path: root/erts/doc/src/notes_history.xml

<?xml version="1.0" encoding="utf-8" ?>
<!DOCTYPE chapter SYSTEM "chapter.dtd">

<chapter>
  <header>
    <copyright>
      <year>2006</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 History</title>
    <prepared>otp_appnotes</prepared>
    <docno>nil</docno>
    <date>nil</date>
    <rev>nil</rev>
    <file>notes_history.xml</file>
  </header>

  <section>
    <title>ERTS 5.4</title>

    <section>
      <title>Fixed Bugs and Malfunctions</title>
      <list type="bulleted">
        <item>
          <p>The BIF <c><![CDATA[is_record(Record,RecordTag)]]></c> has been added to
            the run-time system. It checks that <c><![CDATA[Record]]></c> is a tuple
            whose first element is the atom <c><![CDATA[RecordTag.]]></c> No check
            is made of the size of the tuple (because the run-time
            system doesn't know anything about records).</p>
          <p>Note that normally the compiler translates calls to
            <c><![CDATA[is_record/2]]></c> to code that also verify the size of the
            tuple, in addition to verifying the first element. The
            BIF version will be used <c><![CDATA[is_record/2]]></c> is applied or
            if the second argument is not a literal atom (e.g. a
            variable or another term type, in which case the BIF will
            generate a <c><![CDATA[badarg]]></c> exception).</p>
          <p>Own Id: OTP-4812</p>
        </item>
        <item>
          <p>Guards of mach specifications are corrected to resemble
            the semantics of guards in real code more closely. The
            implementation now corresponds to the documentation in
            ERTS User's Guide. The following things are corrected:</p>
          <list type="bulleted">
            <item>Guard semantics was wrong when it came to logical
             operators and exceptions.
            <c><![CDATA[{'or',{'is_integer','$1'},{'or','$1','$1'}}]]></c>
             evaluated to <c><![CDATA[true]]></c> with <c><![CDATA['$1']]></c> bound to an
             integer.</item>
            <item>Unary + and - was not implemented.</item>
            <item>Calling operators as Bif's was not supported by
            <c><![CDATA[ets/dbg:fun2ms]]></c> (<c><![CDATA[erlang:'or'(A,B)]]></c> etc).</item>
            <item>Old typetests (like <c><![CDATA[integer(X)]]></c> instead of
            <c><![CDATA[is_integer(X))]]></c> was not supported by
            <c><![CDATA[ets/dbg:fun2ms]]></c>.</item>
            <item>Semicolon (;) in guards was not supported by
            <c><![CDATA[ets/dbg:fun2ms]]></c>.</item>
          </list>
          <p>*** POTENTIAL INCOMPATIBILITY ***</p>
          <p>Own Id: OTP-4927</p>
        </item>
        <item>
          <p>A potential initialization failure when using threads and
            elib_malloc has been removed.</p>
          <p>Own Id: OTP-5125</p>
        </item>
        <item>
          <p>Several problems in the 64-bit emulator has been
            corrected. For instance, the emulator could crash while
            running the Debugger.</p>
          <p>Own Id: OTP-5146</p>
        </item>
        <item>
          <p>The match spec parse transform <c><![CDATA[ms_transform]]></c> no
            longer accepts the <c><![CDATA[andalso]]></c> and <c><![CDATA[orelse]]></c>
            constructs in guards for consistency with the standard
            Erlang language. A future release of Erlang/OTP may allow
            <c><![CDATA[andalso]]></c> and <c><![CDATA[orelse]]></c> in guards.</p>
          <p>*** POTENTIAL INCOMPATIBILITY ***</p>
          <p>Own Id: OTP-5149</p>
        </item>
        <item>
          <p>In rare circumstances in a process that has caught
            exceptions and uses funs, the process would be killed
            when changing code because the code server would think
            that the process still held references to the funs.</p>
          <p>Own Id: OTP-5153</p>
        </item>
        <item>
          <p><c><![CDATA[erlang:system_monitor/2]]></c> no longer sends any
            monitoring messages to the system monitor process from
            itself. This behavior is more consistent with other trace
            functionality.</p>
          <p>*** POTENTIAL INCOMPATIBILITY ***</p>
          <p>Own Id: OTP-5183</p>
        </item>
      </list>
    </section>

