| Age | Commit message (Collapse) | Author |
|---|
|
* rickard/ethread-rewrite/OTP-8544:
Rewrite ethread library
|
|
Large parts of the ethread library have been rewritten. The
ethread library is an Erlang runtime system internal, portable
thread library used by the runtime system itself.
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
`{read_concurrency, true}' option upon table creation. See the
documentation of `ets:new/2' for more information.
The ethread library can now also use the libatomic_ops library
for atomic memory accesses. This makes it possible for the
Erlang runtime system to utilize optimized atomic operations
on more platforms than before. Use the
`--with-libatomic_ops=PATH' configure command line argument
when specifying where the libatomic_ops installation is
located. The libatomic_ops library can be downloaded from:
http://www.hpl.hp.com/research/linux/atomic_ops/
The changed API of the ethread 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 ethread library.
Note: When building for x86, the ethread 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
`--enable-ethread-pre-pentium4-compatibility' configure command
line argument when configuring the system.
|
|
* sverker/one_offheap_list/OTP-8737:
One off-heap list, to eliminate two words per ETS object.
|
|
Merging the three off-heap lists (binaries, funs and externals) into
one list. This reduces memory consumption by two words (pointers) per
ETS object.
|
|
* sverker/ets_select_hang/OTP-8732:
Fix ets:select hanging on ordered_set with empty list as key.
|
|
erl_db_tree.c incorrectly used NIL (empty list) as "lastkey" to mark
start of the iteration. A real NIL key could then cause a
select or match iteration to be restarted over and over again if the
last key before a trap happended to be NIL.
Changed NIL to THE_NON_VALUE. Should be ok as the initial key value can
never be put into any continuation tuple.
|
|
* cf/timer_tc:
Add timer:tc/2 to measure the elapsed time of anonymous functions
Conflicts:
lib/stdlib/doc/src/timer.xml
|
|
warn_unused_import works correctly (does not give warnings when overridden).
Local call in guard gives its own error pointing out the local/imported
function.
Use of the phrase "overridden auto-imported bif" instead of "redefined
auto-imported bif" in textual error messages.
|
|
The Erlang code preprocessor (epp) did not correctly handle premature
end-of-input when defining macros. This bug, introduced in STDLIB 1.16, has
been fixed.
|
|
|
|
|
|
Wrote and changed some tests in stdlib:erl_lint_SUITE
nowarn_bif_clash is obsoleted but will remove warnings and errors about bif
clashes. The recommended way is to use no_auto_import directives instead.
Hopefully erlang.erl is the only user in the world of nowarn_bif_clash.
|
|
Expressions evaluating to integers can now be used in types and function
specifications where hitherto only integers were allowed
("Erlang_Integer").
|
|
The erl_expand_records compiler pass translates the
following code:
h(X) when X#r1.a =:= 1 -> ok.
to (essentially):
h({r1,V1,V2}=X) when element(2, X) =:= 1 -> ok.
Since the guard can only be executed when the pattern matching
has succeeded, we know that the second element in the tuple X
must have been bound to V2. Thus we can eliminate the call
to element/2 like this:
h({r1,V1,V2}=X) when V1 =:= 1 -> ok.
|
|
The Erlang scanner has been augmented with two new tokens: .. and ....
|
|
The gen_fsm, gen_server, and wx_object format_status implementations
fail to handle global names of the form {global, term()} where term()
is something other than an atom, pid, or list. Change these
format_status implementations to treat names that are atoms, pids, or
lists as before, but for all other terms, set the header property of
the function return value to a tuple whose first element is a string
describing the return value and whose second element is the name term.
