a number of warnings are emitted when compiling pyOpenSSL

[Jean-Paul Calderone]

When compiling the C source in pyOpenSSL, a handful of compiler warnings are emitted. These are largely harmless, but fixing them is still a good idea (because an important warning might pop up later and be hidden in the crowd, for example).

[Jean-Paul Calderone (exarkun)]

Some discussions about this kind of thing:

<http://lists.mplayerhq.hu/pipermail/ffmpeg-devel/2008-January/041354.html> - ffmpeg developers decide that fixing warnings is fine, as long as doing so doesn't introduce bugs or performance regressions.

<http://kerneltrap.org/node/8252> - Linus rants that warnings should not be silenced without understanding why they're happening, and that warnings are better off left in place than being mindlessly silenced.

<http://lkml.org/lkml/2005/6/19/14> - Implies that Linux kernel now disables warnings about incompatible signed/unsignedness of pointers.

<http://discuss.fogcreek.com/joelonsoftware4/default.asp?cmd=show&ixPost=133392&ixReplies=21> - Random people talking about how they use -Werror all the time and other people talking about how some compiler warnings are utterly without value.

[Jean-Paul Calderone (exarkun)]

Merged the branch into trunk. Changes will be included in the next release.