[MPlayer-dev-eng] [PATCH] Fix libass with gcc 2.95

[Uoti Urpala]

On Thu, 2006-08-17 at 16:43 +0200, Guillaume Poirier wrote:
> BTW, one thing that doesn't make sense to me in Uoti Urpala's
> argumentation, is: "this or that compiler/arch has been broken for xx
> weeks, it's useless, and should be dropped"... well, with this kind of

No, my argumentation is basically "compiler/arch has already been broken
for xx weeks/months and that has had no negative effects, thus no effort
should be wasted to fix it or to keep it from breaking again".

> argument, we'd probably only support gcc4.0 and 3.3/4 on x86/linux.
> That'd be a pitty.

False. Supporting only gcc 3.4 and up (obviously including 4.1 which you
missed) might be ok, but platforms other than x86/Linux have a lot more
users than gcc-2.95.

> Also, your argumentation sounds like how software corporation work:
> "this software is discontinued, do upgrade (and throw us some $$)",
> whereas in free software, things are more like "this software is
> supported as long as ppl care do provide support for it".

No, the current MPlayer situation is not "supported as long as there are
people willing and able to support it". If that were the case the people
who want the support would be doing the work. In the current situation
they aren't. If you want to maintain gcc-2.95 compatibility and do all
the work yourself (and keep the parts where the gcc-2.95 version would
be less readable/maintainable/efficient in a separate patch) that's OK
by me. If you want me to do gcc-2.95 workarounds even though you're not
able to sufficiently justify why that would be necessary or even useful
that is not OK.