Add unit tests for gen_server and gen_fsm to verify sys:get_status
calls work successfully for globally registered instances.
|
|
* sv/format_status_error_info:
Add support for the format_status callback to gen_event
Extend format_status for gen_server/gen_fsm termination error logging
OTP-8630 sv/format_status_error_info
When a gen_server, gen_fsm process, or gen_event terminates abnormally,
sometimes the text representation of the process state can occupy many
lines of the error log, depending on the definition of the state term. A
mechanism to trim out parts of the state from the log has been added (using
a format_status/2 callback). See the documentation.
|
|
Correct behaviour of copy/2 witn 0 copies.
|
|
Even more tests added to cover more error cases, some tests
made easier on systems with small memories and timeouts made longer
due to timetraps on solaris daily builds.
|
|
Add testcases and clean up to get better code coverage.
|
|
|
|
Add the gc_bif's to the VM.
Add infrastructure for gc_bif's (guard bifs that can gc) with two and.
three arguments in VM (loader and VM).
Add compiler support for gc_bif with three arguments.
Add compiler (and interpreter) support for new guard BIFs.
Add testcases for new guard BIFs in compiler and emulator.
|
|
Add testcases for encode/decode_unsigned/1,2.
|
|
Add testcases for referenced_byte_size/1.
Add failure tests for referenced_byte_size.
|
|
Add testcases for binary:list_to_bin/1 and binary:copy/1,2.
Add reference implementation of list_to_bin/1.
|
|
Add testcases for bin_to_list.
Teach binref.erl bin_to_list.
|
|
Add allcoator parameter to erts_get_aligned_binary_bytes_extra.
Add testcases for the functions above.
Add reference implementation for the functions above.
|
|
Change name of the 'scope' option for binary:match/matches.
Add split and replace to binary.erl.
Cleanup comments etc in binary.erl and atom.names
Add testcases for part, split, replace and scopes.
|
|
Set loop factors to 10.
Teach erts_debug:set_internal_state to limit loop factor for binary.
Add random tests for matches and match with multiple searchstrings.
|
|
Fix heap-hole when trapping in binary.c
Fix boyer more segfaulting when searchstring is longer than haystack
|
|
Add testcase embryos and reference implementation.
Change name of compile function according to EEP31.
|
|
The gen_server and gen_fsm behaviors support the format_status
callback to allow developers to specialize how callback module state
appears within the return value of sys:get_status and within logged
output resulting from abnormal process termination. This patch adds
similar support to gen_event.
Event handlers that export a format_status/2 function, which is an
optional callback, and are registered with an event manager will now
have their format_status callbacks invoked when sys:get_status is
called on the event manager. The term returned from format_status for
this case replaces the default handler state in the sys:get_status
return value.
This patch also extends gen_event to call an event handler's
format_status function (if it exports one) should the handler
terminate abnormally. The term returned from format_status is logged
in place of the handler's state. This is intended to allow developers
to control how much output is logged in the case of abnormal
termination.
The documentation is appropriately extended and new unit tests are
added to cover the new gen_event format_status functionality.
|
|
When a gen_server or gen_fsm process terminates abnormally, sometimes
the text representation of the process state can occupy many lines of
the error log, depending on the definition of the state
term. Developers sometimes would like a way to trim out parts of the
state from the log if those parts don't contribute to documenting the
circumstances of the error, thereby helping to reduce the amount of
logged output.
Since the format_status callback can already format and specialize
gen_server and gen_fsm state for inclusion in the term returned from
sys:get_status, this patch extends format_status in a
backward-compatible way to also be able to specialize the state logged
for abnormal gen_server and gen_fsm termination, and includes new unit
tests to verify the new functionality.
This patch also eliminates the previous restriction that the status
returned by format_status must be a list. It can now be any term.
The documentation is extended to cover the new usage for
format_status, and it's been improved to recommend a form for the
normal case that allows the returned status to fit well with the rest
of the term returned by sys:get_status. The documentation is clear
that this form is only recommended, not required.
|
|
* bg/opt-receive:
Test that gen_server:call/2,3 are fast even with a huge message queue
erts: Add tests for the receive optimization
Update primary bootstrap
erts: Implement recv_mark/1 and recv_set/1 for real
compiler tests: Cover the error handling code in beam_receive
compiler test: Test optimization of receive statements
Optimize selective receives in the presence of a large message queue
Introduce the new recv_mark/1 and recv_mark/1 instructions
Compile tests that communicate with R12 nodes with the r12 option
Move p_run/2 to test_lib
gen: Inline wait_resp_mon/2 to help the compiler optimize
OTP-8623 bg/opt-receive
reveive 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 gen_server:call(). (See gen:do_call/4
for an example of a receive statement that will be optimized.)
