CVS: sml-dist/READMES 110.53-README,NONE,1.1 110.53-README.html,NONE,1.1
Matthias Blume <[email protected]>
| Newsgroups | gmane.comp.lang.sml.smlnj.commits |
|---|---|
| Message-ID | <[email protected]> |
Update of /cvsroot/smlnj/sml-dist/READMES
In directory sc8-pr-cvs1.sourceforge.net:/tmp/cvs-serv3575
Added Files:
110.53-README 110.53-README.html
Log Message:
readmes for 110.53
--- NEW FILE: 110.53-README ---
S M L / N J
1 1 0 . 5 3 N E W S
February 25, 2005
WARNING
This working version is believed to be stable, but
we have not run our full regression-test suite.
HOME: http://www.smlnj.org/index.html
FILES: http://smlnj.cs.uchicago.edu/dist/working/110.53/
Summary:
Bugfixes and minor feature improvements.
--------------------------------------------------------------------------
Details:
Basis
-----
- SMLofNJ.Susp is back.
CML
---
- Thread-safe version of structure Atom matches original signature
and re-uses the original implementation. Moreover, since an mvar
is used as a lock (instead of a separate server thread) to protect
the global hashtable, it is now possible to create atoms even before
RunCML.doit has started executing.
Library
-------
- Moved source tree from smlnj-lib/TraceDebugProf to a new
tools/TraceDebugProf.
NLFFI
-----
- Handles access to signed and unsigned "long long" data.
- Handles "long long" function arguments and results (but so far
only on the PowerPC).
- Some organizational cleanup of the library code.
ML-Yacc
-------
- Fixed formatting bug that sometimes caused syntactically incorrect
code from being formed.
Runtime
-------
- Fixed bug in PC-sampling profiler.
- Fixed off-by-1 bug in ML_STRING macro (which caused the names
of certain exceptions to be garbled).
Scripts
-------
- Added an experimental heap2exec script (with a helper heap2asm
script in the background). This is experimental and so far
only works under Mac OS X.
- Made ml-build script slightly smarter. (One can now omit the
3rd argument -- which is then taken to be the same as the 1st
argument without the extension.)
- _run-sml now converts SMLNJ_HOME to a POSIX pathname on cygwin.
CM
--
- Fixed its incorrect assumption about the PowerPC being little-endian.
(PPC couldn be little-endian, but on the only supported platform --
Mac OS X -- it is big-endian.)
MLRISC
------
- Bugfixes in c-calls implementation.
--- NEW FILE: 110.53-README.html ---
<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<html>
<head>
<title>SML/NJ 110.53 NEWS</title>
</head>
<body>
<pre>
S M L / N J
1 1 0 . 5 3 N E W S
Februar 25, 2005
WARNING
This working version is believed to be stable, but
we have not run our full regression-test suite.
</pre>
<dl>
<dt>HOME:</dt>
<dd><a href="http://www.smlnj.org/index.html">http://www.smlnj.org/index.html</a></dd>
<dt>FILES:</dt>
<dd><a href="http://smlnj.cs.uchicago.edu/dist/working/110.53/">http://smlnj.cs.uchicago.edu/dist/working/110.53/</a></dd>
</dl>
<h2>Summary:</h2>
Bugfixes and minor feature improvements.
<hr>
<h2>Details:</h2>
<dl>
<dt>Basis:</dt>
<dd>
<ul>
<li><tt>SMLofNJ.Susp</tt> is back.</li>
</ul>
</dd>
<dt>CML</dt>
<dd>
<ul>
<li>Thread-safe version of <tt>structure Atom</tt> matches
original signature and re-uses the original implementation.
Moreover, since an <i>mvar</i> is used as a lock (instead of a separate
server thread) to protect the global hashtable, it is now
possible to create atoms even before <tt>RunCML.doit</tt> has started
executing.</li>
</ul>
</dd>
<dt>Library</dt>
<dd>
<ul>
<li>Moved source tree from <tt>smlnj-lib/TraceDebugProf</tt> to a new
<tt>tools/TraceDebugProf</tt>.</li>
</ul>
</dd>
<dt>NLFFI</dt>
<dd>
<ul>
<li>Handles access to signed and unsigned <tt>long long</tt> data.</li>
<li>Handles <tt>long long</tt> function arguments and results (but so far
only on the PowerPC).</li>
<li>Some organizational cleanup of the library code.</li>
</ul>
</dd>
<dt>ML-Yacc</dt>
<dd>
<ul>
<li>Fixed formatting bug that sometimes caused syntactically incorrect
code from being formed.</li>
</ul>
</dd>
<dt>Runtime</dt>
<dd>
<ul>
<li>Fixed bug in PC-sampling profiler.</li>
<li>Fixed off-by-1 bug in <tt>ML_STRING</tt> macro (which caused the names
of certain exceptions to be garbled).</li>
</ul>
</dd>
<dt>Scripts</dt>
<dd>
<ul>
<li>Added an experimental <tt>heap2exec</tt> script (with a helper
<tt>heap2asm</tt> script in the background). This is experimental
and so far only works under Mac OS X.</li>
<li>Made <tt>ml-build</tt> script slightly smarter. (One can now omit
the 3rd argument -- which in this case is taken to be the same as
the 1st argument without the extension.</li>
<li> <tt>_run-sml</tt> now converts SMLNJ_HOME to a POSIX pathname on
cygwin.</li>
</ul>
</dd>
<dt>CM</dt>
<dd>
<ul>
<il>Fixed its incorrect assumption about the PowerPC being
little-endian. (PPC couldn be little-endian, but on the only
supported platform -- Mac OS X -- it is big-endian.)</li>
</ul>
</dd>
<dt>MLRISC</dt>
<dd>
<ul>
<li>Bugfixes in <tt>c-calls</tt> implementation.</li>
</ul>
</dd>
</dl>
</body>
</html>
-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now.
http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click