[Bf-committers] Python bug (possible "show-stopper" bug) or user mistake?

[Yann Vernier]

On Thu, Dec 30, 2004 at 06:57:57PM +0400, Juan Julio Peña Mena wrote:
> > By the way, for the FAQ purpose, when the final user is presented with
> > various glibc versions to choose from (like it was the case on the
> > 2.36 release day), what is the rule?  Should he pick the one with the
> > lowest glibc available or the one with the glibc version number
> > nearest to the one already installed on his computer?
> > 
> > (My machine has glibc-2.3.3 installed, so I whent for the one closer
> > to the one in my computer )

We didn't quite have a consistent build environment. The glibc 2.2
versions currently available were built using an environment I made for
this purpose, hopefully as compatible as we can get it. There has been
work on polishing it and duplicating it so this problem won't have to
reoccur - the previously available glibc 2.3 and 2.2 builds you observed
were built in my absence (sorry about that).

As a rule, you would want to have the closer versions (as long as only
the patchlevel may be higher, i.e. a 2.3.3 binary on a 2.3.2 system
might work, but not a 2.3 binary on a 2.2 system). However, glibc is one
of the most standardized libraries while others are much more varying -
including the C++ and Python libraries. I should try to verify LSB
compliance for the build, but even that would not cover very much.

-- 
PGP fingerprint = 9242 DC15 2502 FEAB E15F  84C6 D538 EC09 5380 5746
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: Digital signature
Url : http://projects.blender.org/pipermail/bf-committers/attachments/20041231/facb4e80/attachment.pgp