PATCH: py-opengl didn't build on Tiger.

[jsnitow@…]

The latest py-opengl would not build on Tiger. I submitted the following patch to the Python patch manager (the URL has a more full description, as well):

​http://sourceforge.net/tracker/index.php?func=detail&aid=1203329&group_id=5470&atid=305470

With this patch applied to the Include/pyport.h in python 2.4.1, py-opengl compiles and runs.

Until the main Python project accepts this patch (and in case they don't), I'm sending it here, too, as the patch works Right Now. :-)

[jsnitow@…]

Context patch for python 2.4.1's Include/pyport.h

[mww@…]

excellent, thanks! commited;

[jsnitow@…]

Upon further reflection (and see comment by loewis in the aforementioned URL), this patch IMHO should be considered a temporary workaround, and not a fix. It may be masking a real bug in the build process of either python, py-opengl (likely), or possibly py-numeric.

[py-opengl probably shouldn't even have to worry about sys/stat.h at all.]

[Note also that a dummy .c file with _POSIX_C_SOURCE undefined and sys/stat.h included compiles just fine. Something screwy must have been happening during the overall build process of py-opengl (either it was toggling the state of _POSIX_C_SOURCE, or otherwise messing with the definition of the tokens _TIMESPEC or timespec, which was causing the sys/stat.h included in pyport.h, included in the build process of py-opengl, to emit a compile error.]

In any event, py-opengl now seems to compile without this patch, so this patch should probably be removed from pyport.h, especially if it also compiles for you, sans patch.

[blb@…]

Shooting over to the maintainer...

[markd@…]

Have subsequent Python updates taken care of this? Can this bug be closed?

[markd@…]

This works for me and appears to be fixed. Reopen if you still experience problems.