p>
<p>
Own Id: OTP-7590</p>
</item>
</list>
</section>
</section>
<section><title>Erts 5.6.4.2</title>
<section><title>Fixed Bugs and Malfunctions</title>
<list>
<item>
<p>
A process calling one of the following BIFs could under
very rare conditions deadlock in the runtime system with
SMP support: <c>check_process_code/2</c>,
<c>garbage_collect/1</c>, <c>process_info/[1,2]</c>,
<c>system_flag/2</c>, and
<c>erlang:suspend_process/[1,2]</c>.</p>
<p>
Own Id: OTP-7582</p>
</item>
<item>
<p>
A couple of statistics variables were not managed in a
thread safe manner in the runtime system with SMP
support.</p>
<p>
Own Id: OTP-7583</p>
</item>
<item>
<p>
An extremely rare race condition when terminating a
process could potentially cause a runtime system crash.</p>
<p>
Own Id: OTP-7584</p>
</item>
<item>
<p>
Under certain conditions and when using run_erl/to_erl,
the terminal Erlang driver (ttsl_drv) could crash the
emulator by doing a division by zero due to incorrect
handling of terminals reporting a zero width. For
terminals reporting zero width, the driver now fallbacks
to a default width of 80 and a default height of 24
(vt100), as a fallback behaviour. This fixes the crashes
and also makes output on "dumb" terminals much more
readable.</p>
<p>
Own Id: OTP-7592 Aux Id: seq11073 </p>
</item>
</list>
</section>
</section>
<section><title>Erts 5.6.4.1</title>
<section><title>Improvements and New Features</title>
<list>
<item>
<p>
A new <c>erts_alloc</c> parameter
<c>+M<S>rmbcmt</c> (relative multiblock carrier
move threshold) has been added. It determines when to
force a moving realloc in a multiblock carrier when a
block is shrunk. For more information see the
<c>erts_alloc(3)</c> documentation.</p>
<p>
Own Id: OTP-7540</p>
</item>
<item>
<p>The new option <c>+d</c> can be given to <c>erl</c> to
suppress the crash dump generation if an internal error
is detected. As a result, a more useful core dump is
produced.</p>
<p>
Own Id: OTP-7578 Aux Id: seq11052 </p>
</item>
</list>
</section>
</section>
<section><title>Erts 5.6.4</title>
<section><title>Fixed Bugs and Malfunctions</title>
<list>
<item>
<p>
Double backslashes in format string passed to the erts
internal printf implementation produced erroneous
results. No such format strings were passed to the erts
internal printf implementation, i.e., the bug was
therefore completely harmless. (Thanks to Perry Smith.)</p>
<p>
Own Id: OTP-7408</p>
</item>
<item>
<p>
Large files are now handled on Windows, where the
filesystem supports it.</p>
<p>
Own Id: OTP-7410</p>
</item>
<item>
<p>
Bug fixed for <c>{packet,http}</c> when space follows
http headers.</p>
<p>
Own Id: OTP-7458</p>
</item>
<item>
<p>
The trace option <c>running</c> could cause an emulator
crash if the current function couldn't be determined.</p>
<p>
Own Id: OTP-7484</p>
</item>
<item>
<p>
Using 16#ffffFFFF as a timeout value in receive...after
would often cause a timeout almost at once due to an
32-bit integer overflow. (Thanks to Serge Aleynikov and
Matthias Lang.)</p>
<p>
Own Id: OTP-7493</p>
</item>
<item>
<p>
For the process that an escript runs in, the
<c>trap_exit</c> process flag is now <c>false</c> instead
of <c>true</c> (as in previous releases). Scripts that
depend on the previous (counter-intuitive) behaviour
might not work. (Thanks to Bengt Kleberg.)</p>
<p>
*** POTENTIAL INCOMPATIBILITY ***</p>
<p>
Own Id: OTP-7517</p>
</item>
<item>
<p>
A bug in the <c>string:to_integer/1</c> builtin made the
emulator unstable. This is now corrected. (Thanks to Lev
Walkin.)</p>
<p>
Own Id: OTP-7526</p>
</item>
</list>
</section>
<section><title>Improvements and New Features</title>
<list>
<item>
<p>Performance for ETS intensive applications should now
be better in the SMP emulator. Also, ETS table
identifiers (as returned by <c>ets:new/2</c>) are now
spread out in wider range than before (using 28 bits in a
32-bit emulator) to make sure that the table identifier
for a deleted table will not be quickly re-used.</p>
<p>NOTE: Table identifiers can now be negative integers.
Programs that (incorrectly) assume that table identifiers
can only be positive integers might stop to work. (The
type of a table identifier is intentionally not
documented, and may change in a future release.)</p>
<p>
Own Id: OTP-7348</p>
</item>
<item>
<p>
New BIF <c>erlang:decode_packet/3</c> that extracts a
protocol packet from a binary. Similar to the socket
option <c>{packet, Type}</c>. Also documented the socket
packet type <c>http</c> and made it official.
<em>NOTE</em>: The tuple format for <c>http</c> packets
sent from an active socket has been changed in an
incompatible way.</p>
<p>
*** POTENTIAL INCOMPATIBILITY ***</p>
<p>
Own Id: OTP-7404</p>
</item>
<item>
<p>
The source code for the documentation for some
applications (erts, kernel, stdlib, and several others)
are now included in the source tar ball. There is
currently no Makefile support for building HTML files
from the source (such support will be included in a
future release).</p>
<p>
Own Id: OTP-7406</p>
</item>
<item>
<p>
A lot of frequently accessed memory counters (erts
internal) have been removed. This since they hurt
performance on the runtime system with SMP support. As a
result <c>erlang:memory/[0,1]</c> will only deliver a
result if all <c>erts_alloc(3)</c> allocators are enabled
(default). The result delivered when all
<c>erts_alloc(3)</c> allocators are enabled are both more
accurate and less accurate than before. More memory than
before are included in the result, but the different
parts that are summed are not gathered atomically. A call
to <c>erlang:memory/[0,1]</c> is much cheaper for the
system than before. This since the information isn't
gathered atomically anymore which was very expensive.</p>
<p>
*** POTENTIAL INCOMPATIBILITY ***</p>
<p>
Own Id: OTP-7468</p>
</item>
<item>
<p>
Pre-allocators used for, for example, timers, and
messages have been rewritten to be scheduler specific.
That is, different schedulers will now allocate from
different pools which reduces lock contention.</p>
<p>
Own Id: OTP-7470</p>
</item>
<item>
<p>
On Mac OS X, file:sync/1 now guarantees that all
filesystem buffers are written to the disk by using the
fcntl() with F_FULLFSYNC option. Previously, file:sync/1
called fsync(), which only guaranteed that the data had
been transferred to the disk drive. (Thanks to Jan
Lehnardt.)</p>
<p>
Own Id: OTP-7471</p>
</item>
<item>
<p>
Termination of a process that takes a long time can now
be preempted, i.e., the terminating process will be
rescheduled for later continuation of termination so that
other processes can execute. Termination of a process
that owns many and/or large ets tables typically takes a
long time.</p>
<p>
Own Id: OTP-7477</p>
</item>
<item>
<p>
A new trace option <c>exiting</c> has been added. The
<c>exiting</c> trace is similar to the <c>running</c>
trace, but for exiting processes. For more information
see the erlang(3) documentation.</p>
<p>
The <c>erlang:trace/3</c> bif now doesn't block other
scheduler threads if only one tracee is specified in the
call to <c>erlang:trace/3</c>.</p>
<p>
Own Id: OTP-7481</p>
</item>
<item>
<p>
The re module is extended with repetitive matches (global
option) and replacement function.</p>
<p>
Own Id: OTP-7494 Aux Id: OTP-7181 </p>
</item>
<item>
<p>
In the section about binary construction, the reference
manual now mentions what happens when an integer value
does not fit into an integer segment of size N (namely,
that the N least significant bits will be put into into
the binary and that the most significant bits will be
silently discarded). (Thanks to Edwin Fine.)</p>
<p>
Own Id: OTP-7506</p>
</item>
<item>
<p>
Setting the <c>{active,once}</c> for a socket (using
inets:setopts/2) is now specially optimized (because the
<c>{active,once}</c> option is typically used much more
frequently than other options).</p>
<p>
Own Id: OTP-7520</p>
</item>
</list>
</section>
<section><title>Known Bugs and Problems</title>
<list>
<item>
<p>
Floating point arithmetics in drivers can cause a runtime
system crash and/or unexpected results on runtime systems
with floating point exceptions enabled. Floating point
exceptions are disabled unless explicitly enabled or if
hipe is enabled.</p>
<p>
Own Id: OTP-7237</p>
</item>
</list>
</section>
</section>
<section><title>Erts 5.6.3.3</title>
<section><title>Fixed Bugs and Malfunctions</title>
<list>
<item>
<p>
Binary construction with an integer field of size 0 at
the end of the constructed binary (and the size given in
a variable), would cause a write of one byte outside the
memory reserved for the binary, which in turn could cause
an emulator crash.</p>
<p>
Own Id: OTP-7422</p>
</item>
<item>
<p>
A race condition in the dynamic driver implementation
could cause an emulator crash. (Thanks to Paul Fisher)</p>
<p>
Own Id: OTP-7464</p>
</item>
<item>
<p>
Calls to <c>erlang:system_info(allocated_areas)</c> could
cause the runtime system with SMP support to crash.</p>
<p>
Own Id: OTP-7474</p>
</item>
<item>
<p>
The <c>env</c> option to <c>open_port()</c> could cause
the runtime system with SMP support to crash.</p>
<p>
Own Id: OTP-7475</p>
</item>
</list>
</section>
<section><title>Improvements and New Features</title>
<list>
<item>
<p>
Operations that needed to block other threads in the
runtime system with SMP support unnecessarily waited for
async threads to block. Most important the
<c>erlang:memory/[0,1]</c> bif, code loading, and the
<c>erlang:trace/3</c> bif.</p>
<p>
Own Id: OTP-7480</p>
</item>
</list>
</section>
</section>
<section><title>Erts 5.6.3.2</title>
<section><title>Fixed Bugs and Malfunctions</title>
<list>
<item>
<p>
Calls to <c>erlang:memory/[0,1]</c> could cause the
runtime system with SMP support to crash.</p>
<p>
Own Id: OTP-7415</p>
</item>
</list>
</section>
</section>
<section><title>Erts 5.6.3.1</title>
<section><title>Fixed Bugs and Malfunctions</title>
<list>
<item>
<p>
Doing local call trace on bit syntax matching code that
has been optimized with delayed sub-binary creation could
crash the emulator.</p>
<p>
Own Id: OTP-7399 Aux Id: seq10978 </p>
</item>
</list>
</section>
</section>
<section><title>Erts 5.6.3</title>
<section><title>Fixed Bugs and Malfunctions</title>
<list>
<item>
<p>
Only one to_erl process at a time is allowed to connect
to the same run_erl pipe. Prevents buggy behaviour when
IO from several to_erl's get interleaved.</p>
<p>
Own Id: OTP-5107</p>
</item>
<item>
<p>
IPv6 name resolving has now been fixed to use
getaddrinfo() patch (thoroughly reworked) courtesy of Love
Hörnquist-Åstrand submitted by Fredrik Thulin. It also
can use gethostname2() patch (also reworked) courtesy of
Mikael Magnusson for debian submitted by Sergei Golovan.</p>
<p>
Own Id: OTP-5382</p>
</item>
<item>
<p>
Improved error handling in run_erl</p>
<p>
Own Id: OTP-7252</p>
</item>
<item>
<p>A permanent fix for the deadlock issue temporarily
fixed by OTP-7260.</p> <taglist><tag>OTP-7260</tag><item>
The runtime system with SMP support could under rare
circumstances deadlock when a distribution channel was
taken down while multiple simultaneous operations were
performed on it. </item></taglist>
<p>
Own Id: OTP-7267 Aux Id: OTP-7260 </p>
</item>
<item>
<p>
./configure has been improved to find 64-bit OpenSSL
libraries.</p>
<p>
Own Id: OTP-7270</p>
</item>
<item>
<p>
A terminating process could under very rare circumstances
trigger a bug which could crash the runtime system with
SMP support.</p>
<p>
Own Id: OTP-7272</p>
</item>
<item>
<p>
SCTP_ADDR_CONFIRMED events are now handled by gen_sctp.</p>
<p>
Own Id: OTP-7276</p>
</item>
<item>
<p>
binary_to_term/1 would crash the emulator if the binary
data contained an external fun with non-atom module
and/or function. Corrected to generate a badarg
exception.</p>
<p>
Own Id: OTP-7281</p>
</item>
<item>
<p>
On Mac OS 10.5 (Leopard), sending to socket which the
other end closes could cause the emulator to consume 100%
CPU time. (Thanks to Matthias Radestock.)</p>
<p>
Own Id: OTP-7289</p>
</item>
<item>
<p>
The vanilla driver used on Windows could crash the
emulator and sometimes produced corrupt files. The
vanilla driver is the driver that is used when one only
pass a filename as first argument to <c>open_port/2</c>.
<em>NOTE</em>: This use of <c>open_port/2</c> is
<em>obsolete</em>, and the documentation of this use has
previously been removed. The functionality is only
present for backward compatibility reasons and
<em>will</em> eventually be removed.</p>
<p>
Own Id: OTP-7301</p>
</item>
<item>
<p>
Faulty matching in binaries larger than 512Mb on 64bit
machines fixed.(On 32bit, the size limit for binaries is
still 512Mb). Thanks to Edwin Fine and Per Gustafsson for
finding fault and fix.</p>
<p>
Own Id: OTP-7309</p>
</item>
<item>
<p>
crypto:start() on Windows caused emulator to hang on
error popup window if openssl DLL was not found. Windows
error popups now suppressed.</p>
<p>
Own Id: OTP-7325</p>
</item>
<item>
<p>
Configuration option <c>without-termcap</c> can be used to
disable the use of termcap libraries for terminal cursor
control in the shell.</p>
<p>
Own Id: OTP-7338</p>
</item>
<item>
<p>
to_erl reports its terminal window size to run_erl in
order to get output formatted accordingly</p>
<p>
Own Id: OTP-7342</p>
</item>
<item>
<p>
On Solaris, the <c>compressed</c> option for file
operations did not work if the file descriptor happened
to be greater than 255 (a problem with fopen() and
friends in Solaris itself).</p>
<p>
Own Id: OTP-7343 Aux Id: seq10949 </p>
</item>
<item>
<p>
A race condition in the runtime system with SMP support
causing an erroneous removal of a newly created ets table
has been fixed.</p>
<p>
The race occurred when a process removed a table during
termination simultaneously as another process removed the
same table via <c>ets:delete/1</c> and a third process
created a table that accidentally got the same internal
table index as the table being removed.</p>
<p>
Own Id: OTP-7349</p>
</item>
<item>
<p>
<c>zlib:inflate</c> failed when the size of the inflated
data was an exact multiple of the internal buffer size
(4000 bytes by default).</p>
<p>
Own Id: OTP-7359</p>
</item>
<item>
<p>
If the total number of allowed atoms is exceeded, there
will now be a controlled termination of the emulator with
a crash dump file. The emulator used to simply crash.
