[Distutils] GCC versions and binary egg names

[David Cournapeau]

On Tue, Jul 27, 2010 at 2:04 AM, Chris Withers <chris at simplistix.co.uk> wrote:
> Hi All,
>
> In addition to the UCS2/4 problems already described by MAL and which I've
> bumped into myself, I now have a problem with binary linux eggs where the
> GCC version doesn't match that of the system the egg is being installed on.

That really should not be a problem, unless you use vastly different
versions of gcc.

> Current egg metadata doesn't take either of these into account or even
> provide a way for the nature of the binary egg to be recorded.
>
> How're people coping with this?
> What are the future plans in this area?

If you need to follow non trivial ABI aspects, the current python
infrastructure
is too simplistic. Also, I think that short of preparing rpm/deb for
the distributions you target, you are going into a very time consuming
process, and automated solution are completely beyond reach (the
closest being the build service from open suse, but that still depends
on the system packaging system).

Projects who manage to distribute cross distributions binaries usually
put a lot of manpower behind it (only the biggest ones or the ones
which strong financial incentive do it). Or they provide the whole
stack (EPD, Activestate distributions, Sage) so that they depend as
little as possible on the target environment.

cheers,

David