[Numpy-discussion] Moving to gcc 4.* for win32 installers ?

[David Cournapeau]

On Thu, Oct 27, 2011 at 5:18 PM,  <josef.pktd at gmail.com> wrote:
> On Thu, Oct 27, 2011 at 9:02 AM, David Cournapeau <cournape at gmail.com> wrote:
>> Hi,
>>
>> I was wondering if we could finally move to a more recent version of
>> compilers for official win32 installers. This would of course concern
>> the next release cycle, not the ones where beta/rc are already in
>> progress.
>>
>> Basically, the pros:
>>  - we will have to move at some point
>>  - gcc 4.* seem less buggy, especially C++ and fortran.
>>  - no need to maintain msvcr90 vodoo
>> The cons:
>>  - it will most likely break the ABI
>>  - we need to recompile atlas (but I can take care of it)
>>  - the biggest: it is difficult to combine gfortran with visual
>> studio (more exactly you cannot link gfortran runtime to a visual
>> studio executable). The only solution I could think of would be to
>> recompile the gfortran runtime with Visual Studio, which for some
>> reason does not sound very appealing :)
>
> What does the last mean in practice? (definition of linking in this case?)
> If numpy and scipy are compiled with MingW gcc 4.*, then it cannot be
> used with the standard python?

It would of course work with the standard python, it would be rather
useless otherwise :)

The main difference is that you could build numpy/scipy with MS
compilers for C/C++ code and use g77 for the fortran part. This has
always been hackish, but kinda worked. Not so much with gfortran
anymore. IOW, the choice for people building extensions on top of
numpy becomes full MingW or full MS/Intel compilers.

> Or does it just mean we cannot combine fortran extensions that are
> build with MingW with extension build with visual studio?

I hope that you could mix them as long as the extension does not
contain any fortran code.

>
> another example: would Matplotlib compiled against visual studio work
> with a new MingW compiled numpy? I guess that's what the ABI break
> would prevent?

This, I am not so sure, this would need testing.

> Since we will have to update MingW sooner or later anyway, I'm in
> favor of doing it. And given the comments on the mailing list about
> the Linux transition to gfortran, I expect that the transition will
> take some time.
>
> Thanks for your successful effort that installation on Windows was
> without problems for years for a user like me.

You're welcome !

David