(Thanks Howard Yeh and Thomas Lindgren.)</p>
<p>
Own Id: OTP-7372</p>
</item>
<item>
<p>
The break handler in werl on Windows could cause the
emulator to hang or crash, that is now corrected.</p>
<p>
Own Id: OTP-7394 Aux Id: seq10969 </p>
</item>
<item>
<p>
The configure script now tests for an serious
optimization bug in gcc-4.3.0. If the bug is present, the
configure script will abort (if this happens, the only
way to build Erlang/OTP is to change to another version
of gcc). (Thanks to Mikael Pettersson.)</p>
<p>
Own Id: OTP-7397</p>
</item>
</list>
</section>
<section><title>Improvements and New Features</title>
<list>
<item>
<p>
On Unix, the emulator now notices when the width of the
terminal has changed. (Thanks to Matthew Dempsky and
Patrick Mahoney.)</p>
<p>
Own Id: OTP-7290</p>
</item>
<item>
<p>
There is a new function <c>init:stop/1</c> which can be
used to shutdown the system cleanly AND generate a
non-zero exit status or crash dump. (Thanks to Magnus
Froberg.)</p>
<p>
Own Id: OTP-7308</p>
</item>
<item>
<p>
process_info(Pid, garbage_collection) now returns more
information</p>
<p>
Own Id: OTP-7311</p>
</item>
<item>
<p>
The <c>hide</c> option for <c>open_port/2</c> is now
documented. (Thanks to Richard Carlsson.)</p>
<p>
Own Id: OTP-7358</p>
</item>
</list>
</section>
<section><title>Known Bugs and Problems</title>
<list>
<item>
<p>
Floating point arithmetics in drivers can cause a runtime
system crash on runtime systems with floating point
exceptions enabled. Floating point exceptions are
disabled unless explicitly enabled or if hipe is enabled.</p>
<p>
Own Id: OTP-7237</p>
</item>
</list>
</section>
</section>
<section><title>Erts 5.6.2</title>
<section><title>Fixed Bugs and Malfunctions</title>
<list>
<item>
<p>
The maximum length of an atom of 255 characters is now
strictly enforced. <c>binary_to_term/1</c> will now fail
with a badarg if an encoded term contains an atom longer
than 255 characters. Atoms created by drivers will now be
truncated to 255 characters if necessary. (Thanks to
Matthew Dempsky.)</p>
<p>
Own Id: OTP-7147</p>
</item>
<item>
<p>
A bug in "bignum handling" on some 64bit architectures
could cause rem and div operations on large numbers to
hang indefinitely. Rem operations involving the smallest
negative number representable in 28 bits or 60 bits could
also cause access violation and emulator crash. Both
errors are corrected.</p>
<p>
Own Id: OTP-7177</p>
</item>
<item>
<p>
When doing the initial garbage collection after waking a
hibernated process, a fullsweep garbage collection was
unnecessarily triggered.</p>
<p>
Own Id: OTP-7212</p>
</item>
<item>
<p>The beta testing module <c>gen_sctp</c> now supports
active mode as stated in the documentation. Active mode
is still rather untested, and there are some issues about
what should be the right semantics for
<c>gen_sctp:connect/5</c>. In particular: should it be
blocking or non-blocking or choosable. There is a high
probability it will change semantics in a (near) future
patch.</p> <p>Try it, give comments and send in bug
reports!</p>
<p>
Own Id: OTP-7225</p>
</item>
<item>
<p>
Invalid arguments to <c>ets:update_counter/3</c> were not
handled correctly. A tuple position (<c>Pos</c>) less
than 1 caused the element directly following the key to
be updated (as if no position at all had been specified).
All invalid values for <c>Pos</c> will now fail with
<c>badarg</c>.</p>
<p>
Own Id: OTP-7226</p>
</item>
<item>
<p>
The runtime system with SMP support could under rare
circumstances deadlock when a distribution channel was
taken down while multiple simultaneous operations were
performed on it.</p>
<p>
Own Id: OTP-7260</p>
</item>
</list>
</section>
<section><title>Improvements and New Features</title>
<list>
<item>
<p>
More checksum/hash algorithms from the zlib library are
now available as built in functions (like md5 hashes has
been for a long time).</p>
<p>
Own Id: OTP-7128</p>
</item>
<item>
<p>
Minor improvements in the garbage collector.</p>
<p>
Own Id: OTP-7139 Aux Id: OTP-7132 </p>
</item>
<item>
<p>
The switch "-detached" to the windows werl program now
can create an erlang virtual machine without any main
window and without a temporary console showing.</p>
<p>
Own Id: OTP-7142</p>
</item>
<item>
<p><c>erlang:system_info/1</c> now accepts the
<c>logical_processors</c>, and <c>debug_compiled</c>
arguments. For more info see the, <c>erlang(3)</c>
documentation.</p> <p>The scale factor returned by
<c>test_server:timetrap_scale_factor/0</c> is now also
effected if the emulator uses a larger amount of
scheduler threads than the amount of logical processors
on the system. </p>
<p>
Own Id: OTP-7175</p>
</item>
<item>
<p>
A new BIF ets:update_element/3. To update individual
elements within an ets-tuple, without having to read,
update and write back the entire tuple.</p>
<p>
Own Id: OTP-7200</p>
</item>
<item>
<p>
A process executing the <c>processes/0</c> BIF can now be
preempted by other processes during its execution. This
in order to disturb the rest of the system as little as
possible. The returned result is, of course, still a
consistent snapshot of existing processes at a time
during the call to <c>processes/0</c>.</p>
<p>
The documentation of the <c>processes/0</c> BIF and the
<c>is_process_alive/1</c> BIF have been updated in order
to clarify the difference between an existing process and
a process that is alive.</p>
<p>
Own Id: OTP-7213</p>
</item>
</list>
</section>
</section>
<section><title>Erts 5.6.1.1</title>
<section><title>Fixed Bugs and Malfunctions</title>
<list>
<item>
<p>
Not enough parameters were passed when sending an error
report in erl_check_io.c (Thanks to Matthew Dempsky).</p>
<p>
Own Id: OTP-7176</p>
</item>
<item>
<p>
In rare circumstances, complex binary matching code could
cause the emulator to crash or not match when it should.
(Thanks to Rory Byrne.)</p>
<p>
Own Id: OTP-7198</p>
</item>
</list>
</section>
<section><title>Improvements and New Features</title>
<list>
<item>
<p>The <c>{allocator_sizes, Alloc}</c> and
<c>alloc_util_allocators</c> arguments are now accepted
by <c>erlang:system_info/1</c>. For more information see
the <c>erlang(3)</c> documentation.</p>
<p>
Own Id: OTP-7167</p>
</item>
<item>
<p>
The finishing reallocation of the heap block when
hibernating a process is now always moving the heap block
since it drastically reduces memory fragmentation when
hibernating large amounts of processes.</p>
<p>
Own Id: OTP-7187</p>
</item>
</list>
</section>
</section>
<section><title>Erts 5.6.1</title>
<section><title>Fixed Bugs and Malfunctions</title>
<list>
<item>
<p>
The SMP emulator on sparc64 erroneously used the sparc32
atomic and the sparc32 spinlock implementations which
caused it to crash.</p>
<p>
Own Id: OTP-7006</p>
</item>
<item>
<p>
Call tracing the new guard BIFs <c>byte_size</c>,
<c>bit_size</c>, or <c>tuple_size</c> and the loading a
module that uses one of those functions, could cause the
emulator to terminate.</p>
<p>
Own Id: OTP-7008</p>
</item>
<item>
<p>
configuring --enable-darwin-universal or
--enable-darwin-64bit on MacOSX could result in a non
optimized emulator. Top level configure script now
corrected.</p>
<p>
Own Id: OTP-7014</p>
</item>
<item>
<p>
configuring --with-gd did not produce correct include
flags for percept.</p>
<p>
Own Id: OTP-7015</p>
</item>
<item>
<p>
Environment variables weren't handled in thread safe
manner in the runtime system with SMP support on Windows.</p>
<p>
<c>erl_drv_putenv()</c>, and <c>erl_drv_getenv()</c> has
been introduced for use in drivers. Do <em>not</em> use
putenv(), or getenv() directly in drivers. For more
information see the <c>erl_driver</c> documentation.</p>
<p>
Own Id: OTP-7035</p>
</item>
<item>
<p>
HIPE: Corrected the choice of interface to the send/3 and
setnode/3 BIFs for native-compiled code. Using the
incorrect interface could, in unusual circumstances, lead
to random runtime errors.</p>
<p>
Own Id: OTP-7067</p>
</item>
<item>
<p>
Garbage collections could become extremely slow when
there were many keys in the process dictionary. (Thanks
to Fredrik Svahn.)</p>
<p>
Own Id: OTP-7068</p>
</item>
<item>
<p>
The duplicate documentation directory in the windows
installation is removed.</p>
<p>
Own Id: OTP-7070</p>
</item>
<item>
<p>Documentation bugfixes and clarifications.</p> (Thanks
to Joern ([email protected]), Matthias Lang, and Richard
Carlsson.)
<p>
Own Id: OTP-7079</p>
</item>
<item>
<p>
The runtime system with SMP support <em>not</em> using
the native atomic integer implementation part of OTP
could deadlock when run on a system with more than one
logical processor. That is, only the runtime system with
SMP support on <em>other</em> hardware platforms than
x86, x86_64, sparc32, and powerpc32 were effected by this
bug.</p>
<p>
Own Id: OTP-7080</p>
</item>
<item>
<p>
The break handling code (run when Ctrl-C is hit) could
could potentially deadlock the runtime system with SMP
support.</p>
<p>
Own Id: OTP-7104</p>
</item>
<item>
<p>
The sctp driver has been updated to work against newer
lksctp packages e.g 1.0.7 that uses the API spelling
change adaption -> adaptation. Older lksctp (1.0.6) still
work. The erlang API in gen_sctp.erl and inet_sctp.hrl
now spells 'adaptation' regardless of the underlying C
API.</p>
<p>
*** POTENTIAL INCOMPATIBILITY ***</p>
<p>
Own Id: OTP-7120</p>
</item>
<item>
<p>A bug in <c>erlang:phash2/1</c> on 64-bit platforms
has been fixed. (Thanks to Scott Lystig Fritchie.)</p>
<p>
Own Id: OTP-7127</p>
</item>
<item>
<p>
The emulator could under rare circumstances crash while
garbage collecting.</p>
<p>
Own Id: OTP-7132</p>
</item>
</list>
</section>
<section><title>Improvements and New Features</title>
<list>
<item>
<p>The documentation has been updated so as to reflect
the last updates of the Erlang shell as well as the minor
modifications of the control sequence <c>p</c> of the
<c>io_lib</c> module.</p> <p>Superfluous empty lines have
been removed from code examples and from Erlang shell
examples.</p>
<p>
Own Id: OTP-6944 Aux Id: OTP-6554, OTP-6911 </p>
</item>
<item>
<p>
Bit syntax construction with a small integer in a
non-byte aligned field wider than the CPU's word size
could cause garbage bits in the beginning of the field.</p>
<p>
Own Id: OTP-7085</p>
</item>
<item>
<p>
All Windows versions older than Windows 2000 are now
<em>not supported</em> by the Erlang runtime system. This
since there was a need for usage of features introduced
in Windows 2000.</p>
<p>
Own Id: OTP-7086</p>
</item>
<item>
<p>Memory management improvements especially for the
runtime system with SMP support:</p> <list> <item> The
runtime system with SMP support can now use multiple,
thread specific instances of most memory allocators. This
improves performance since it reduces lock contention in
the memory allocators. It may however increase memory
usage for some applications. The runtime system with SMP
support will by default enable this feature on most
allocators. The amount of instances used can be
configured. </item> <item> <c>driver_alloc()</c>,
<c>driver_realloc()</c>, and <c>driver_free()</c> now use
their own erts specific memory allocator instead of the
default <c>malloc()</c> implementation on the system.
</item> <item> The default configuration of some
allocators have been changed to fit applications that use
much memory better. </item> <item> Some new
<c>erts_alloc</c> configuration parameters have been
added. </item> <item> <c>erts_alloc_config</c> has been
modified to be able to create configurations suitable for
multiple instances of allocators. </item> <item> The
returned value from <c>erlang:system_info({allocator,
Alloc})</c> has been changed. This since an allocator may
now run in multiple instances. </item> </list> <p>If you
for some reason want the memory allocators to be
configured as before, you can pass the <c>+Mea r11b</c>
command-line argument to <c>erl</c>.</p> <p>For more
information see the <c>erts_alloc(3)</c>, the
<c>erts_alloc_config(3)</c>, and the <c>erlang(3)</c>
documentation.</p>
<p>
*** POTENTIAL INCOMPATIBILITY ***</p>
<p>
Own Id: OTP-7100</p>
</item>
<item>
<p>
On Unix, denormalized floating point numbers could not be
created using <c>list_to_float/1</c> or
<c>binary_to_term/1</c>. (Thanks to Matthew Dempsky.)</p>
<p>
Own Id: OTP-7122</p>
</item>
<item>
<p>
Native atomic integers and spin-locks are now also
available for the runtime system with SMP support on
sparc64.</p>
<p>
Own Id: OTP-7130</p>
</item>
<item>
<p>
FP exceptions support for sparc64 userspace on Linux has
been added. Note that FP exception support is now turned
off by default, so to actually enable it you need to do
'<c>./configure --enable-fp-exceptions</c>'.</p>
<p>
Own Id: OTP-7131</p>
</item>
</list>
</section>
</section>
<section><title>Erts 5.6</title>
<section><title>Fixed Bugs and Malfunctions</title>
<list>
<item>
<p>
A bug for raw files when reading 0 bytes returning 'eof'
instead of empty data has been corrected.</p>
<p>
Own Id: OTP-6291 Aux Id: OTP-6967 </p>
</item>
<item>
<p>
All exported functions in gzio.c have now been renamed to
avoid conflict with drivers that are indirectly linked
with an external zlib library.</p>
<p>
Own Id: OTP-6816 Aux Id: OTP-6591 </p>
</item>
<item>
<p>
On the 64-bit Erlang emulator, bit syntax construction
with integers containing more than 60 bits ("big
numbers") into fields with more than 60 bits could
produce incorrect results.</p>
<p>
Own Id: OTP-6833</p>
</item>
<item>
<p>
When the runtime system failed to allocate memory for
binaries, it could dead lock while writing the
<c>erl_crash.dump</c>.</p>
<p>
Own Id: OTP-6848</p>
</item>
<item>
<p>
The runtime system with SMP support could deadlock if a
process called the <c>erlang:suspend_process(Pid)</c> BIF
or the <c>erlang:garbage_collect(Pid)</c> BIF while the
process identified by <c>Pid</c> was currently running
and the process calling the BIFs was terminated during
the call to the BIFs.</p>
<p>
Processes suspending each other via the
<c>erlang:suspend_process/1</c> BIF or garbage collecting
each other via the <c>erlang:garbage_collect/1</c> BIF
could deadlock each other when the runtime system with
SMP support was used.</p>
<p>
Own Id: OTP-6920</p>
</item>
<item>
<p>
<c>dbg</c> could leave traced processes in a suspended
state if the tracer process was killed with exit reason
<c>kill</c>.</p>
<p>
<c>erlang:suspend_process/2</c> has been introduced which
accepts an option list as second argument. For more
information see the <c>erlang(3)</c> documentation.</p>
<p>
Processes suspended via
<c>erlang:suspend_process/[1,2]</c> will now be
automatically resumed if the process that called
<c>erlang:suspend_process/[1,2]</c> terminates.</p>
<p>
Processes could previously be suspended by one process
and resumed by another unless someone was tracing the
suspendee. This is <em>not</em> possible anymore. The
process resuming a process <em>has</em> to be the one
that suspended it.</p>
<p>
*** POTENTIAL INCOMPATIBILITY ***</p>
<p>
Own Id: OTP-6946</p>
</item>
<item>
<p>file:write_file/3, file:write/2 and file:read/2 could
crash (contrary to documentation) for odd enough file
system problems, e.g write to full file system. This bug
has now been corrected.</p> <p>In this process the file
module has been rewritten to produce better error codes.