    <section>
      <title>Improvements and New Features</title>
      <list type="bulleted">
        <item>
          <p>The documentation stated that <c><![CDATA[ets:safe_fixtable/2]]></c>
            could return <c><![CDATA[true]]></c> or <c><![CDATA[false,]]></c> that was wrong,
            it always returns <c><![CDATA[true]]></c>.</p>
          <p>Own Id: OTP-4830</p>
        </item>
        <item>
          <p>The unary '+' operator has been changed to throw an
            <c><![CDATA[badarith]]></c> exception if its argument is not numeric (or
            fail in a guard). It used its argument unchanged whatever
            the type. Given the new meaning, unary '+' can now be
            used to test whether a term is numeric.</p>
          <p>*** POTENTIAL INCOMPATIBILITY ***</p>
          <p>Own Id: OTP-4928</p>
        </item>
        <item>
          <p>Process identifiers and port identifiers have been
            made more unique. Previously 18 bits were used as id in
            the internal representation of process and port
            identifiers. Now 28 bits are used.</p>
          <p>The maximum limit on the number of concurrently existing
            processes due to the representation of pids has been
            increased to 268435456 processes. The same is true for
            ports. This limit will at least on a 32-bit architecture be
            impossible to reach due to memory shortage.</p>
          <p><em>NOTE:</em> By default, the <c><![CDATA[ERTS]]></c>, and the
            <c><![CDATA[erl_interface]]></c>, <c><![CDATA[ei]]></c>, and <c><![CDATA[jinterface]]></c>
            libraries are now only guaranteed to be compatible with
            other Erlang/OTP components from the same release. It is
            possible to set each component in compatibility mode of
            an earlier release, though. See the documentation for
            respective component on how to set it in compatibility
            mode.</p>
          <p>*** POTENTIAL INCOMPATIBILITY ***</p>
          <p>Own Id: OTP-4968 Aux Id: OTP-4196 </p>
        </item>
        <item>
          <p>A new internal thread library for the ERTS has been
            added.</p>
          <p>Own Id: OTP-5048</p>
        </item>
        <item>
          <p>The system's performance could degrade severely if one
            process held numerous links or monitors. The issue is
            resolved.</p>
          <p>Own Id: OTP-5079</p>
        </item>
        <item>
          <p>A new function, <c><![CDATA[string:to_integer/1]]></c>, has been added.</p>
          <p>Own Id: OTP-5081 Aux Id: OTP-5136 </p>
        </item>
        <item>
          <p>A new function, <c><![CDATA[string:to_float/1]]></c>, has been added.</p>
          <p>Own Id: OTP-5136 Aux Id: OTP-5081 </p>
        </item>
        <item>
          <p>The exception code for calling a fun with wrong number of
            arguments has been changed from simply <c><![CDATA[badarity]]></c> to
            <c><![CDATA[{badarity,{Fun,Args}}]]></c>.</p>
          <p>*** POTENTIAL INCOMPATIBILITY ***</p>
          <p>Own Id: OTP-5139</p>
        </item>
        <item>
          <p>The long-awaited <c><![CDATA[try]]></c>...<c><![CDATA[catch]]></c> construction
            is included in this release. However, its use in
            production code is not yet supported as there are several
            known cases of legal code crashing the compiler. We plan
            to release a patch to the compiler (including the
            documentation) and at that time <c><![CDATA[try]]></c>...<c><![CDATA[catch]]></c>
            will be supported.</p>
          <p>Own Id: OTP-5150</p>
        </item>
      </list>
    </section>
  </section>

  <section>
    <title>ERTS 5.3.6.6</title>

    <section>
      <title>Fixed Bugs and Malfunctions</title>
      <list type="bulleted">
        <item>
          <p>A bug that caused an emulator crash when using system
            monitor of long GC has been fixed.</p>
          <p>Own Id: OTP-5123</p>
        </item>
      </list>
    </section>
  </section>

  <section>
    <title>ERTS 5.3.6.5</title>

    <section>
      <title>Fixed Bugs and Malfunctions</title>
      <list type="bulleted">
        <item>
          <p><c><![CDATA[erlang:memory(ets)]]></c> and <c><![CDATA[c:memory(ets)]]></c>
            sometimes reported erroneous values. This bug has now been
            fixed.</p>
          <p>Own Id: OTP-5115 Aux Id: seq9063 </p>
        </item>
        <item>
          <p>There is now a packet size limit option for <c><![CDATA[gen_tcp]]></c>
            sockets. See the manual for <c><![CDATA[inet:setopts/2]]></c>.</p>
          <p>The ASN.1 BER packet decoding for <c><![CDATA[gen_tcp]]></c> sockets
            can now decode indefinite length packets.</p>
          <p>Own Id: OTP-5128</p>
        </item>
      </list>
    </section>

    <section>
      <title>Improvements and New Features</title>
      <list type="bulleted">
        <item>
          <p>Port index was unnecessarily incremented by port table
            size when port table got full. This unnecessary increment
            has now been removed.</p>
          <p>Own Id: OTP-5119</p>
        </item>
      </list>
    </section>
  </section>

