[MPlayer-dev-eng] [PATCH] fix seeking in demux_lavf

[Uoti Urpala]

On Wed, 2008-08-13 at 02:19 +0200, Aurelien Jacobs wrote:
> Uoti Urpala wrote:
> > On Tue, 2008-08-12 at 05:17 +0200, Michael Niedermayer wrote:
> > > Its not as if i would want some kind of monarchy, its just that half of
> > > the new mails here where about a bunch of monkeys fighting about a piece of a
> > > 7 year old banana while they had plenty of fresh ones.
> > > Why exactly did diego fight so fiercely to remove 1 harmless line from
> > > a file maintained by ivan?
> > 
> > What else do you suggest he could have done?
> 
> He could have added a check in configure as suggested. It may seem a bit
> overkill, but it would have avoided any potential issue with old Xfree,
> and it would have avoided flames.

The way Diego initially posted the patch was in no way provocative. The
the first time Ivan mentioned XFree 4.1 was 5 days later. Before that
there was no mention of any system where the patch could cause problems.
If Ivan's first reply had been "XFree 4.1 lacks the prototype and I
still want to support it" then what you say could be applicable; however
Ivan chose not to do that and trolled instead.

> Another possiblity would have been to simply disregard the harmless warning
> generated by this line and switch to more important tasks.

While the warning is mostly "harmless" the line is clearly questionable.
I don't think it would be a good idea to just leave such issues in the
code.

> > >  Have we run out of dirty code that needs to
> > > be cleaned up And that can be cleaned up easily in a non provocative way?
> > > Must be so ...
> > 
> > How was the way he did it provocative?
> 
> He committed, ignoring maintainer disagreement and braving rules.
> This is provocative.

But that was only AFTER most of the flaming, not the cause of it. I
think it's understandable he didn't want to bother trying to cooperate
with Ivan after that (after Ivan had refused any rational discussion).
The unnecessary controversy caused by Ivan's trolling was the reason not
to cooperate, not the other way around.

> Note that I don't care about the technical side of this issue.
> But it feels very wrong to me to ignore maintainer's decision.
> IMO, this has the same effect as svn rm DOCS/tech/MAINTAINERS.

Even if you want to give a lot of control to per-file maintainers (which
I think is not a good idea) it still doesn't make much sense to leave
the declarations in one file only. Having them in "Ivan's" vo_xvmc.c but
not in vo_x11.c would not do much good. In fact it seems that vo_xvmc.c
wouldn't even use the XShmGetEventBase prototype for anything anyway.