Posix error codes now originate from the OS file system
calls or are generated only for very similar causes (for
example 'enomem' is generated if a memory allocation
fails, and 'einval' is generated if the file handle in
Erlang is a file handle but currently invalid).</p>
<p>More Erlang-ish error codes are now generated. For
example <c>{error,badarg}</c> is now returned from
<c>file:close/1</c> if the argument is not of a file
handle type. See file(3).</p> <p>The possibility to write
a single byte using <c>file:write/2</c> instead of a list
or binary of one byte, contradictory to the
documentation, has been removed.</p>
<p>
*** POTENTIAL INCOMPATIBILITY ***</p>
<p>
Own Id: OTP-6967 Aux Id: OTP-6597 OTP-6291 </p>
</item>
<item>
<p>
Monitor messages produced by the system monitor
functionality, and garbage collect trace messages could
contain erroneous heap and/or stack sizes when the actual
heaps and/or stacks were huge.</p>
<p>
As of erts version 5.6 the <c>large_heap</c> option to
<c>erlang:system_monitor/[1,2]</c> has been modified. The
monitor message is sent if the sum of the sizes of all
memory blocks allocated for all heap generations is equal
to or larger than the specified size. Previously the
monitor message was sent if the memory block allocated
for the youngest generation was equal to or larger than
the specified size.</p>
<p>
*** POTENTIAL INCOMPATIBILITY ***</p>
<p>
Own Id: OTP-6974 Aux Id: seq10796 </p>
</item>
<item>
<p>
<c>inet:getopts/2</c> returned random values on Windows
Vista.</p>
<p>
Own Id: OTP-7003</p>
</item>
</list>
</section>
<section><title>Improvements and New Features</title>
<list>
<item>
<p>
The emulator internal process lock implementation has
been rewritten and optimized. A slight risk of starvation
existed in the previous implementation. This risk has
also been eliminated in the new implementation.</p>
<p>
Own Id: OTP-6500</p>
</item>
<item>
<p>
Bitstrings (bit-level) binaries and binary comprehensions
are now part of the language. See the Reference Manual.</p>
<p>
Own Id: OTP-6558</p>
</item>
<item>
<p>
The windows version of erlang now has SMP support. The
SMP emulator is run by default on machines which shows
more than one virtual or physical processor.</p>
<p>
Own Id: OTP-6560 Aux Id: OTP-6925 </p>
</item>
<item>
<p>
The details of the compressed term format has been
documented in erl_ext_dist.txt. (Thanks to Daniel
Goertzen.)</p>
<p>
Own Id: OTP-6755</p>
</item>
<item>
<p>
The runtime system with SMP support is now started by
default if more than one logical processor are detected.
For more information, see the <c>erl(3)</c>
documentation.</p>
<p>
Own Id: OTP-6756</p>
</item>
<item>
<p>
The external format for Erlang terms and the distribution
protocol are now documented in ERTS User's Guide.</p>
<p>
Own Id: OTP-6779</p>
</item>
<item>
<p>
New BIF's erlang:system_profile/1 and
erlang:system_profile/2. These BIF's controls concurrency
profiling options for processes, ports and schedulers.</p>
<p>
Own Id: OTP-6783 Aux Id: OTP-6285 </p>
</item>
<item>
<p>
The <c>ErlDrvTermData</c> term types used by
<c>driver_output_term()</c> and <c>driver_send_term()</c>
have been extended with the term types
<c>ERL_DRV_BUF2BINARY</c>, <c>ERL_DRV_EXT2TERM</c>, and
<c>ERL_DRV_UINT</c>. <c>ERL_DRV_BUF2BINARY</c> is used
for passing and creating a binary,
<c>ERL_DRV_EXT2TERM</c> is used for passing terms encoded
with the external term format, and <c>ERL_DRV_UINT</c> is
used for passing unsigned integers.</p>
<p>
Also the data types <c>ErlDrvUInt</c> and
<c>ErlDrvSInt</c> have been added which makes it more
obvious how arguments to term types are interpreted with
regards to width and signedness.</p>
<p>
The incorrect data types <c>ErlDriverTerm</c>,
<c>ErlDriverBinary</c>, and <c>ErlDriverPort</c> in the
<c>erl_driver(3)</c> documentation have been replaced
with the correct data types <c>ErlDrvTermData</c>,
<c>ErlDrvBinary</c>, and <c>ErlDrvPort</c>.</p>
<p>
For more information see the <c>erl_driver(3)</c>
documentation.</p>
<p>
Own Id: OTP-6823</p>
</item>
<item>
<p>
Miscellaneous improvements of the erts internal thread
library.</p>
<p>
It now support optimized atomic operations and spin-locks
on windows.</p>
<p>
Fall-backs based on mutexes and/or spin-locks for missing
optimized atomic operations, spin-locks, or rwlocks has
been implemented. This makes it possible to compile the
runtime system with SMP support on a lot more platforms.</p>
<p>
Default stack size on OpenBSD has been increased to 256
kilo-words.</p>
<p>
Own Id: OTP-6831 Aux Id: OTP-6560 </p>
</item>
<item>
<p>Many bit syntax operations, both construction and
matching, are faster. For further information, see the
Efficiency Guide.</p>
<p>
Own Id: OTP-6838</p>
</item>
<item>
<p>Literal lists, tuples, and binaries are no longer
constructed at run-time as they used to be, but are
stored in a per-module constant pool. Literals that are
used more than once are stored only once.</p>
<p>This is not a change to the language, only in the
details of its implementation. Therefore, the
implications of this change is described in the
Efficiency Guide.</p>
<p>Example 1: In the expression <c>element(BitNum-1,
{1,2,4,8,16,32,64,128})</c>, the tuple used to be
constructed every time the expression was executed, which
could be detrimental to performance in two ways if the
expression was executed in a loop: the time to build the
tuple itself and the time spent in garbage collections
because the heap filled up with garbage faster.</p>
<p>Example 2: Literal strings, such as <c>"abc"</c>, used
to be stored in the compiled code compactly as a byte
string and expanded to a list at run-time. Now all
strings will be stored expanded to lists (such as
<c>[$a,$b,$c]</c>) in the constant pool. That means that
the string will be faster to use at run-time, but that it
will require more space even when not used. If space is
an issue, you might want to use binary literals (that is,
<c><<"abc">></c>) instead of string literals for
infrequently used long strings (such as error
messages).</p>
<p>
Own Id: OTP-6850</p>
</item>
<item>
<p>The Erlang driver API has been extended with a
portable POSIX thread like API for multi-threading. The
Erlang driver thread API provides:</p> <list>
<item>Threads</item> <item>Mutexes</item> <item>Condition
variables</item> <item>Read/Write locks</item>
<item>Thread specific data</item> </list> <p>For more
information see the <c>erl_driver(3)</c>
documentation.</p>
<p>
Own Id: OTP-6857</p>
</item>
<item>
<p>
Recursive calls now usually consume less stack than in
R11B. See the Efficiency Guide.</p>
<p>
Own Id: OTP-6862 Aux Id: seq10746 </p>
</item>
<item>
<p>
The deprecated BIFs <c>erlang:old_binary_to_term/1</c>
and <c>erlang:info/1</c> have been removed.</p>
<p>
*** POTENTIAL INCOMPATIBILITY ***</p>
<p>
Own Id: OTP-6876</p>
</item>
<item>
<p>
Calls to driver call-backs triggered by external events
are now scheduled and interleaved with execution of
Erlang processes also on the runtime system without SMP
support.</p>
<p>
Own Id: OTP-6878</p>
</item>
<item>
<p>
Faster arithmetic of integers of more than 27 bits signed
(or 60 bits signed on an 64-bit CPU), and also faster
integer multiplication. (Thanks to Tony Rogvall.)</p>
<p>
Own Id: OTP-6891</p>
</item>
<item>
<p>Significant improvements of the <c>process_info</c>
BIFs:</p> <list> <item> <c>process_info/2</c> can now be
called with a list of items as second argument in order
to atomically retrieve information about multiple items.
</item> <item> <c>process_info/[1,2]</c> has been
optimized in the runtime system with SMP support. The
whole scheduler could previously be blocked for a
significant period of time in <c>process_info/[1,2]</c>
waiting for a lock on the process being inspected. The
Erlang process calling <c>process_info/[1,2]</c> can
still be blocked for a significant period of time waiting
for the lock, but the scheduler will now be able to run
other processes while the process calling
<c>process_info/[1,2]</c> waits for the lock. </item>
<item> <c>process_info/2</c> now accept a few more items
than before. </item> <item> The documentation of
<c>process_info/[1,2]</c> has been improved. </item>
</list> <p>For more information see the <c>erlang(3)</c>
documentation.</p>
<p>
Own Id: OTP-6899</p>
</item>
<item>
<p>
<c>open_port({}, [])</c> could crash the emulator.
(Thanks to Matthew Dempsky.)</p>
<p>
Own Id: OTP-6901</p>
</item>
<item>
<p>Two new guard BIFs have been introduced as a
recommended replacement for <c>size/1</c>. (The
<c>size/1</c> BIF will be removed no earlier than in
R14B.) The BIFs are <c>tuple_size/1</c> to calculate the
size of a tuple and <c>byte_size/1</c> to calculate the
number of bytes needed for the contents of the binary or
bitstring (rounded up to the nearest number of bytes if
necessary).</p>
<p>There is also a new <c>bit_size/1</c> BIF that returns
the exact number of bits that a binary or bitstring
contains.</p>
<p>
Own Id: OTP-6902</p>
</item>
<item>
<p>The <c>ets:fixtable/2</c> function, which has been
deprecated for several releases, has been removed.</p>
<p>The <c>ets:info/1</c> function has been reimplemented
as a BIF, which guarantees that information returned is
consistent.</p>
<p>The <c>ets:info/2</c> function now fails with reason
<c>badarg</c> if the second argument is invalid.
(Dialyzer can be used to find buggy code where the second
argument is misspelled.)</p>
<p>
*** POTENTIAL INCOMPATIBILITY ***</p>
<p>
Own Id: OTP-6906</p>
</item>
<item>
<p>
As the linux kernel may generate a minor fault when
tracing with CPU timestamps, and there exists no patch to
the Linux kernel that fixes the problem, cpu timestamps
are disabled on Linux for now.</p>
<p>
*** POTENTIAL INCOMPATIBILITY ***</p>
<p>
Own Id: OTP-6922</p>
</item>
<item>
<p>The functions io:columns/0, io:columns/1, io:rows/0
and io:rows/1 are added to allow the user to get
information about the terminal geometry. The shell takes
some advantage of this when formatting output. For
regular files and other io-devices where height and width
are not applicable, the functions return
{error,enotsup}.</p>
<p>Potential incompatibility: If one has written a custom
io-handler, the handler has to either return an error or
take care of io-requests regarding terminal height and
width. Usually that is no problem as io-handlers, as a
rule of thumb, should give an error reply when receiving
unknown io-requests, instead of crashing.</p>
<p>
*** POTENTIAL INCOMPATIBILITY ***</p>
<p>
Own Id: OTP-6933</p>
</item>
<item>
<p>
<c>driver_caller()</c> can now also be used from the
<c>start</c> callback of a driver.</p>
<p>
Own Id: OTP-6951</p>
</item>
<item>
<p>
The emulator can now be compiled for 64bit intel, as well
as a 32bit universal binary on darwin/MacOSX 10.4 and
10.5.</p>
<p>
Own Id: OTP-6961</p>
</item>
<item>
<p>
If <c>open_port</c> fails because all available ports are
already in use, it will now throw a <c>system_limit</c>
exception instead of an <c>enfile</c> exception.
(<c>enfile</c> might still be thrown if the operating
system would return ENFILE.)</p>
<p>
Own Id: OTP-6968</p>
</item>
<item>
<p>
The <c>spawn_monitor/1</c> and <c>spawn_monitor/3</c> BIFs
are now auto-imported (i.e. they no longer need an
<c>erlang:</c> prefix).</p>
<p>
Own Id: OTP-6975</p>
</item>
<item>
<p>
On Windows, the werl window now handles resize, so that
the whole window can be utilized. Text selection is also
updated to be line oriented instead of rectangle oriented
as in earlier versions.</p>
<p>
Own Id: OTP-6994 Aux Id: OTP-6933 </p>
</item>
<item>
<p>
Kqueue support (kernel-poll) has been enabled on FreeBSD.
The problem with kqueue not detecting writes on a pipe on
FreeBSD was actually not a kqueue issue, but a writev on
pipes issue. Neither poll(), nor select() detected the
write when the bug hit. NetBSD and DragonFlyBSD probably
have or have had the same bug. This bug should have been
fixed in FreeBSD 6.3 and FreeBSD 7.0 thanks to
Jean-Sebastien Pedron.</p>
<p>
Own Id: OTP-7001</p>
</item>
</list>
</section>
</section>
<section><title>Erts 5.5.5.5</title>
<section><title>Fixed Bugs and Malfunctions</title>
<list>
<item>
<p>
Hanging writes on temporarily unavailable NFS
filesystems could cause the execution of (not file
related) erlang code to get blocked even though I/O
threads were used. This is now corrected.</p>
<p>
Own Id: OTP-6907 Aux Id: seq10771 </p>
</item>
</list>
</section>
</section>
<section><title>Erts 5.5.5.4</title>
<section><title>Fixed Bugs and Malfunctions</title>
<list>
<item>
<p>
Data passed to a driver via <c>erlang:port_call</c> could
be corrupted when the runtime system with SMP support was
used. (Thanks to YAMASHINA Hio.)</p>
<p>
Own Id: OTP-6879</p>
</item>
<item>
<p>
In the SMP emulator, if several processes called
ets:update_counter/3 (even for different tables) when the
counter values exceeded 27 bits, the counter values could
be corrupted or the emulator could crash.</p>
<p>
Own Id: OTP-6880 Aux Id: seq10760 </p>
</item>
</list>
</section>
</section>
<section><title>Erts 5.5.5.3</title>
<section><title>Fixed Bugs and Malfunctions</title>
<list>
<item>
<p>Creating a named table using <c>ets:new/2</c> or
renaming a named table using <c>ets:rename/2</c> could in
rare circumstances succeed, meaning that there would be
two or more tables with the same name. Now such call will
fail with a <c>badarg</c> exception as it is supposed to
do.</p>
<p><c>ets:delete/1</c> used on a named table now removes
the name immediately so that a new table with the same
name can be created.</p>
<p>Turning on call trace on the internal BIF that
implements <c>ets:delete/1</c> would crash the
emulator.</p>
<p>SMP emulator only: Using <c>ets:rename/2</c> on a
table that <c>ets:safe_fixtable/2</c> has been used on
could cause an emulator crash or undefined behaviour
because of a missing lock.</p>
<p>
Own Id: OTP-6872 Aux Id: seq10756, seq10757 </p>
</item>
</list>
</section>
</section>
<section><title>Erts 5.5.5.2</title>
<section><title>Known Bugs and Problems</title>
<list>
<item>
<p>
ets:select/3 on ordered_set and with a chunksize a
multiple of 1000 gave all elements instead of just 1000.
Also ets:slot/2 on ordered set could give unexpected
results on SMP emulator. Both problems are corrected.</p>
<p>
Own Id: OTP-6842</p>
</item>
</list>
</section>
</section>
<section><title>Erts 5.5.5.1</title>
<section><title>Fixed Bugs and Malfunctions</title>
<list>
<item>
<p>
All exported functions in gzio.c have now been renamed to
avoid conflict with drivers that are indirectly linked
with an external zlib library.</p>
<p>
Own Id: OTP-6816 Aux Id: OTP-6591 </p>
</item>
<item>
<p>
Calling binary_to_term/1 with certain invalid binaries
would crash the emulator.</p>
<p>
Own Id: OTP-6817</p>
</item>
</list>
</section>
<section><title>Improvements and New Features</title>
<list>
<item>
<p>
Restored speed of bit-syntax matching of 32 bits
integers.</p>
<p>
Own Id: OTP-6789 Aux Id: seq10688 </p>
</item>
</list>
</section>
</section>
<section>
<title>Erts 5.5.5</title>
<section>
<title>Fixed Bugs and Malfunctions</title>
<list type="bulleted">
<item>
<p>The functions in gzio.c have been renamed to avoid
conflict with drivers that indirectly linked with an
external zlib library.</p>
<p>Own Id: OTP-6591</p>
</item>
<item>
<p>The emulator without SMP support dumped core if an
async-thread finished a job after the emulator had begun
writing an <c><![CDATA[erl_crash.dump]]></c>. </p>
<p>Own Id: OTP-6685</p>
</item>
<item>
<p>In bit syntax matching, integer fields with greater
size than 16Mb would fail to match. (Thanks to Bertil
Karlsson and Francesco Pierfederici.)</p>
<p>Matching out a 32-bit integer not aligned on a byte
boundary from a binary could cause an heap overflow (and
subsequent termination of the emulator).</p>
<p>A module that contained bit syntax construction with a
literal field size greater than 16Mb would fail to
load.</p>
<p>Several other similar bugs having to do with huge
field sizes were eliminated.</p>
<p>Attempting to construct a binary longer than 536870911
bytes will now fail with a <c><![CDATA[system_limit]]></c> exception
(rather than fail in mysterious ways or construct an
binary with incorrect contents). Similarily, attempting
to match a binary longer than 536870911 bytes will now
fail (instead of producing an incorrect result). This
limitation has been documented in the Efficiency Guide.