  <section>
    <title>ERTS 5.3.6.3</title>

    <section>
      <title>Fixed Bugs and Malfunctions</title>
      <list type="bulleted">
        <item>
          <p>Resolved a build problem on Mac OS 10.3 ("Panther").
            Because of a conflict with the zlib sources included in
            the Erlang run-time system and the zlib library included
            in Panther, linking would fail.</p>
          <p>Minor optimization on all Unix systems: caching the
            system name returned from the uname() system call.
            (Thanks to David N. Welton.)</p>
          <p>Own Id: OTP-5069</p>
        </item>
      </list>
    </section>

    <section>
      <title>Improvements and New Features</title>
      <list type="bulleted">
        <item>
          <p>The ability to set system wide options for TCP sockets is
            added through the kernel application variables
            <c><![CDATA[inet_default_listen_options]]></c> and
            <c><![CDATA[inet_default_connect_options]]></c>, see the <c><![CDATA[inet]]></c>
            manual page for details.</p>
          <p>Own Id: OTP-5080</p>
        </item>
      </list>
    </section>
  </section>

  <section>
    <title>ERTS 5.3.6.2</title>

    <section>
      <title>Fixed Bugs and Malfunctions</title>
      <list type="bulleted">
        <item>
          <p>A few portability enhancements for the R9C-1 Open Source
            release: The installer for Windows can now be built with
            NSIS 2.0 (as well as with the NSIS 2.0b3). The driver
            header files updated to allowed drivers to be built which
            the MinGW compiler on Windows. Minor portability
            enhancement in <c><![CDATA[io_lib:fread]]></c>.</p>
          <p>Own Id: OTP-4789</p>
        </item>
        <item>
          <p>Conversion of extremely small floating point numbers in
            the external format (distribution) could sometimes fail
            leading to unexpected closing of distribution channels,
            i.e. generating nodedowns for healthy nodes.</p>
          <p>Own Id: OTP-5026 Aux Id: seq8631 EABln12478 </p>
        </item>
      </list>
    </section>
  </section>

  <section>
    <title>ERTS 5.3.6.1</title>

    <section>
      <title>Fixed Bugs and Malfunctions</title>
      <list type="bulleted">
        <item>
          <p>Conversion of extremely small floating point numbers in
            the external format (distribution) could sometimes fail
            leading to unexpected closing of distribution channels,
            i.e. generating nodedowns for healthy nodes.</p>
          <p>Own Id: OTP-5026 Aux Id: seq8631 EABln12478 </p>
        </item>
      </list>
    </section>

    <section>
      <title>Improvements and New Features</title>
      <list type="bulleted">
        <item>
          <p>Remote spawn on a nonreachable node now gives warning
            instead of error in the error_log.</p>
          <p>Own Id: OTP-5030 Aux Id: seq8663] </p>
        </item>
      </list>
    </section>

    <section>
      <title>Known Bugs and Problems</title>
      <list type="bulleted">
        <item>
          <p>Emulator with elib_malloc enabled could hang when many
            I/O threads were in use.</p>
          <p>Own Id: OTP-5028 Aux Id: EABln13041, EABln12253 </p>
        </item>
      </list>
    </section>
  </section>

  <section>
    <title>ERTS 5.3.6</title>

    <section>
      <title>Fixed Bugs and Malfunctions</title>
      <list type="bulleted">
        <item>
          <p>Distributed monitoring in combination with nodes
            restarting did not behave correctly in rare
            circumstances.</p>
          <p>Own Id: OTP-4914</p>
        </item>
        <item>
          <p>A module containing code similar to <c><![CDATA[receive after 4.1]]></c>
            could not be loaded.</p>
          <p>Own Id: OTP-4963 Aux Id: seq8344 </p>
        </item>
        <item>
          <p>Problems fixed in <c><![CDATA[file:open/2]]></c>: <c><![CDATA["/dev/null"]]></c>
            can now be opened. Opening a FIFO will now return an error
            instead of hanging the emulator. The documentation has been
            updated to point out that <c><![CDATA[file:open/2]]></c> returns
            the error code <c><![CDATA[eisdir]]></c> when the pathname is not a
            regular file (the pathname is not necessarily a directory).</p>
          <p>Own Id: OTP-4992</p>
        </item>
      </list>
    </section>

