I haven't used it, but you might try Rational Purify from IBM as a valgrind alternative on Windows. They used to have a free trial. Btw. Are there any docs that I can read on the issues involved in fortran-c-python bindings? Anything related to gfortran (fortran 95/2003) and python in particular would be much appreciated!!!<br>
<br><br>Cheers,<br>William<br><br><div class="gmail_quote">On Mon, Jun 22, 2009 at 2:39 AM, David Cournapeau <span dir="ltr"><<a href="mailto:cournape@gmail.com">cournape@gmail.com</a>></span> wrote:<br><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
<div class="im">On Mon, Jun 22, 2009 at 10:15 AM, Bruce Southey<<a href="mailto:bsouthey@gmail.com">bsouthey@gmail.com</a>> wrote:<br>
><br>
><br>
> On Sun, Jun 21, 2009 at 4:01 AM, David Cournapeau<br>
> <<a href="mailto:david@ar.media.kyoto-u.ac.jp">david@ar.media.kyoto-u.ac.jp</a>> wrote:<br>
>><br>
>> (Continuing the discussion initiated in the neighborhood iterator thread)<br>
>><br>
>> Hi,<br>
>><br>
>> I would like to gather people's opinion on what to target for numpy<br>
>> 1.4.0.<br>
>> - Chuck suggested to drop python < 2.6 support from now on. I am<br>
>> against it without a very strong and detailed rationale, because many OS<br>
>> still don't have python 2.6 (RHEL, Ubuntu LTS).<br>
>> - Even if not many new features have been implemented since 1.3.0,<br>
>> there were several changes which would be quite useful (using npy_math<br>
>> from numpy for scipy.special, neighborhood iterator for scipy.signal).<br>
>> So releasing 1.4.0 soon would be useful so that scipy 0.8.0 could depend<br>
>> on it.<br>
>> - Fixing crashes on windows 64 bits: I have not made any progress on<br>
>> this. I am out of ideas on how to debug the problem, to be honest.<br>
>><br>
><br>
> I think this is an essential requirement for numpy.<br>
<br>
</div>Yes. Windows 64 bits is finally usable (enough drivers) for the end<br>
users. And the (totally unusable) numpy 1.3 64 bits binaries have been<br>
downloaded quite a bit, confirming the interest.<br>
<div class="im"><br>
> I am prepared to attempt<br>
> to try to help as I now have a 64-bit windows system I can use. Just that I<br>
> do not like the windows environment for programming one bit.<br>
<br>
</div>Me neither. It is really awful unless you are in the IDE. Ideally, a<br>
primarily windows developer would be in charge.<br>
<div class="im"><br>
> So I would<br>
> appreciate any pointers to get the necessary functional system for 64-bit<br>
> windows.<br>
<br>
</div>To summarize the issue. We have two alternatives:<br>
- compiling with VS 2008 works more or less. There are a few bugs,<br>
but those are solvable through the debugger I guess with enough<br>
motivation.<br>
- compiling with mingw-w64 crashes randomly. Sometimes it crashes<br>
at import, sometimes during the unit tests, sometimes even before<br>
calling init_multirarray (the first numpy C extension). It also<br>
depends on how you launch python (IDLE vs. cmd.exe).<br>
<br>
The big problem with VS 2008 is that there is no free fortran compiler<br>
compatible with VS 2008. I can not even compile a trivial fortran + C<br>
project with the VS 2008-gfortran combination. A numpy only build<br>
could still be useful (for matplotlib, for example), but I am<br>
reluctant to do that if we later go the mingw route, since both built<br>
would be more than likely ABI incompatible.<br>
<br>
With mingw compilers, we get gfortran "for free" (I can compile scipy,<br>
for example), but there is this very serious crash which happens<br>
randomly. The debugger does not work (maybe some stack corruption),<br>
and there is no valgrind on windows, so it is very difficult to track<br>
down. Of course, mingw debugging symbols are not compatible with MS<br>
compilers, so the MS debugger is no option either.<br>
<div><div></div><div class="h5"><br>
David<br>
_______________________________________________<br>
Numpy-discussion mailing list<br>
<a href="mailto:Numpy-discussion@scipy.org">Numpy-discussion@scipy.org</a><br>
<a href="http://mail.scipy.org/mailman/listinfo/numpy-discussion" target="_blank">http://mail.scipy.org/mailman/listinfo/numpy-discussion</a><br>
</div></div></blockquote></div><br>