[SciPy-Dev] Release blocker?

[Pauli Virtanen]

On Wed, 16 Feb 2011 14:49:51 -0800, Christoph Gohlke wrote:
[clip]
> I tested msvc9/MKL builds of rc3 with 8 different Python versions and
> did not notice this issue.

Thanks a lot for your trouble!

> I think the problem is not Win7 but that the
> official scipy binaries link against two different C runtime libraries,
> which is a bad idea on any platform. For example, interpnd.pyd correctly
> uses msvcr90.dll, while qhull.pyd uses msvcrt.dll.

Hmm, yes, different C runtimes would make sense as the cause -- I was 
wondering how just crossing DLL borders could cause problems for 
malloc/free...

> Actually, interpnd.pyd uses functions from both C runtime libraries.

That's probably not a good sign.

I managed to reproduce the issue by building Scipy with Mingw against the 
official Numpy 1.5.1 binary, so at least that sort of build setup appears 
to produce strange binaries. The build is driven by distutils, so I've 
little idea what's happening in it. This might be some sort of Mingw 
problem.

So at least two solutions seem possible: (i) apply the patches from my 
github branch to replace malloc/free by static allocations, or 
(ii) fix the build setup.

	Pauli