(The limit is in the 32-bit emulator; use the 64-bit
emulator if you need to handle larger binaries than
536870911.)</p>
<p>Own Id: OTP-6686</p>
</item>
<item>
<p>Bugs in rem and div of very large numbers are corrected.</p>
<p>Own Id: OTP-6692</p>
</item>
<item>
<p><c><![CDATA[erlang:system_info({allocator, Alloc})]]></c> didn't
allocate enough heap when a bignum was part of the result
which could cause an emulator crash.</p>
<p>Own Id: OTP-6693</p>
</item>
<item>
<p>It was previously not possible to pass
<c><![CDATA[erts_alloc]]></c> the same configuration via the
command-line, as used by default. </p>
<p>A <c><![CDATA[+M*]]></c> command-line argument that configure a
size of some sort can now be passed a value that equals
the size of the address space. The value used, in this
case, will be <c><![CDATA["the size of the address space" - 1]]></c>.</p>
<p>Own Id: OTP-6699</p>
</item>
<item>
<p><c><![CDATA[SysIOVec* driver_peekq(ErlDrvPort port, int *vlen)]]></c> did not update <c><![CDATA[*vlen]]></c> if <c><![CDATA[port]]></c> was
invalid. <c><![CDATA[*vlen]]></c> is now set to <c><![CDATA[-1]]></c> if the
<c><![CDATA[port]]></c> is invalid.</p>
<p>The <c><![CDATA[efile]]></c> driver
expects <c><![CDATA[*vlen]]></c> to be updated also when the
<c><![CDATA[port]]></c> is invalid. This situation occurs seldom, but
when the runtime system has async-threads enabled and
ports are killed it can. When it occurred the runtime
system crashed.</p>
<p>Own Id: OTP-6729</p>
</item>
</list>
</section>
<section>
<title>Improvements and New Features</title>
<list type="bulleted">
<item>
<p>For scripts written using <c><![CDATA[escript]]></c>, there is a new
function <c><![CDATA[escript:script_name/0]]></c>, which can be used
to retrieve the pathame of the script. The documentation
has been clarified regarding pre-defined macros such as
?MODULE and the module name.</p>
<p>Own Id: OTP-6593</p>
</item>
<item>
<p>The section Guards in the chapter The Abstract Format
of the ERTS User's Guide has been updated.</p>
<p>Own Id: OTP-6600</p>
</item>
<item>
<p>Corrected protocol layer flue for socket options
SO_LINGER, SO_SNDBUF and SO_RCVBUF, for SCTP.</p>
<p>Own Id: OTP-6625 Aux Id: OTP-6336 </p>
</item>
<item>
<p>The behaviour of the inet option {active,once} on peer
close is improved and documented.</p>
<p>Own Id: OTP-6681</p>
</item>
<item>
<p>The inet option send_timeout for connection oriented
sockets is added to allow for timeouts in communicating
send requests to the underlying TCP stack.</p>
<p>Own Id: OTP-6684 Aux Id: seq10637 OTP-6681 </p>
</item>
<item>
<p>The command line flag <c><![CDATA[-args_file FileName]]></c>, and
the environment variables <c><![CDATA[ERL_AFLAGS]]></c>, and
<c><![CDATA[ERL_ZFLAGS]]></c> for the <c><![CDATA[erl]]></c> command have been
added. For more information see the <c><![CDATA[erl(1)]]></c>
documentation.</p>
<p>Own Id: OTP-6697</p>
</item>
<item>
<p>The <c><![CDATA[is_constant/1]]></c> type test has been deprecated.
<c><![CDATA[is_constant/1]]></c> is improperly named and almost
entirely undocumented.</p>
<p>Own Id: OTP-6731</p>
</item>
</list>
</section>
</section>
<section>
<title>Erts 5.5.4.3</title>
<section>
<title>Fixed Bugs and Malfunctions</title>
<list type="bulleted">
<item>
<p><c><![CDATA[process_flag(trap_exit, Bad)]]></c> where <c><![CDATA[Bad]]></c>
was a term not equal to <c><![CDATA[true]]></c> or <c><![CDATA[false]]></c>,
didn't fail with <c><![CDATA[badarg]]></c> as it should; instead, the
failure was silently ignored. This bug was introduced in
<c><![CDATA[erts-5.5.2]]></c>.</p>
<p>Own Id: OTP-6627 Aux Id: OTP-6160 </p>
</item>
<item>
<p>The minimum and default stack size for async-threads has
been increased to 16 kilowords. This since the previous
minimum and default stack size of 8 kilowords proved to
be too small (introduced in <c><![CDATA[erts-5.5.4.2]]></c>).</p>
<p>Own Id: OTP-6628 Aux Id: OTP-6580, Seq10633 </p>
</item>
</list>
</section>
<section>
<title>Improvements and New Features</title>
<list type="bulleted">
<item>
<p>process_flag/2 accepts the new flag <c><![CDATA[sensitive]]></c>.</p>
<p>Own Id: OTP-6592 Aux Id: seq10555 </p>
</item>
</list>
</section>
</section>
<section>
<title>Erts 5.5.4.2</title>
<section>
<title>Fixed Bugs and Malfunctions</title>
<list type="bulleted">
<item>
<p>When a port steals control over a file descriptor from
another port, the stealing port tests if the other port
is alive. This in order to be able to give an accurate
error message. In the runtime system with SMP support,
this test was done without appropriate locks held. This
could in worst case lead to an erroneous error message;
therefore, this bug is to be considered harmless.</p>
<p>Own Id: OTP-6602</p>
</item>
</list>
</section>
<section>
<title>Improvements and New Features</title>
<list type="bulleted">
<item>
<p>The default stack size for threads in the async-thread
pool has been shrunk to 8 kilowords, i.e., 32 KB on
32-bit architectures. This small default size has been
chosen since the amount of async-threads might be quite
large. The default stack size is enough for drivers
delivered with Erlang/OTP, but might not be sufficiently
large for other dynamically linked in drivers that use
the <c><![CDATA[driver_async()]]></c> functionality. A suggested
stack size for threads in the async-thread pool can be
configured via the <c><![CDATA[+a]]></c> command line argument of
<c><![CDATA[erl(1)]]></c>.</p>
<p>*** POTENTIAL INCOMPATIBILITY ***</p>
<p>Own Id: OTP-6580</p>
</item>
</list>
</section>
</section>
<section>
<title>Erts 5.5.4.1</title>
<section>
<title>Fixed Bugs and Malfunctions</title>
<list type="bulleted">
<item>
<p>Setting the time on the system while using heart on a
linux machine where the emulator was built with
clock_gettime support (default from Linux 2.6/erts-5.5.4
and upwards), could make the heart command fire. This was
due to bug in the heart executable which is now
corrected.</p>
<p>Own Id: OTP-6598 Aux Id: seq10614 </p>
</item>
</list>
</section>
</section>
<section>
<title>Erts 5.5.4</title>
<section>
<title>Fixed Bugs and Malfunctions</title>
<list type="bulleted">
<item>
<p>Corrected misspelling of '<c><![CDATA[-pz]]></c>' in the help text
for <c><![CDATA[erlc]]></c>. (Thanks to Ulf Wiger.)</p>
<p>Own Id: OTP-6433</p>
</item>
<item>
<p>The MD5 calculation of a BEAM file done by
<c><![CDATA[code:module_md5/1]]></c>, <c><![CDATA[beam_lib:md5/1]]></c>, and by
the compiler for the default value of the <c><![CDATA[vsn]]></c>
attribute have all been changed so that its result will
be the same on all platforms; modules containing funs
could get different MD5s on different platforms.</p>
<p>Own Id: OTP-6459</p>
</item>
<item>
<p>The emulator could dump core while writing an
<c><![CDATA[erl_crash.dump]]></c> file if there were funs with a
large terms in its environment. Since there is no way to
inspect a fun's environment in the crashdump_viewer
application anyway, a variables in the environment are
now set to [] before dumping the fun. (Thanks to
Jean-Sebastien Pedron.)</p>
<p>Own Id: OTP-6504</p>
</item>
<item>
<p><c><![CDATA[{Port, {exit_status, S}}]]></c> messages from ports
opened with the <c><![CDATA[exit_status]]></c> option could under
rare circumstances be delayed. This bug was present on
Erlang runtime systems without SMP support on all unix
operating systems other than SunOS.</p>
<p>Own Id: OTP-6528</p>
</item>
<item>
<p>A bug in linuxthreads could cause the emulator to dump
core when dlerror() was called before the first call to
dlopen(). As a workaround the emulator always makes a
call to dlopen() on initialization when linuxthreads is
used as thread library.</p>
<p>Own Id: OTP-6530</p>
</item>
<item>
<p>file:sync/1 did not do anything on Windows. Now it calls
the system function for flushing buffers
(FlushFileBuffers()). (Thanks to Matthew Sackman.)</p>
<p>Own Id: OTP-6531</p>
</item>
<item>
<p><c><![CDATA[open_port/2]]></c> could on the runtime system with SMP
support fail with the wrong exit reason when a port
couldn't be created. When this happened the exit reason
was typically <c><![CDATA[eintr]]></c>, or <c><![CDATA[ebusy]]></c> instead of
<c><![CDATA[eagain]]></c>.</p>
<p>Own Id: OTP-6536</p>
</item>
<item>
<p>The file driver (efile_drv) did not flush data written
using the option 'delayed_write' after the set timeout
time, rather at the next file operation. This bug has now
been corrected.</p>
<p>Own Id: OTP-6538</p>
</item>
</list>
</section>
<section>
<title>Improvements and New Features</title>
<list type="bulleted">
<item>
<p>An interface towards the SCTP Socket API Extensions
has been implemented.It is an Open Source patch courtesy
of Serge Aleynikov and Leonid Timochouk. The Erlang code
parts has been adapted by the OTP team, changing the
Erlang API somewhat.</p>
<p>The Erlang interface consists of the module
<c><![CDATA[gen_sctp]]></c> and an include file
<c><![CDATA[-include_lib("kernel/include/inet_sctp.hrl").]]></c> for
option record definitions. The <c><![CDATA[gen_sctp]]></c> module is
documented.</p>
<p>The delivered Open Source patch, before the OTP team
rewrites, was written according to
http://tools.ietf.org/html/draft-ietf-tsvwg-sctpsocket-13
and was claimed to work fine, tested on Linux Fedora Core
5.0 (kernel 2.6.15-2054 or later) and on Solaris 10 and
11. The OTP team rewrites used the same standard document
but might have accidentally broken some functionality. If
so it will soon be patched to working state. The tricky
parts in C and the general design has essentially not
changed. During the rewrites the code was hand tested on
SuSE Linux Enterprise Server 10, and briefly on Solaris
10. Feedbach on code and docs is very much
appreciated.</p>
<p>The SCTP interface is in beta state. It has only been
hand tested and has no automatic test suites in OPT
meaning everything is most certainly not tested. Socket
active mode is broken. IPv6 is not tested. The
documentation has been reworked due to the API changes,
but has not been proofread after this.</p>
<p>Thank you from the OTP team to Serge Aleynikov and
Leonid Timochouk for a valuable contribution. We hope we
have not messed it up too much.</p>
<p>Own Id: OTP-6336</p>
</item>
<item>
<p>A <c><![CDATA[{minor_version,Version}]]></c> option is now recognized
by <c><![CDATA[term_to_binary/2]]></c>. {minor_version,1} will cause
floats to be encoded in an exact and more space-efficient
way compared to the previous encoding.</p>
<p>Own Id: OTP-6434</p>
</item>
<item>
<p>There is a new <c><![CDATA[escript]]></c> program that can be used
for writing scripts in Erlang. Erlang scripts don't need
to be compiled and any arguments can be passed to them
without risk that they are interpreted by the Erlang
system.</p>
<p>Own Id: OTP-6505</p>
</item>
<item>
<p>Setting and getting socket options in a "raw" fashion is
now allowed. Using this feature will inevitably produce
non portable code, but will allow setting ang getting
arbitrary uncommon options on TCP stacks that do have
them.</p>
<p>Own Id: OTP-6519</p>
</item>
<item>
<p>Miscellaneous signal handling improvements on the Erlang
runtime system with SMP support.</p>
<p>The fallback implementation of spin locks and atomic
operations are now implemented using pthread spin locks
when pthread spin locks are found on the system.</p>
<p>The Erlang runtime system with SMP support can now run on
Linux systems that has Linuxthreads instead of NPTL
(Native POSIX Thread Library). Note that the SMP support
hasn't been as thoroughly tested with Linuxthreads as
with NPTL. A runtime system with SMP support will
therefore not be built by default on Linux when NPTL
isn't found. In order to force a build of the runtime
system with SMP support, pass <c><![CDATA[--enable-smp-support]]></c>
to <c><![CDATA[configure]]></c> when building OTP.</p>
<p>Own Id: OTP-6525</p>
</item>
</list>
</section>
</section>
<section>
<title>Erts 5.5.3.1</title>
<section>
<title>Fixed Bugs and Malfunctions</title>
<list type="bulleted">
<item>
<p><c><![CDATA[erlang:system_flag(multi_scheduling, block)]]></c> could
cause the emulator with SMP support to deadlock.</p>
<p>Own Id: OTP-6431 Aux Id: OTP-6403 </p>
</item>
<item>
<p>The runtime system with SMP support failed to call the
driver timeout callback of ports in state closing. This
could cause ports to fail flushing their I/O queues.</p>
<p>Own Id: OTP-6432</p>
</item>
<item>
<p>The <c><![CDATA[{Port, closed}]]></c> message from a closed port
could arrive at the port owner before <c><![CDATA[Port]]></c> had
been removed from the result of <c><![CDATA[erlang:ports/0]]></c> in
the runtime system with SMP support.</p>
<p>Own Id: OTP-6437</p>
</item>
<item>
<p>The async id of async jobs created via
<c><![CDATA[driver_async]]></c> wasn't created in a thread safe
manner in the runtime system with SMP support. This could
in worst case cause <c><![CDATA[driver_async_cancel()]]></c> to
cancel another async job than intended.</p>
<p>Own Id: OTP-6438</p>
</item>
<item>
<p>Under rare circumstances a terminating connection between
two nodes could cause an instantaneous reconnect between
the two nodes to fail on the runtime system with SMP
support.</p>
<p>Own Id: OTP-6447</p>
</item>
<item>
<p>In the documentation of the driver entry field
<c><![CDATA[extended_marker]]></c> of erts version 5.5.3
(<c><![CDATA[driver_entry(3)]]></c>), the following is stated: "The
following fields are ignored if this field is equal to
<c><![CDATA[0]]></c>". This is a documentation bug and has been
changed to: "If this field is equal to <c><![CDATA[0]]></c>, all the
fields following this field also <em>have</em> to be
<c><![CDATA[0]]></c>, or <c><![CDATA[NULL]]></c> in case it is a pointer field".</p>
<p>The runtime check for detection of old incompatible
drivers made when loading drivers has been improved. The
emulator can, however, not make sure that a driver that
doesn't use the extended driver interface isn't
incompatible. Therefore, when loading a driver that
doesn't use the extended driver interface, there is a
risk that it will be loaded also when the driver is
incompatible. When the driver use the extended driver
interface, the emulator can verify that it isn't of an
incompatible driver version. You are therefore advised to
use the extended driver interface.</p>
<p>For more information see the <c><![CDATA[erl_driver(3)]]></c> and
<c><![CDATA[driver_entry(3)]]></c> documentation.</p>
<p>Own Id: OTP-6452 Aux Id: OTP-6330 </p>
</item>
<item>
<p>When terminating ports on the runtime system with SMP
support, removal of links to processes was done without
locking the link lock on processes. This could cause an
emulator crash.</p>
<p>Own Id: OTP-6475</p>
</item>
<item>
<p>The emulator with SMP support could crash when a port
flushed its I/O queue by calling <c><![CDATA[driver_async()]]></c>
from the timeout driver callback.</p>
<p>Own Id: OTP-6479</p>
</item>
<item>
<p>Large exit reasons could under rare circumstances cause
the runtime system with SMP support to crash.</p>
<p>Own Id: OTP-6521</p>
</item>
</list>
</section>
<section>
<title>Improvements and New Features</title>
<list type="bulleted">
<item>
<p>Faster system calls for keeping the time accurate are
used on newer Linux kernels, which can result in a
significant speed-up of the emulator on those systems.</p>
<p>Own Id: OTP-6430</p>
</item>
<item>
<p>Added number of async threads and number of scheduler
threads to the system information that can be retrieved
via <c><![CDATA[driver_system_info()]]></c>. For more information see
the <c><![CDATA[erl_driver(3)]]></c> documentation.</p>
<p>Own Id: OTP-6440</p>
</item>
<item>
<p>When <c><![CDATA[SIGUSR1]]></c> is received by the runtime system
with SMP support, the <c><![CDATA[erl_crash.dump]]></c> is now
written by a special thread, instead of as before
directly from the signal handler.</p>
<p>Own Id: OTP-6465</p>
</item>
<item>
<p>term_to_binary/2 with compression is now faster.</p>
<p></p>
<p>term_to_binary/2 now accepts the option
'<c><![CDATA[{compressed,Level}]]></c>' for specifying the
compression level. <c><![CDATA[Level]]></c> must be in the range 0
(no compression) through 9 (highest compression level).