    <section>
      <title>Improvements and New Features</title>
      <list type="bulleted">
        <item>
          <p>The Solaris kernel poll feature was changed from a
            compile time option to a runtime option. The kernel poll
            feature can be enabled, by passing the command-line
            argument <c><![CDATA[+K true]]></c> to an emulator (see <c><![CDATA[erl(1)]]></c>)
            that have kernel poll support, i.e. an emulator for
            Solaris 8. By default the kernel poll feature is disabled.</p>
          <p>Own Id: OTP-4979 Aux Id: seq8478 </p>
        </item>
        <item>
          <p>Before the Erlang emulator writes an <c><![CDATA[erl_crash.dump]]></c>
            file (for any reason), it will close all open files and
            sockets.</p>
          <p>Own Id: OTP-4985 Aux Id: EABln10730, EABln11277,
            EABln11279 </p>
        </item>
        <item>
          <p>The <c><![CDATA[+c]]></c> switch has been added to disable time
            correction in the runtime system, this should be used on
            systems where one is certain no dramatic wall clock time
            changes will occur and the time correction algorithm is too
            costly (namely very fast Linux systems where loads of
            <c><![CDATA[erlang:now()]]></c> are executed).</p>
          <p>Own Id: OTP-4986</p>
        </item>
        <item>
          <p>The <c><![CDATA[process_flag/2,3]]></c> BIFs now take a flag
            <c><![CDATA[min_heap_size]]></c> that allows changing an existing
            process's minimum heap size. The actual size will only be
            changed when the next garbage collection occurs.</p>
          <p>Own Id: OTP-4991 Aux Id: seq8515, OTP-4987 </p>
        </item>
      </list>
    </section>
  </section>

  <section>
    <title>ERTS 5.3.4</title>

    <section>
      <title>Improvements and New Features</title>
      <list type="bulleted">
        <item>
          <p>A possibility to make distribution messages be qued up
            during running of erlang code, so that larger packages is
            sent over the network is added.</p>
          <p>Own Id: OTP-4916</p>
        </item>
        <item>
          <p>When code loading failed it was impossible to know
            exactly what caused it, only <c><![CDATA[{undef,[{M,F,A}|...]}]]></c>
            would be reported. Now the primitive loader lets the
            error logger print an error report if a file operation
            fails. All file errors except <c><![CDATA[enoent]]></c> and
            <c><![CDATA[enotdir]]></c> are reported this way.</p>
          <p>Own Id: OTP-4925 Aux Id: OTP-4952 </p>
        </item>
      </list>
    </section>
  </section>

  <section>
    <title>ERTS 5.3.3</title>

    <section>
      <title>Fixed Bugs and Malfunctions</title>
      <list type="bulleted">
        <item>
          <p>The driver for dynamically linked in drivers has been
            fixed to delete loaded drivers when its Erlang server
            dies. The Erlang server has also been updated to improve
            the start-on-demand behaviour.</p>
          <p>Own Id: OTP-4876 Aux Id: OTP-4855 seq8272 </p>
        </item>
        <item>
          <p><c><![CDATA[erlang:register/2]]></c> does no longer generate an ERROR
            REPORT to the error logger when the name already is
            registered. If the name is already registered the process
            function will crash with <c><![CDATA[{'EXIT',Reason}]]></c> and that is
            enough. It is up to the caller to decide if it is an
            error that the name is already registered.</p>
          <p>Own Id: OTP-4892</p>
        </item>
        <item>
          <p>When using <c><![CDATA[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 has now been
            corrected.</p>
          <p>Own Id: OTP-4903 Aux Id: seq8379 </p>
        </item>
      </list>
    </section>

    <section>
      <title>Improvements and New Features</title>
      <list type="bulleted">
        <item>
          <p>Starting Erlang with the <c><![CDATA[+Bi]]></c> flag (to ignore ^C), now
            also disables the quit ('q') option in the JCL menu.</p>
          <p>Own Id: OTP-4897</p>
        </item>
      </list>
    </section>
  </section>

  <section>
    <title>ERTS 5.3.2</title>

    <section>
      <title>Improvements and New Features</title>
      <list type="bulleted">
        <item>
          <p>The elib_malloc alternative dynamic memory allocator has
            been improved to use an address order best fit strategy.
            The instrumented emulator has been improved to be able to
            catch memory allocations done by external libraries. The
            emulator flag <c><![CDATA[+r]]></c> (stands for "relocate") makes
            <c><![CDATA[ets]]></c> updates always result in an object relocation,
            which significantly lessens the memory fragmentation in
            certain systems. The <c><![CDATA[erlang:system_info/1]]></c> bif can
            now be called with the argument <c><![CDATA[ets_realloc_moves]]></c>
            and will return <c><![CDATA[true]]></c> if the <c><![CDATA[+r]]></c> emulator flag
            is in effect, <c><![CDATA[false]]></c> otherwise.</p>
          <p>Own Id: OTP-4838 Aux Id: seq8156 </p>
        </item>
      </list>
    </section>
  </section>
</chapter>