[Python-Dev] Linker problems on Linux
Sun, 16 Jul 2000 11:29:44 +0200
Greg Stein wrote:
> On Sat, Jul 15, 2000 at 07:09:01PM -0400, Tim Peters wrote:
> > [M.-A. Lemburg]
> > > ...
> > > Why hasn't anyone objected to simply dropping such an important
> > > API without proper backwards compatibility workaround ?
> > If I knew it was happening, I would have whined. And I expect Guido would
> > have screamed. So let's fix it, and move on. I expect patch review isn't
> > as good as it used to be because of the clumsy SourceForge interface, so
> > nobody noticed.
> That has nothing to do with it. A full discussion occurred here on what to
> do. The consensus was reached and the code was changed.
I don't believe that Guido reached any consensus on this... he
has always tried to make releases as binary compatible as possible
(which is a very user friendly way of designing software).
> I knew it was happening, and I hardly ever look at the Patch Manager. In
> other words, you can't place the blame there.
> Personally, I had no issue with it at all. Python 1.5 extensions are not
> compatible with Python 1.6. We knew that already. So if they have to
> recompile anyways, then they are going to pick up the macro redefinition.
Sure they are... what makes you think that 1.5 extensions won't
work with 1.6 ? Ok, you get annoying warnings, but it's still better
as temporary solution than putting out a new version of Python
which won't work together with most existing extensions simply due
to linker problems.
> No big deal.
Ok, Greg, you pay the support bill ;-)
> > > The situation on Windows is already bad enough with extensions
> > > compiled against 1.5 causing a segfault...
> Unrelated FUD.
Why unrelated ? Why FUD ? The problem is real (until we check in
Anyway, enough ranting ;-) ... I'll review Thomas Wouters patch
and then check it in.
Python Pages: http://www.lemburg.com/python/