Default is 6.</p>
<p>Future compatibility bugfix: binary_to_term/1 did not
handle the <c><![CDATA[Uniq]]></c> and <c><![CDATA[Index]]></c> fields
correctly.</p>
<p>Own Id: OTP-6494</p>
</item>
<item>
<p>Removed unnecessary reallocation when initializing
kernel-poll set.</p>
<p>Own Id: OTP-6516</p>
</item>
</list>
</section>
</section>
<section>
<title>Erts 5.5.3</title>
<section>
<title>Fixed Bugs and Malfunctions</title>
<list type="bulleted">
<item>
<p>Node and fun table entries could under rare circumstances
be deallocated multiple times on the emulator with SMP
support.</p>
<p>Own Id: OTP-6369</p>
</item>
<item>
<p>epoll_wait() can repeat entries for the same file
descriptor in the result array. This could cause the
ready_input, ready_output, or event callbacks of a driver
to unintentionally be called multiple times. We have only
noted repeated entries when an error condition has
occurred on the file descriptor. In this case, the
repeated entries should normally not be a problem for the
driver since it should detect the error and de-select the
file descriptor. Therefore this bug should be considered
harmless. The emulator now coalesce repeated entries into
one.</p>
<p>You are only affected by this bug if you are using
erts-5.5.2.* and the kernel-poll feature on linux.</p>
<p>Own Id: OTP-6376 Aux Id: OTP-6222 </p>
</item>
<item>
<p>If a process that was waiting in gen_tcp:accept/1 was
killed, calling gen_tcp:accept/1 again on the same listen
socket would fail with '<c><![CDATA[{error,einval}]]></c>'.</p>
<p>Own Id: OTP-6381 Aux Id: seq10535 </p>
</item>
<item>
<p>The emulator failed to start on Solaris 10 when
kernel-poll support was enabled and the maximum number of
filedescriptors configured was less than or equal to 256.</p>
<p>Own Id: OTP-6384 Aux Id: OTP-6222 </p>
</item>
<item>
<p>The R10B compiler could generate unsafe
<c><![CDATA[bs_save/bs_restore]]></c> instructions that could cause
memory corruption. (The R11B compiler does not have that
problem.) The erlang emulator will now refuse to load
R10B-compiled modules that contain such unsafe
<c><![CDATA[bs_save/bs_restore]]></c> instructions. In addition, the
beam_validator module in the compiler will also reject
such instructions (in case it is used to validate R10B
code). (Thanks to Matthew Reilly.)</p>
<p>Own Id: OTP-6386</p>
</item>
<item>
<p>Process and port timers could fail to work properly on
the runtime system with SMP support. Many thanks to
Dmitriy Kargapolov and Serge Aleynikov who tracked down
this bug.</p>
<p>Own Id: OTP-6387</p>
</item>
<item>
<p>Bit syntax code compiled by an R10B compiler that matched
out a floating point number would not properly check that
the floating point number was valid; if the float was,
for instance, NaN the emulator could crash.</p>
<p>Own Id: OTP-6395</p>
</item>
<item>
<p>statistics(runtime) on Windows used to return the elapsed
system time for the process, instead of the user time.
Corrected. (Thanks to Robert Virding.)</p>
<p>Own Id: OTP-6407</p>
</item>
<item>
<p>A loadable driver (loaded by erl_ddll) which used
driver_async() would cause the emulator to crash. (Thanks
to Dmitriy Kargapolov.)</p>
<p>Own Id: OTP-6410</p>
</item>
<item>
<p>Under rare circumstances the emulator on unix platforms
could fail to terminate the Erlang port corresponding to
a port program opened with the <c><![CDATA[exit_status]]></c> option.</p>
<p>Own Id: OTP-6411</p>
</item>
<item>
<p>A link removed via <c><![CDATA[unlink/1]]></c> could under rare
circumstances transfer exit signals for a short period of
time after the call to <c><![CDATA[unlink/1]]></c> had returned when
the runtime system with SMP support was used.</p>
<p>Own Id: OTP-6425 Aux Id: OTP-6160 </p>
</item>
</list>
</section>
<section>
<title>Improvements and New Features</title>
<list type="bulleted">
<item>
<p>In the runtime system with SMP support, ports are now
being scheduled on the scheduler threads interleaved with
processes instead of being run in a separate I/O thread.</p>
<p>Own Id: OTP-6262</p>
</item>
<item>
<p>More interfaces are added in erl_ddll, to support
different usage scenarios.</p>
<p>Own Id: OTP-6307 Aux Id: OTP-6234 </p>
</item>
<item>
<p>In the runtime system with SMP support, the global
I/O lock has been replaced with a more fine grained port
locking scheme. Port locking is either done on driver
level, i.e., all ports executing the same driver share a
lock, or on port level, i.e., each port has its own lock.
Currently the inet driver, the efile driver, and the
spawn driver use port level locking and all other
statically linked in drivers use driver level locking. By
default dynamically linked in drivers will use locking on
driver level. For more information on how to enable port
level locking see the <c><![CDATA[erl_driver(3)]]></c> and the
<c><![CDATA[driver_entry(3)]]></c> man pages. </p>
<p>As of erts
version 5.5.3 the driver interface has been extended. The
extended interface introduce version management, the
possibility to pass capability flags to the runtime
system at driver initialization, and some new driver API
functions. For example, the <c><![CDATA[driver_system_info()]]></c>
function which can be used to determine if the driver is
run in a runtime system with SMP support or not. The
extended interface doesn't have to be used, <em>but</em>
dynamically linked in driver <em>have</em> to be
recompiled. For information see the <c><![CDATA[erl_driver(3)]]></c>
and the <c><![CDATA[driver_entry(3)]]></c> man pages. </p>
<p><em>NOTE:</em> Dynamically linked in drivers
<em>have</em> to be recompiled. </p>
<p>*** POTENTIAL INCOMPATIBILITY ***</p>
<p>Own Id: OTP-6330 Aux Id: OTP-6262 </p>
</item>
<item>
<p>A test and debug feature which modifies the timing of the
runtime system has been added. For more information, see
documentation of the <c><![CDATA[+T<Level>]]></c> command line
argument of <c><![CDATA[erl(1)]]></c>.</p>
<p>Own Id: OTP-6382</p>
</item>
<item>
<p>The version of zlib (http://zlib.net) linked into
run-time system has been updated to version 1.2.3.</p>
<p>Own Id: OTP-6394</p>
</item>
<item>
<p>The <c><![CDATA[erlc]]></c> program now passes on the <c><![CDATA[-smp]]></c>
and <c><![CDATA[-hybrid]]></c> options to the Erlang emulator it
starts. This is mainly useful for compiling native code,
because native code must be compiled with same type of
run-time system as in which it will be run.</p>
<p>If the <c><![CDATA[-s]]></c> option is given, <c><![CDATA[erlc]]></c> now
prints a warning that it is deprecated and that it will
be removed in R12B.</p>
<p>Own Id: OTP-6398</p>
</item>
<item>
<p>The <c><![CDATA[schedulers]]></c> option of
<c><![CDATA[erlang:system_flag/2]]></c> has been removed, i.e., the
number of scheduler threads cannot be changed after
emulator boot time any more.</p>
<p>A <c><![CDATA[multi_scheduling]]></c> option has been added to
<c><![CDATA[erlang:system_flag/2]]></c>. This option can be used for
blocking and unblocking multi-scheduling. For more
information see the <c><![CDATA[erlang(3)]]></c> documentation.</p>
<p>Own Id: OTP-6403</p>
</item>
<item>
<p>A port program that had been started with the
<c><![CDATA[exit_status]]></c> option and closed one of the pipes
used for communication with the emulator caused the
emulator to continuously poll for termination of the port
program. This only became a problem when the emulator
didn't have other things to do and the port program
closed a pipe much earlier than it terminated. When the
emulator had other things to do, such as running Erlang
processes, the emulator polled for termination in between
scheduling of processes.</p>
<p>Now the emulator doesn't poll for termination of the port
program at all; instead, it waits for the child signal
from the terminated port program to arrive and then
schedules the Erlang port for termination.</p>
<p>The termination of any port programs have also been
optimized. Previously the termination of any port program
did always cause a scan of a table of the size equal to
maximum number of file descriptors. If the maximum number
of file descriptors was large, this scan could be quite
expensive. Now the search have been reduced to the set of
ports started with the <c><![CDATA[exit_status]]></c> option.</p>
<p>Note, all of the above only applies to Erlang emulators
on Unix platforms.</p>
<p>Own Id: OTP-6412 Aux Id: seq10194 </p>
</item>
<item>
<p>* BEAM: added support for floating-point exceptions on
FreeBSD (32-bit x86)</p>
<p>* SMP: made locking procedures work even when native
lock operations aren't implemented</p>
<p>* SMP: improved timing accuracy in the timer thread
(if enabled)</p>
<p>Own Id: OTP-6424</p>
</item>
</list>
</section>
</section>
<section>
<title>Erts 5.5.2.2</title>
<section>
<title>Fixed Bugs and Malfunctions</title>
<list type="bulleted">
<item>
<p>A bug in the kernel poll implementation could cause the
emulator to omit polling for events on file descriptors.
The bug was only present when using the kernel poll
implementation based on epoll or kqueue. This bug was
introduced in erts-5.5.2.</p>
<p>Own Id: OTP-6344 Aux Id: OTP-6222 </p>
</item>
</list>
</section>
</section>
<section>
<title>Erts 5.5.2.1</title>
<section>
<title>Fixed Bugs and Malfunctions</title>
<list type="bulleted">
<item>
<p>The run_erl program now acquires its pseudo-ttys using
openpty(3), so that it will work on newer Linux
distributions that don't have the traditional pseudo-tty
devices in the file system. On platforms that don't have
openpty(3), run_erl will still search for pseudo-tty
devices in the file system.</p>
<p>The run_erl program will now wait using waitpid(3) to
prevent the program it spawned to become defunct. run_erl
will also terminate after a delay of 5 seconds (to allow
any pending output to be written to the log file) if the
spawned program terminates even if some child of it still
holds stdin and/or stdout open.</p>
<p>Own Id: OTP-6225 Aux Id: seq10500 </p>
</item>
<item>
<p>A bug in ordered_set ETS datatyp caused ets:select (and
match) to return faulty results when the table contained
process id's from another node.</p>
<p>Own Id: OTP-6338</p>
</item>
</list>
</section>
</section>
<section>
<title>Erts 5.5.2</title>
<section>
<title>Fixed Bugs and Malfunctions</title>
<list type="bulleted">
<item>
<p>erlc: A typo in the help text for '-pa path' was
corrected.</p>
<p>Own Id: OTP-6218</p>
</item>
<item>
<p>Failure in port command/control driver callbacks could
crash the non-SMP emulator. This bug was introduced in
the 5.5 version of erts.</p>
<p>Own Id: OTP-6224</p>
</item>
<item>
<p>Erroneous "<c><![CDATA[Input driver gone away without deselecting!]]></c>" error reports could sometimes occur
when a driver instance terminated in the ready_output()
callback of the driver. This bug was only present in
emulators that used poll(). Note, that this bug was
harmless, the only effect it had was the erroneous error
report.</p>
<p>Own Id: OTP-6229 Aux Id: OTP-3993, Seq5266, Seq7247,
OTP-4307 </p>
</item>
<item>
<p>The emulator could cause a type assertion failure while
writing an erl_crash.dump, causing the erl_crash.dump to
be terminated and a core dump generated.</p>
<p>Own Id: OTP-6235 Aux Id: seq10444 </p>
</item>
<item>
<p>The registered name of a process is now the last
observable resource removed before links and monitors are
triggered when a process terminates.</p>
<p>Previously ets tables were removed after the registered
name. This could cause problems on the runtime system
with SMP support for code that expected that ets tables
owned by a specific process had been removed if the name
of the process had been removed.</p>
<p>Own Id: OTP-6237</p>
</item>
<item>
<p>Failure to fork() a new (os) process could cause the
emulator to deadlock. This bug affect all emulators with
SMP support, and emulators with async thread support on
SunOS.</p>
<p>Own Id: OTP-6241 Aux Id: OTP-3906 </p>
</item>
<item>
<p>Fprof traces could become truncated for the SMP emulator.
This bug has now been corrected.</p>
<p>Own Id: OTP-6246</p>
</item>
<item>
<p>The undocumented functions inet:getiflist/0,1
inet:ifget/2,3 and inet:getif/1 were completely broken on
Windows. That has been fixed.</p>
<p>Own Id: OTP-6255</p>
</item>
<item>
<p>Behavior in case of disappeared nodes when using the
dist_auto_connect once got changed in R11B-1. The
timeouts regarding normal distributed operations is now
reverted to the old (pre R11B-1).</p>
<p>Own Id: OTP-6258 Aux Id: OTP-6200, seq10449 </p>
</item>
<item>
<p>The bsl and bsr operators could cause the emulator to
crash if given invalid arguments. (Thanks to datacompboy
and Per Gustafsson.)</p>
<p>Own Id: OTP-6259</p>
</item>
<item>
<p>driver_cancel_timer() could under certain circumstances
fail to cancel the timer on the runtime system with SMP
support.</p>
<p>Own Id: OTP-6261</p>
</item>
<item>
<p>A call to erlang:system_info(info) could deadlock the
runtime system with SMP support.</p>
<p>Own Id: OTP-6268</p>
</item>
</list>
</section>
<section>
<title>Improvements and New Features</title>
<list type="bulleted">
<item>
<p>Exit signals are now truly asynchronous in the runtime
system with SMP support. This simplifies locking in bifs
a lot, and makes process termination cheaper.</p>
<p>Own Id: OTP-6160</p>
</item>
<item>
<p>When tracing on the runtime system with SMP support it
can be difficult to know when a trace has been delivered
to the tracer. A new built in function
<c><![CDATA[erlang:trace_delivered/1]]></c> has been introduced in
order to make it easier to know when the trace has been
delivered. See the erlang(3) man page for more
information.</p>
<p>Own Id: OTP-6205 Aux Id: OTP-6269 </p>
</item>
<item>
<p>Kernel poll support can now be combined with SMP support.