|
|
|
|
R12 nodes cannot load code that use the optimized receive that
we are about to implement.
|
|
|
|
The function erl_scan:reserved_word/1 no longer returns true when given the
word spec. This bug was introduced in STDLIB-1.15.3 (R12B-3).
|
|
|
|
Works like timer:tc/3 but for anonymous functions.
|
|
When defining macros the closing right parenthesis before the dot is now
mandatory.
|
|
The Erlang parser no longer duplicates the singleton type undefined in the
type of record fields without initial value.
|
|
Both reltool and rebar needs to parse escripts. They are currently
using an undocumented function called escript:foldl/3. It folds a
function over all files in the body of an escript. If the body
contains source code the function compiles it and the gives debug
compiled beam code to the fold fun. If the body is an archive the fun
is applied for all files in the archive.
Instead of making the undocumented function public, the new functions
escript:create/2 and escript:extract/2 has been introduced. Together
with the new zip:foldl/3 function they have the same functionality as
escript:foldl/3 in a more flexible and generic way.
escript:foldl/3 should be removed as soon as reltool and rebar has
been adopted to use the new functions. The simplest way for reltool
and rebar to do this is to just copy the code from
escript_SUITE:escript_foldl/3, which happens to provide a future
compatible implementation of an emulated escript:foldl/3 function.
I was quite hesitant when I introduced the compile_source option. It
feels that it does not belong there but the alternative felt worse.
The rationale for the compile_source option is that it is a bit
cumbersome to compile the source code, as the source in most cases is
partial. In order to do compile the source you need to know about some
internals in escript. Without compile_source I think that these
internals should be documented. Further you need to duplicate parts of
the code.
Without the compile_source option you need to first parse the source
to forms, using an undocumented function in epp with an extended
format of predefined macros which also is undocumented. Then you need
to investigate the forms to see if you need to add an export form for
main. When that is done you can run the rest of the compiler passes as
usual. It is not so much code (60 lines or so) to write, but I do not
want to urge people to write it. I actually wrote the code (see
escript_SUITE:escript_foldl/3) before I decided to introduce the
compile_source option.
|
|
This is the public interface of prim_zip:open/3,
which has been used in earlier releases by both
erl_prim_loader and escript. The new function
can be used as a replacement for the undocumented
function escript:foldl/3 that is likely to be
removed without further notice.
The error handling of prim_zip:open/3 (and
prim_zip:foldl/3) has been improved in order to
better suite a public interface. For example it
could happen that a file or a zlib port could be
left open in some errors cases.
The documentation of the FileSpec parameter to
zip:create/3 has been updated to show that
file info can be explicitly specified. A FileSpec
may contain {Filename, binary(), #file_info{}}
elements. The function zip:create/3 was already
prepared to partly support this, but now after
a few minor fixes it is fully supported.
|
|
|
|
The empty record (no fields) is now considered typed.
It is more consistent than before; the base case is
the logical one.
A record is typed iff all its fields are typed.
A record is tagged 'typed' iff it is typed.
|
|
The abstract type 'fun' could not be printed. This bug has been fixed.
|
|
The Erlang code preprocessor (epp) sent extra messages on the form
{eof,Location} to the client when parsing the file attribute. This bug,
introduced in R11B, has been fixed.
|
|
* ms/pcre-security:
Fix CVE-2008-2371 (outer level option with alternatives caused crash).
OTP-8438 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.)
|
|
* jn/supervisor_child_count_only:
Add count_children/1 to supervisor.erl to determine the number of
OTP-8436 Added supervisor:count_children/1 to count the number of children
being managed without the memory impact of which_children/1.
(Thanks to Jay Nelson.)
|