Currently the following kernel poll versions exist:
<c><![CDATA[/dev/poll]]></c>, <c><![CDATA[epoll]]></c>, and <c><![CDATA[kqueue]]></c>. Linux
<c><![CDATA[kpoll]]></c> has been replaced with <c><![CDATA[epoll]]></c>. Some
time in the future there will also be a kernel poll
version using Solaris event ports.</p>
<p>The "check io" implementation for unix has been
completely rewritten. The current kernel poll
implementation reduce the amount of system calls needed
compared to the old kernel poll implementation.</p>
<p>When <c><![CDATA[epoll]]></c> or <c><![CDATA[kqueue]]></c> is used either
<c><![CDATA[poll]]></c> or <c><![CDATA[select]]></c> is used as fallback.
Previously only <c><![CDATA[poll]]></c> could be used as fallback.
Since <c><![CDATA[select]]></c> now can be used as fallback, kernel
poll support is now also available on newer MacOSX. Note
however, when <c><![CDATA[select]]></c> is used as fallback, the
maximum number of file descriptors is limited to
<c><![CDATA[FD_SETSIZE]]></c>. </p>
<p>Kernel poll support is now enabled by default if
<c><![CDATA[/dev/poll]]></c>, <c><![CDATA[epoll]]></c>, or <c><![CDATA[kqueue]]></c> is found
when building OTP, i.e. you do not have to pass the
<c><![CDATA[--enable-kernel-poll]]></c> argument to <c><![CDATA[configure]]></c>.
As before, kernel poll is disabled by default in the
runtime system. In order to enable it, pass the
<c><![CDATA[+Ktrue]]></c> command line argument to <c><![CDATA[erl]]></c>.</p>
<p>Note: <c><![CDATA[configure]]></c> will refuse to enable kernel poll
support on FreeBSD since <c><![CDATA[kqueue]]></c> have problems with
(at least) pipes on all version of FreeBSD that we have
tested.</p>
<p>Own Id: OTP-6222 Aux Id: seq10380 </p>
</item>
<item>
<p>The <c><![CDATA[erl_ddll]]></c> module and the code in the emulator
have been completely rewritten; several bugs were fixed.</p>
<p>Own Id: OTP-6234</p>
</item>
<item>
<p>The SMP emulator now avoids locking for the following
operations (thus making them as fast as in the UP
emulator): <c><![CDATA[atom_to_list/1]]></c>, atom comparison, atom
hashing, <c><![CDATA[erlang:apply/3]]></c>.</p>
<p>Own Id: OTP-6252</p>
</item>
<item>
<p>There are new BIFs <c><![CDATA[erlang:spawn_monitor/1,3]]></c>,
and the new option <c><![CDATA[monitor]]></c> for
<c><![CDATA[spawn_opt/2,3,4,5]]></c>.</p>
<p>The <c><![CDATA[observer_backend]]></c> module has been updated to
handle the new BIFs.</p>
<p>Own Id: OTP-6281</p>
</item>
</list>
</section>
</section>
<section>
<title>Erts 5.5.1.1</title>
<section>
<title>Improvements and New Features</title>
<list type="bulleted">
<item>
<p>There is now an option read_packets for UDP sockets that
sets the maximum number of UDP packets that will be read
for each invocation of the socket driver.</p>
<p>Own Id: OTP-6249 Aux Id: seq10452 </p>
</item>
</list>
</section>
</section>
<section>
<title>Erts 5.5.1</title>
<section>
<title>Fixed Bugs and Malfunctions</title>
<list type="bulleted">
<item>
<p>Complex pattern matching of strings would fail in the 64
bits emulator because of a bug in the loader. (Thanks to
Igor Goryachev.)</p>
<p>Own Id: OTP-6142</p>
</item>
<item>
<p><c><![CDATA[-134217728 div 134217728]]></c> and <c><![CDATA[-134217728 rem 134217728]]></c> would be calculated incorrectly.
<c><![CDATA[abs(-2147483648)]]></c> could in unlucky circumstances
cause a heap overflow, as could size(Binary) when size of
the binary was larger than 128Mb.</p>
<p>Own Id: OTP-6154</p>
</item>
<item>
<p>erlang:display/1 displayed erroneous values for negative
integers.</p>
<p>Big integers (both positive and negative) were previously
displayed in hexadecimal form while small integers were
displayed in decimal form. All integers are now displayed
in decimal form.</p>
<p>NOTE: erlang:display/1 should only be used for debugging.</p>
<p>Own Id: OTP-6156</p>
</item>
<item>
<p>A call to erlang:trace/3 with erroneous flags caused the
SMP emulator to deadlock instead of exiting the calling
process with badarg.</p>
<p>Own Id: OTP-6175</p>
</item>
<item>
<p>A bug causing the emulator to hang when exiting a process
that is exception traced has been fixed.</p>
<p>Own Id: OTP-6180</p>
</item>
<item>
<p>ets:rename/1 could deadlock, or crash the SMP emulator
when the table wasn't a named table.</p>
<p>ets:next/2, and ets:prev/2 could return erroneous results
on the SMP emulator.</p>
<p>Own Id: OTP-6198 Aux Id: seq10392, seq10415 </p>
</item>
<item>
<p>A memory allocation bug could cause the SMP emulator to
crash when a process had executed a <c><![CDATA[receive after]]></c>
with a larger timeout than 10 minutes.</p>
<p>Own Id: OTP-6199</p>
</item>
<item>
<p>The runtime system with SMP support did not slowly adjust
its view of time when the system time suddenly changed.</p>
<p>Timeouts could sometimes timeout too early on the runtime
system with SMP support.</p>
<p>Own Id: OTP-6202</p>
</item>
</list>
</section>
<section>
<title>Improvements and New Features</title>
<list type="bulleted">
<item>
<p>The smp runtime system now automatically detects the
number of logical processors on MacOSX (darwin) and
OpenBSD.</p>
<p>The smp runtime system is now built by default on MacOSX
(darwin) on x86.</p>
<p>Own Id: OTP-6119</p>
</item>
<item>
<p>The <c><![CDATA[-smp]]></c> command line argument now take the
following options: <c><![CDATA[enable]]></c>, <c><![CDATA[auto]]></c>, or
<c><![CDATA[disable]]></c>.</p>
<p>Especially the <c><![CDATA[-smpauto]]></c> argument is useful since
it starts the Erlang runtime system with SMP support if
it is available and more than one logical processor are
detected; otherwise, it starts the Erlang runtime system
without SMP support. For more information see the
<c><![CDATA[erl(1)]]></c> man page.</p>
<p>Own Id: OTP-6126</p>
</item>
<item>
<p>Increased the reduction cost for sending messages in the
SMP emulator so it behaves more like the non-SMP
emulator.</p>
<p>Own Id: OTP-6196</p>
</item>
<item>
<p>A port running a dynamically linked-in driver that exits
due to the driver being unloaded now exits with exit
reason <c><![CDATA[driver_unloaded]]></c>. Previously the port exited
with exit reason <c><![CDATA[-1]]></c>.</p>
<p>*** POTENTIAL INCOMPATIBILITY ***</p>
<p>Own Id: OTP-6204</p>
</item>
<item>
<p>Changed name of the <c><![CDATA[erlang:system_info/1]]></c> argument
<c><![CDATA[scheduler]]></c> to <c><![CDATA[scheduler_id]]></c>. This since the
<c><![CDATA[scheduler]]></c> argument so easily could be mixed up
with the <c><![CDATA[schedulers]]></c> argument (both returning
integers).</p>
<p>*** POTENTIAL INCOMPATIBILITY ***</p>
<p>Own Id: OTP-6208</p>
</item>
<item>
<p>The changes below were made by Mikael Pettersson, HiPE.</p>
<p>HiPE runtime system:</p>
<p>Reduce overheads in the HiPE runtime system's BIF glue
code.</p>
<p>Fix bug when exceptions are thrown from BEAM to HiPE.</p>
<p>Support SPARC on Linux.</p>
<p>Support x86 on FreeBSD.</p>
<p>Floating-point exceptions:</p>
<p>Reduce overheads in checking results of floating-point
operations.</p>
<p>Minor bug fix in SSE2 floating-point exception
handling.</p>
<p>Support SSE2 floating-point exceptions on 32-bit x86
machines.</p>
<p>Make FP exceptions work in the SMP runtime system on
FreeBSD/x86.</p>
<p>Support floating-point exceptions on SPARCs running
Linux.</p>
<p>Runtime system:</p>
<p>Minor scheduler optimisation in the non-SMP runtime
system.</p>
<p>Substantial reduction of I/O thread overheads in the
SMP runtime system if the separate timer thread is used.
(In R11B-1, the separate timer thread is not used.)</p>
<p>Own Id: OTP-6211</p>
</item>
</list>
</section>
</section>
<section>
<title>ERTS 5.5</title>
<section>
<title>Fixed Bugs and Malfunctions</title>
<list type="bulleted">
<item>
<p>Previously <c><![CDATA[unlink/1]]></c> and <c><![CDATA[erlang:demonitor/2]]></c>
behaved completely asynchronous. This had one undesirable
effect, though. You could never know when you were
guaranteed <em>not</em> to be affected by a link that you
had unlinked or a monitor that you had demonitored.</p>
<p>The new behavior of <c><![CDATA[unlink/1]]></c> and
<c><![CDATA[erlang:demonitor/2]]></c> can be viewed as two operations
performed atomically. Asynchronously send an unlink
signal or a demonitor signal, and ignore any future
results of the link or monitor.</p>
<p><em>NOTE</em>: This change can cause some obscure code
to fail which previously did not. For example, the
following code might hang:</p>
<code type="none"><![CDATA[
Mon = erlang:monitor(process, Pid),
%% ...
exit(Pid, bang),
erlang:demonitor(Mon),
receive
{'DOWN', Mon , process, Pid, _} -> ok
%% We were previously guaranteed to get a down message
%% (since we exited the process ourself), so we could
%% in this case leave out:
%% after 0 -> ok
end,
]]></code>
<p>*** POTENTIAL INCOMPATIBILITY ***</p>
<p>Own Id: OTP-5772</p>
</item>
<item>
<p>Two bugs fixed: If the environment variable <c><![CDATA[ERL_FLAGS]]></c>
was set, its contents would be appended to the end of the
command line even if the command line had an <c><![CDATA[-extra]]></c>
options. Changed to place the options from <c><![CDATA[ERL_FLAGS]]></c>
just before <c><![CDATA[-extra]]></c>. Also, the <c><![CDATA[-smp]]></c> and
<c><![CDATA[-hybrid]]></c> flags no longer have any effect if placed
after <c><![CDATA[-extra]]></c>.</p>
<p>Own Id: OTP-6054</p>
</item>
</list>
</section>
<section>
<title>Improvements and New Features</title>
<list type="bulleted">
<item>
<p>The documentation for writing drivers in the ERTS User's
Guide has been expanded and updated.</p>
<p>Own Id: OTP-5192</p>
</item>
<item>
<p>The <c><![CDATA[andalso]]></c> and <c><![CDATA[orelse]]></c> operators are
now allowed to be used in guards. That also applies to
match specifications.</p>
<p>Own Id: OTP-5894 Aux Id: OTP-5149 </p>
</item>
<item>
<p>There is a new trace match spec function
<c><![CDATA[{exception_trace}]]></c> and a corresponding trace
message <c><![CDATA[exception_from]]></c> that can be used to trace
on any exit from a function, both normal function return
and exception return. See the documentation for
details.</p>
<p>The <c><![CDATA[return_to]]></c> trace message is now also
generated when the execution returns to a function due to
catching an exception.</p>
<p>Own Id: OTP-5956</p>
</item>
<item>
<p>Erlang runtime system with SMP (symmetric multi processing)
support.</p>
<p>The runtime system with SMP support is in this release
focused on stability and there are a number of steps
with optimizations to follow before it will take
full advantage of multi processor systems.
The released system is however truly multi threaded
and you will notice increased performance
for many applications already.
We recommend that you evaluate your application on
the SMP version of the runtime system and wait for some
more optimizations before you use it in a real product.
You will then discover if there are any problems in
your application that needs to be fixed in order for
it to work properly in a multi threaded environment.
More optimized versions of the runtime system
with SMP support will be included in the R11B
maintenance releases.</p>
<p>The SMP enabled runtime system will be started if
the <c><![CDATA[-smp]]></c> command line argument is passed to
the <c><![CDATA[erl]]></c> command. In order to make use of more than
one processor core, multiple scheduler threads are used. By
default, the number of scheduler threads will equal
the number of processor cores. The number of scheduler
threads can be set with the <c><![CDATA[+S]]></c> command line argument.
For more information see the <c><![CDATA[erl(1)]]></c> man page.</p>
<p>A runtime system with SMP support is by default built on
the following platforms if posix threads, and a gcc
compiler of at least version 2.95 is found:</p>
<list type="bulleted">
<item>
<p>Linux with at least kernel version 2.6 and the Native
POSIX Thread Library on x86, x86_64, and 32-bits
PowerPC.</p>
</item>
<item>
<p>Solaris of at least version 8 on 32-bits SPARC-V9.</p>
</item>
<item>
<p>MacOSX of at least version 10.4 (Darwin 8.0) on
32-bits PowerPC.</p>
</item>
</list>
<p>The runtime system with SMP support is known <em>not</em>
to build on:</p>
<list type="bulleted">
<item>
<p>Windows.</p>
</item>
<item>
<p>Linux with kernel versions less than 2.4, or without
the Native POSIX Thread Library.</p>
</item>
<item>
<p>Other hardware platforms than x86, x86_64, 32-bits
SPARC-V9 and 32-bits PowerPC.</p>
</item>
</list>
<p>Windows will be supported in a future release.</p>
<p>The runtime system with SMP support might build on other
operating systems in combination with supported hardware.
In order to force a build of a runtime system with SMP
support, pass the <c><![CDATA[--enable-smp-support]]></c> command line
argument to configure. Note, however, that it is not enough
that it builds. The underlying thread library and operating
system has to provide SMP support as well. If the thread
library does not distribute scheduler threads over multiple
processor cores then the runtime system will only seemingly
provide SMP support. If the runtime system is not built by
default on a specific platform, we have <em>not</em> tested
it on that platform.</p>
<p><em>NOTE</em>: The design of SMP support for drivers is
ongoing. There will probably be incompatible driver
changes (only affecting drivers run on the runtime system
with SMP support) released as patches for R11B.</p>
<p><em>Potential incompatibility</em>: Previously, specific
driver call-backs were always called from the same thread.
This is <em>not</em> true in the runtime system with SMP
support. Calls to call-backs will be made from different
threads, e.g., two consecutive calls to exactly the same
call-back can be made from two different threads. This
will in <em>most</em> cases not be a problem. All calls
to call-backs are synchronized, i.e., only one call-back
will be called at a time.</p>
<p>In the future the default behavior will <em>probably</em>
be the following: Calls to call-backs will, as now, be
made from different threads. Calls to call-backs in the
same driver <em>instance</em> will be synchronized. It
will probably be possible to configure so that all calls
to call-backs in all driver instances of a specific
driver type will be synchronized. It may be possible to
configure so that all calls to call-backs of a driver
instance or a of a specific driver type will be made from
the same thread.</p>
<p>Parallelism in the Erlang code executed is a necessity
for the Erlang runtime system to be able to take
advantage of multi-core or multi-processor hardware.
There need to be at least as many Erlang processes
runnable as processor cores for the Erlang runtime system
to be able to take advantage of all processor cores.</p>
<p>An Erlang runtime system with SMP support with only one
Erlang process runnable all the time will almost always be
slower than the same Erlang runtime system without SMP
support. This is due to thread synchronization overhead.</p>
<p>Known major bottleneck in the Erlang runtime system:</p>
<list type="bulleted">
<item>
<p>Currently the I/O system uses one "big lock", i.e. only
one thread can do I/O at a time (with the exception of
async threads and threads created by users own linked-in
drivers). This is high on the list of things to
optimize. Note, code that does not do I/O can be executed
at the same time as one thread does I/O.</p>
</item>
</list>
<p>Some pitfalls which might cause Erlang programs that work on
the non-SMP runtime system to fail on the SMP runtime
system:</p>
<list type="bulleted">
<item>
<p>A newly spawned process will often begin executing
immediately. Code that expects that the parent process
will be able to execute for a while before the child
process begins executing is likely to fail.</p>
</item>
<item>
<p>High priority processes could previously provide
mutual exclusion (bad programming style) by preventing
normal and low priority processes from being run. High
priority processes cannot be used this way to provide
mutual exclusion.</p>
</item>
<item>
<p><c><![CDATA[erlang:yield()]]></c> could be used to provide some
kind of temporary mutual exclusion (also bad programming
style). <c><![CDATA[erlang:yield()]]></c> cannot be used to provide
any kind of mutual exclusion.</p>
</item>
<item>
<p>Obscure pitfall, only if a process being traced also
sends normal messages to the tracer:<br></br>
The order between trace messages and normal
messages is undefined. I.e. the order between normal
messages sent from a tracee to a tracer and the trace
messages generated from the same tracee to the same
tracer is undefined. The internal order of normal
messages and the internal order of trace messages will,
of course, be preserved as before.</p>
</item>
</list>
<p>The kernel poll feature is currently not supported by
the runtime system with SMP support. It will probably be
supported in a future release.</p>
<p>*** POTENTIAL INCOMPATIBILITY ***</p>
<p>Own Id: OTP-6006 Aux Id: OTP-6095 </p>
</item>
<item>
<p>Linked-in driver modifications.</p>
<list type="bulleted">
<item>
<p>Linked-in drivers must be recompiled.</p>
</item>
<item>
<p>The <c><![CDATA[refc]]></c> field in the <c><![CDATA[ErlDrvBinary]]></c> type
has been removed. The reference count can be accessed
via API functions. For more information see
the <c><![CDATA[erl_driver(1)]]></c> man page.</p>
</item>
</list>
<p>*** POTENTIAL INCOMPATIBILITY ***</p>
<p>Own Id: OTP-6095 Aux Id: OTP-6006 </p>
</item>
</list>
</section>
</section>
<section>
<title>ERTS 5.4.13</title>
<section>
<title>Fixed Bugs and Malfunctions</title>
<list type="bulleted">
<item>
<p>Large files (more than 2 GBytes) are now handled on
Solaris 8.</p>
<p>Own Id: OTP-5849 Aux Id: seq10157 </p>
</item>
<item>
<p>A failing bit syntax construction could fail with the
PREVIOUS exception reason that had occurred in the process
(instead of with <c><![CDATA[badarg]]></c>).</p>
<p>Own Id: OTP-5911</p>
</item>
<item>
<p>When building OTP, the Kernel application was built in
both the primary and secondary bootstrap steps, which
would cause problems if OTP including its bootstrap is
checked into a version control system (such as CVS).
(Thanks to Sebastian Strollo.)</p>
<p>Own Id: OTP-5921</p>
</item>
<item>
<p><c><![CDATA[binary_to_term(<<131,109,255,255,255,255)]]></c> and
similar expressions used to crash the emulator instead of
causing a <c><![CDATA[badarg]]></c> exception. (Thanks to Matthias
Lang.)</p>
<p>Own Id: OTP-5933</p>
</item>
<item>
<p><c><![CDATA[erlang:hibernate/3]]></c> could sometimes crash the emulator
when no heap was needed.</p>
<p>Own Id: OTP-5940</p>
</item>
<item>
<p>Execution of match specs could under rare circumstances
cause the emulator to dump core.</p>
<p>Execution of match specs could cause memory leaks in the
hybrid emulator.</p>
<p>Own Id: OTP-5955</p>
</item>
<item>
<p>A bug in <c><![CDATA[erlang:trace_info/2]]></c> when getting info for a
function in a deleted module resulting in an emulator
crash, has been corrected.</p>
<p>Own Id: OTP-5957</p>
</item>
<item>
<p>Different (and old) <c><![CDATA[config.guess]]></c> files in the
ERTS and Erl_Interface applications would cause build
problems on the new Intel-based iMacs.
(Thanks to Sebastion Strollo.)</p>
<p>Own Id: OTP-5967</p>
</item>
<item>
<p>pthread header and library mismatch on Linux systems (at
least some SuSE and Debian) with both NPTL and
Linuxthreads libraries installed.</p>
<p>Own Id: OTP-5981</p>
</item>
</list>
</section>
<section>
<title>Improvements and New Features</title>
<list type="bulleted">
<item>
<p>The driver_set_timer did not change the previous timeout
if called a second time. Now it works as specified,
changing the timeout.</p>
<p>Own Id: OTP-5942</p>
</item>
<item>
<p>The undocumented <c><![CDATA[{packet,http}]]></c> option (for the
<c><![CDATA[gen_tcp]]></c> module) did not not work correctly when
there were multiple continuation lines. (Thanks to Per
Hedeland.)</p>
<p>Own Id: OTP-5945</p>
</item>
<item>
<p>The setuid_socket_wrap program was corrected to work for
C compilers that treat the <c><![CDATA[char]]></c> type as unsigned.
(Thanks to Magnus Henoch.)</p>
<p>Own Id: OTP-5946</p>
</item>
</list>
</section>
</section>
<section>
<title>ERTS 5.4.12</title>
<section>
<title>Fixed Bugs and Malfunctions</title>
<list type="bulleted">
<item>
<p>Fixed a minor build problem on Windows.</p>
<p>Own Id: OTP-5819 Aux Id: OTP-5382 OTP-5540 OTP-5577 </p>
</item>
<item>
<p>The <c><![CDATA[delay_send]]></c> option for <c><![CDATA[gen_tcp]]></c> was
broken on Windows.</p>
<p>Own Id: OTP-5822</p>
</item>
<item>
<p>If there were user-defined variables in the boot
script, and their values were not provided using the
<c><![CDATA[-boot_var]]></c> option, the emulator would refuse to
start with a confusing error message. Corrected to show a
clear, understandable message.</p>
<p>The <c><![CDATA[prim_file]]></c> module was modified to not depend
on the <c><![CDATA[lists]]></c> module, to make it possible to start
the emulator using a user-defined loader. (Thanks to
Martin Bjorklund.)</p>
<p>Own Id: OTP-5828 Aux Id: seq10151 </p>
</item>
</list>
</section>
<section>
<title>Improvements and New Features</title>
<list type="bulleted">
<item>
<p>The HiPE compiler identifies more leaf functions,
giving slightly faster code.</p>
<p>Corrected problems in HiPE's coalescing register
allocating that would cause it to fail when compiling
very large functions (e.g. some of parse modules in the
Megaco application).</p>
<p>Own Id: OTP-5853</p>
</item>
</list>
</section>
</section>
<section>
<title>ERTS 5.4.11</title>
<section>
<title>Fixed Bugs and Malfunctions</title>
<list type="bulleted">
<item>
<p>Timers could sometimes timeout too early. This bug has
now been fixed.</p>
<p>Automatic cancellation of timers created by
<c><![CDATA[erlang:send_after(Time,pid(),Msg)]]></c> and
<c><![CDATA[erlang:start_timer(Time,pid(),Msg)]]></c> has been
introduced. Timers created with the receiver specified by a
pid, will automatically be cancelled when the receiver
exits. For more information see the <c><![CDATA[erlang(3)]]></c> man
page.</p>
<p>In order to be able to maintain a larger amount of timers
without increasing the maintenance cost, the internal
timer wheel and bif timer table have been enlarged.</p>
<p>Also a number of minor bif timer optimizations have been
implemented.</p>
<p>Own Id: OTP-5795 Aux Id: OTP-5090, seq8913, seq10139,
OTP-5782 </p>
</item>
<item>
<p><c><![CDATA[erlang:monitor(process,Pid)]]></c> hanged if <c><![CDATA[Pid]]></c>
referred to a process on a non-existing node with the same
nodename as the nodename of node on which the call was made.
This bug has now been fixed.</p>
<p>Own Id: OTP-5827</p>
</item>
</list>
</section>
<section>
<title>Improvements and New Features</title>
<list type="bulleted">
<item>
<p>By setting Unix environment variables, the priority for
the emulator can be lowered when it is writing crash
dumps and the time allowed for finishing writing a crash
dump can be set to a certain number of seconds. See the
documentation for <c><![CDATA[erl]]></c> in the ERTS application.
(Also, a few other previously undocumented environment
variables are now documented.)</p>
<p>Own Id: OTP-5818</p>
</item>
<item>
<p>Documentation improvements:</p>
<p>- documentation for <c><![CDATA[erlang:link/1]]></c> corrected</p>
<p>- command line flag <c><![CDATA[-code_path_cache]]></c> added</p>
<p>- <c><![CDATA[erl]]></c> command line flags clarifications</p>
<p>- <c><![CDATA[net_kernel(3)]]></c> clarifications</p>
<p>Own Id: OTP-5847</p>
</item>
</list>
</section>
</section>
<section>
<title>ERTS 5.4.10</title>
<section>
<title>Fixed Bugs and Malfunctions</title>
<list type="bulleted">
<item>
<p>-D_GNU_SOURCE is now always passed on the compile command
line on linux. This in order to ensure that all included
system headers see _GNU_SOURCE defined.</p>
<p>_GNU_SOURCE is now also defined on linux in configure
when looking for features.</p>
<p>Some minor (harmless) configure bugs were also fixed.</p>
<p>Own Id: OTP-5749</p>
</item>
<item>
<p>Some compiler warnings and Dialyzer warnings were
eliminated in the Tools application.</p>
<p>When tracing to a port (which <c><![CDATA[fprof]]></c> does),
there could be fake schedule out/schedule in messages
sent for a process that had exited.</p>
<p>Own Id: OTP-5757</p>
</item>
</list>
</section>
<section>
<title>Improvements and New Features</title>
<list type="bulleted">
<item>
<p>The BIFs <c><![CDATA[iolist_size/1]]></c> and <c><![CDATA[iolist_to_binary/1]]></c>
has been added.</p>
<p>The BIF <c><![CDATA[list_to_existing_atom/1]]></c> has been added.</p>
<p>Minor bug fix: The exception reason could be changed
to <c><![CDATA[error]]></c> inside nested try/catch constructs if the
<c><![CDATA[erlang:raise/3]]></c> BIF was called with an empty
stacktrace. (Calling <c><![CDATA[erlang:raise/3]]></c> with an empty
stacktrace is NOT recommended.)</p>
<p>Minor bugfix: On Windows, <c><![CDATA[file:open/2]]></c> will now
return the documented error reason <c><![CDATA[{error,eisdir}]]></c>
if the filename refers to a directory (it used to return
<c><![CDATA[{error,eacces}]]></c>).</p>
<p>The message in the documentation for
<c><![CDATA[erlang:system_monitor/2]]></c>, description of
<c><![CDATA[busy_dist_port]]></c>, was corrected.</p>
<p>Own Id: OTP-5709 Aux Id: seq10100 </p>
</item>
<item>
<p>The previously undocumented and UNSUPPORTED <c><![CDATA[zlib]]></c>
module has been updated in an incompatible way and many
bugs have been corrected. It is now also documented.</p>
<p>*** POTENTIAL INCOMPATIBILITY ***</p>
<p>Own Id: OTP-5715</p>
</item>
<item>
<p>New socket options <c><![CDATA[priority]]></c> and <c><![CDATA[tos]]></c> for
platforms that support them (currently only Linux).</p>
<p>Own Id: OTP-5756</p>
</item>
<item>
<p>Only the emulator is now linked with termcap library in
order to decrease library dependencies for other otp
programs.</p>
<p>Own Id: OTP-5758</p>
</item>
</list>
</section>
</section>
<section>
<title>ERTS 5.4.9.2</title>
<section>
<title>Fixed Bugs and Malfunctions</title>
<list type="bulleted">
<item>
<p>The native resolver has gotten an control API for
extended debugging and soft restart. It is:
<c><![CDATA[inet_gethost_native:control(Control)]]></c><br></br><c><![CDATA[Control = {debug_level,Level} | soft_restart]]></c><br></br><c><![CDATA[Level = integer() in the range 0-4]]></c>.</p>
<p>Own Id: OTP-5751 Aux Id: EABln25013 </p>
</item>
</list>
</section>
</section>
<section>
<title>ERTS 5.4.9.1</title>
<section>
<title>Improvements and New Features</title>
<list type="bulleted">
<item>
<p>On VxWorks, epmd did not handle file descriptors with
higher numbers than 63. Also, if epmd should get a file
descriptor with a number >= FD_SETSIZE, it will close a
the file descriptor and write a message to the log
(instead of mysteriously fail); the Erlang node that
tried to register will fail with a duplicate_name error
(unfortunately, epmd has no way to indicate to the Erlang
node why the register attempt failed).</p>
<p>Own Id: OTP-5716 Aux Id: seq10070 </p>
</item>
</list>
</section>
</section>
<section>
<title>ERTS 5.4.9</title>
<section>
<title>Fixed Bugs and Malfunctions</title>
<list type="bulleted">
<item>
<p>Timezone data is now initialized better. (was a problem
at least on NetBSD 2.0.2) Thanks to Rich Neswold.</p>
<p>Own Id: OTP-5621</p>
</item>
<item>
<p>The hybrid-heap emulator ('erl -hybrid') is much more
stable. We have corrected all known bugs that caused it
to dump core while running our test suites.</p>
<p>Own Id: OTP-5634</p>
</item>
<item>
<p>Fixed rare memory leaks in <c><![CDATA[erlang:demonitor/1]]></c> when
distributed monitors were removed.</p>
<p>Own Id: OTP-5692</p>
</item>
<item>
<p>Processes were sometimes unnecessarily garbage collected
when terminating. These unnecessary garbage collections
have now been eliminated.</p>
<p>Own Id: OTP-5693</p>
</item>
</list>
</section>
<section>
<title>Improvements and New Features</title>
<list type="bulleted">
<item>
<p>The <c><![CDATA[c:i/0]]></c> function will now run in a paged mode if
there are more than 100 processes in the system. (Thanks
to Ulf Wiger.)</p>
<p><c><![CDATA[erlang:system_info(process_count)]]></c> has
been optimized and does now return exactly the same value
as <c><![CDATA[length(processes())]]></c>. Previously
<c><![CDATA[erlang:system_info(process_count)]]></c> did not include
exiting processes which are included in
<c><![CDATA[length(processes())]]></c>.</p>
<p>The <c><![CDATA[+P]]></c> flag for <c><![CDATA[erl]]></c>, which sets the maximum
number of processes allowed to exist at the same, no longer
accepts values higher than 134217727. (You will still
probably run out of memory before you'll be able to reach
that limit.)</p>
<p>Own Id: OTP-5645 Aux Id: seq9984 </p>
</item>
<item>
<p>The term-building driver functions
<c><![CDATA[driver_output_term()]]></c> and <c><![CDATA[driver_send_term()]]></c>
have been updated:</p>
<p>The ERL_DRV_FLOAT type has been added.</p>
<p>For the ERL_DRV_BINARY type, the length and offset are
now validated against the length of the driver binary.</p>
<p>The ERL_DRV_PID type is now implemented (it was
documented, but not implemented).</p>
<p>Own Id: OTP-5674</p>
</item>
</list>
</section>
</section>
<section>
<title>ERTS 5.4.8</title>
<section>
<title>Fixed Bugs and Malfunctions</title>
<list type="bulleted">
<item>
<p><c><![CDATA[ets:delete/1]]></c> now allows other Erlang process to run
when a large table is being deleted.</p>
<p>Own Id: OTP-5572</p>
</item>
<item>
<p>A bug regarding tcp sockets which results in hanging
<c><![CDATA[gen_tcp:send/2]]></c> has been corrected. To encounter
this bug you needed one process that read from a socket,
one that wrote more date than the reader read out so the
sender got suspended, and then the reader closed the
socket. (Reported and diagnosed by Alexey Shchepin.)</p>
<p>Corrected a bug in the (undocumented and unsupported)
option <c><![CDATA[{packet,http}]]></c> for <c><![CDATA[gen_tcp.]]></c>
(Thanks to Claes Wikstrom and Luke Gorrie.)</p>
<p>Updated the documentation regarding the second argument to
<c><![CDATA[gen_tcp:recv/2]]></c>, the <c><![CDATA[Length]]></c> to receive.</p>
<p>Own Id: OTP-5582 Aux Id: seq9839 </p>
</item>
<item>
<p>Erlang/OTP will now build on Mac OS X 10.4 "Tiger" (the
problem was that 10.4 has a partially implemented poll()
function that can't handle devices). Also, on Mac OS X
10.3 "Panther", Erlang/OTP will now use select() instead
of poll() (because poll() on Mac OS X 10.3 is implemented
using select()).</p>
<p>Own Id: OTP-5595</p>
</item>
<item>
<p>A bug in the file driver when opening a file in
compressed mode, and the returned allocated pointer from
the compressing library was in very high memory (>= 2GB),
causing e.g. <c><![CDATA[file:read/2]]></c> to return
<c><![CDATA[{error,ebadf}]]></c>, has been corrected.</p>
<p>Own Id: OTP-5618</p>
</item>
</list>
</section>
<section>
<title>Improvements and New Features</title>
<list type="bulleted">
<item>
<p>The new fun <c><![CDATA[M:F/A]]></c> construct creates a fun that
refers to the latest version of <c><![CDATA[M:F/A.]]></c> This syntax is
meant to replace tuple funs <c><![CDATA[{M,F}]]></c> which have many
problems.</p>
<p>The new type test <c><![CDATA[is_function(Fun,A)]]></c> (which may be
used in guards) test whether <c><![CDATA[Fun]]></c> is a fun that can be
applied with <c><![CDATA[A]]></c> arguments. (Currently, <c><![CDATA[Fun]]></c>
can also be a tuple fun.)</p>
<p>Own Id: OTP-5584</p>
</item>
<item>
<p>In the HiPE application, there's a new experimental
register allocator (optimistic coalescing), and the
linear scan register allocator is now also available on
ppc. Plus lots of cleanups.</p>
<p>Minor hybrid heap corrections.</p>
<p>The maximum size of a heap used to be artificially
limited so that the size of a heap would fit in 28 bits;
that limitation could cause the emulator to terminate in
a garbage collection even if there still was available
memory. Now the largest heap size for a 32 bit CPU is
1,699,221,830 bytes. (Thanks to Jesper Wilhelmsson.)</p>
<p>Also removed the undocumented <c><![CDATA[+H]]></c> emulator option.</p>
<p>Own Id: OTP-5596</p>
</item>
</list>
</section>
</section>
<section>
<title>ERTS 5.4.7</title>
<section>
<title>Fixed Bugs and Malfunctions</title>
<list type="bulleted">
<item>
<p><c><![CDATA[to_erl]]></c> could close unexpectedly on Linux systems.
Also, output from the emulator could be lost. Corrected.</p>
<p>Own Id: OTP-5561</p>
</item>
</list>
</section>
<section>
<title>Improvements and New Features</title>
<list type="bulleted">
<item>
<p>The <c><![CDATA[cpu_timestamp]]></c> option for <c><![CDATA[erlang:trace/3]]></c>
is now also supported on Linux.</p>
<p>Own Id: OTP-5532 Aux Id: seq9813 </p>
</item>
<item>
<p>The last known <c><![CDATA[werl]]></c> window size/position is now saved
correctly when <c><![CDATA[werl]]></c> is stopped with the window
minimized. A problem with the placement not being saved if
the emulator is halted or stopped from the JCL menu has also
been fixed.</p>
<p>Own Id: OTP-5544 Aux Id: OTP-5522 </p>
</item>
</list>
</section>
</section>
<section>
<title>ERTS 5.4.6</title>
<section>
<title>Fixed Bugs and Malfunctions</title>
<list type="bulleted">
<item>
<p>Some math libraries do not always throw floating-point
exceptions on errors. In order to be able to use these
libraries, floating-point errors are now more thoroughly
checked.</p>
<p>Misc floating-point fixes for Linux and MacOSX.</p>
<p>Own Id: OTP-5467</p>
</item>
<item>
<p>An internal buffer was sometimes not cleared which caused
garbage to appear in error messages sent to the error
logger.</p>
<p><c><![CDATA[errno]]></c> was sometimes clobbered which caused erroneous
error reports about <c><![CDATA[poll()]]></c> errors.</p>
<p>Only emulators on unix platforms were affected by these
bugs.</p>
<p>Own Id: OTP-5492</p>
</item>
<item>
<p>The ethread library now works on OpenBSD.</p>
<p>Own Id: OTP-5515</p>
</item>
<item>
<p>Corrected a bug in the (undocumented and unsupported)
option <c><![CDATA[{packet,http}]]></c> for <c><![CDATA[gen_tcp]]></c>.
(Thanks to Claes Wikstrom and Luke Gorrie.)</p>
<p>Own Id: OTP-5519</p>
</item>
</list>
</section>
<section>
<title>Improvements and New Features</title>
<list type="bulleted">
<item>
<p><c><![CDATA[binary_to_term/1]]></c> could cause the emulator to crash
when given invalid pids or funs.</p>
<p>Own Id: OTP-5484 Aux Id: seq9801 </p>
</item>
<item>
<p>Some more stability problems were fixed in the
hybrid-heap emulator.</p>
<p>Own Id: OTP-5489</p>
</item>
<item>
<p>After <c><![CDATA[werl]]></c> was closed with the window minimized, it
was not possible to restart <c><![CDATA[werl]]></c> with an open
window. A temporary solution has so far been implemented
that restores the initial window settings every time
<c><![CDATA[werl]]></c> is started.</p>
<p>Own Id: OTP-5522</p>
</item>
</list>
</section>
</section>
<section>
<title>ERTS 5.4.5</title>
<section>
<title>Fixed Bugs and Malfunctions</title>
<list type="bulleted">
<item>
<p>If a process had node links (created by
<c><![CDATA[monitor_node/2]]></c>), executing
<c><![CDATA[process_info(Pid,memory)]]></c> for that process would
crash the emulator.</p>
<p>Own Id: OTP-5420</p>
</item>
<item>
<p>Minor corrections to the help text printed by <c><![CDATA[erlc -help]]></c>. The documentation for <c><![CDATA[erlc]]></c> was also
slightly updated.</p>
<p>Own Id: OTP-5428</p>
</item>
<item>
<p>32-bit words were used for offsets in the garbage
collector. This caused the emulator to crash on 64-bit
machines when heaps were moved more than 4 GB during
garbage collection.</p>
<p>Own Id: OTP-5430</p>
</item>
<item>
<p><c><![CDATA[is_boolean(42.5)]]></c> failed to load if optimization was
explicitly turned off.</p>
<p>Own Id: OTP-5448</p>
</item>
<item>
<p>If there was a call to <c><![CDATA[Module:foo/X]]></c> from any loaded
module, the returned by <c><![CDATA[M:module_info(exports)]]></c> would
always include <c><![CDATA[{foo,X}]]></c> (even though
<c><![CDATA[Module:foo/X]]></c> if was not defined).</p>
<p>Own Id: OTP-5450 Aux Id: seq9722 </p>
</item>
</list>
</section>
<section>
<title>Improvements and New Features</title>
<list type="bulleted">
<item>
<p>The <c><![CDATA[c]]></c> option for the <c><![CDATA[+B]]></c> flag has been
introduced which makes it possible to use Ctrl-C
(Ctrl-Break on Windows) to interrupt the shell process
rather than to invoke the emulator break handler. All new
<c><![CDATA[+B]]></c> options are also supported on Windows
(<c><![CDATA[werl]]></c>) as of now. Furthermore, Ctrl-C on Windows has
now been reserved for copying text (what Ctrl-Ins was used
for previously). Ctrl-Break should be used for break
handling. Lastly, the documentation of the system flags has
been updated.</p>
<p>Own Id: OTP-5388</p>
</item>
</list>
</section>
</section>
<section>
<title>ERTS 5.4.4</title>
<section>
<title>Fixed Bugs and Malfunctions</title>
<list type="bulleted">
<item>
<p>The function <c><![CDATA[erlang:float/1]]></c> can now be used in
match specifications and is recognized by
<c><![CDATA[dbg:fun2ms/1]]></c> and <c><![CDATA[ets:fun2ms/1]]></c>. This
addition is part of the work to "harmonize" match
specification guards with Erlang guards.</p>
<p>Own Id: OTP-5297 Aux Id: OTP-4927 </p>
</item>
<item>
<p>The <c><![CDATA[register/2]]></c> BIF would return <c><![CDATA[true]]></c> even if
the second argument was not a pid for living process.
Corrected to cause an exception.</p>
<p>Own Id: OTP-5324 Aux Id: seq9698 </p>
</item>
<item>
<p>In the 'bag' type of ets tables, elements having the same
key were supposed to be order in insertion order. The
would be wrong if a rehash occurred.</p>
<p>Own Id: OTP-5340 Aux Id: seq9704 </p>
</item>
<item>
<p>Linked in drivers in the Crypto and Asn1 applications
are now compiled with the <c><![CDATA[-D_THREAD_SAFE]]></c> and
<c><![CDATA[-D_REENTRANT]]></c> switches on unix when the emulator has
thread support enabled.</p>
<p>Linked in drivers on MacOSX are not compiled with the
undocumented <c><![CDATA[-lbundle1.o]]></c> switch anymore. Thanks to
Sean Hinde who sent us a patch.</p>
<p>Linked in driver in Crypto, and port programs in SSL, now
compiles on OSF1.</p>
<p>Minor makefile improvements in Runtime_Tools.</p>
<p>Own Id: OTP-5346</p>
</item>
<item>
<p>Fixed a bug in the hybrid heap in distributed send
operations.</p>
<p>Own Id: OTP-5361</p>
</item>
<item>
<p>A BIF <c><![CDATA[erlang:raise/3]]></c> has been added. See the manual
for details. It is intended for internal system programming
only, advanced error handling.</p>
<p>Own Id: OTP-5376 Aux Id: OTP-5257 </p>
</item>
<item>
<p>Mikael Pettersson (HiPE) corrected a few bugs in the
emulator that caused problems when compiled with the
experimental gcc-4.0.0.</p>
<p>Own Id: OTP-5386</p>
</item>
</list>
</section>
<section>
<title>Improvements and New Features</title>
<list type="bulleted">
<item>
<p>Minor update of the internal documentation of
the <c><![CDATA[epmd]]></c> protocol.</p>
<p>The listen port of <c><![CDATA[epmd]]></c> has now been registered at
IANA:
<url href="http://www.iana.org/assignments/port-numbers">http://www.iana.org/assignments/port-numbers</url>.</p>
<p>Own Id: OTP-5234</p>
</item>
<item>
<p><c><![CDATA[run_erl.c]]></c> now works on Mac OS X and FreeBSD.</p>
<p>Own Id: OTP-5384</p>
</item>
<item>
<p>A few bugs were corrected in the HiPE application.</p>
<p>Own Id: OTP-5385</p>
</item>
</list>
</section>
</section>
<section>
<title>ERTS 5.4.3</title>
<section>
<title>Fixed Bugs and Malfunctions</title>
<list type="bulleted">
<item>
<p>Font and color preferences for <c><![CDATA[werl.exe]]></c> now can be
saved even after the first time you run <c><![CDATA[werl.exe]]></c>.
The window position and size is also saved. Patch from
James Hague who did all the hard work.</p>
<p>Own Id: OTP-5250</p>
</item>
<item>
<p>OTP archive libraries, e.g. the <c><![CDATA[erl_interface]]></c>
library, on MacOSX could not be used without first
rerunning <c><![CDATA[ranlib]]></c> on them. This bug has now been
fixed.</p>
<p>Own Id: OTP-5274</p>
</item>
<item>
<p>Bugs in <c><![CDATA[erlang:hash/2]]></c> and <c><![CDATA[erlang:phash/2]]></c> on
64-bit platforms have been fixed.</p>
<p>Own Id: OTP-5292</p>
</item>
</list>
</section>
<section>
<title>Improvements and New Features</title>
<list type="bulleted">
<item>
<p>Corrected a crash in the 64-bits emulator.</p>
<p>Corrected a problem in the hybrid heap emulator.</p>
<p>In the chapter about the abstract format in the ERTS User's
Guide, updated the last section about how the abstract
format is stored in BEAM files.</p>
<p>Own Id: OTP-5262</p>
</item>
<item>
<p>The maximum number of concurrent threads in the internal
ethread thread package has been limited to 2048.</p>
<p>Own Id: OTP-5280</p>
</item>
</list>
</section>
</section>
<section>
<title>ERTS 5.4.2.1</title>
<section>
<title>Improvements and New Features</title>
<list type="bulleted">
<item>
<p>If Erlang/OTP was installed in a short directory name,
such as <c><![CDATA[C:\Program\erl5.4.2]]></c>, the emulator would not
start.</p>
<p>Own Id: OTP-5254</p>
</item>
</list>
</section>
</section>
<section>
<title>ERTS 5.4.2</title>
<section>
<title>Fixed Bugs and Malfunctions</title>
<list type="bulleted">
<item>
<p>If one used <c><![CDATA[select/3]]></c> and <c><![CDATA[select/1]]></c> on a
non-fixed ETS table and deleted objects simultaneously,
the emulator could crash. Note that the result of such
simultaneous operations on tables that are not in a fixed
state is still undefined, but the emulator crash is,
needless to say, fixed.</p>
<p>Own Id: OTP-5209 Aux Id: seq9198 </p>
</item>
<item>
<p>Arithmetic with big numbers could crash the emulator.</p>
<p>The HiPE native code compiler and run-time code in the
emulator has been updated. (Note: Native code is still
not supported.)</p>
<p>Eliminated a few bugs that could crash the hybrid emulator
(which is not supported).</p>
<p>Own Id: OTP-5233 Aux Id: seq9587 </p>
</item>
</list>
</section>
</section>
<section>
<title>ERTS 5.4.1</title>
<section>
<title>Fixed Bugs and Malfunctions</title>
<list type="bulleted">
<item>
<p>The ethread library was unnecessarily rebuilt multiple
times during the build process, also a debug version of
the library was build during the install phase. These
unnecessary builds have now been removed. Note, the
content of the installed Erlang/OTP system is not
effected at all by this change.</p>
<p>Own Id: OTP-5203</p>
</item>
<item>
<p>The emulator could fail to clear the memory segment
cache. This could potentially cause memory allocation to
unnecessarily fail when memory usage was close to its
maximum. This bug has now been fixed.</p>
<p>Own Id: OTP-5211</p>
</item>
<item>
<p>std_alloc (std short for standard) was sometimes called
def_alloc (def short for default). std_alloc is now
everywhere referred to as std_alloc.</p>
<p>Own Id: OTP-5216</p>
</item>
<item>
<p>A documentation bug has been corrected in
the <c><![CDATA[erts_alloc(3)]]></c> documentation. It was stated that
some of the memory allocators present were by default
disabled. This is true for Erlang/OTP R9C, but is not true
for Erlang/OTP R10B. In R10B all memory allocators present
are enabled by default.</p>
<p>Own Id: OTP-5217</p>
</item>
</list>
</section>
<section>
<title>Improvements and New Features</title>
<list type="bulleted">
<item>
<p>The emulator now close all open files and sockets
immediately after receiving an USR1 signal. This causes
the emulator to unregister at <c><![CDATA[epmd]]></c> as early as
possible.</p>
<p>Own Id: OTP-5221 Aux Id: OTP-4985, seq9514 </p>
</item>
<item>
<p>Try/catch support in the emulator slightly updated.</p>
<p>Own Id: OTP-5229</p>
</item>
</list>
</section>
</section>
</chapter>
