[vlc-devel] RE : [PATCH] -- fullscreen enhancements
brezhoneg1 at yahoo.fr
Mon Feb 2 22:12:06 CET 2009
> -----Message d'origine-----
> De : vlc-devel-bounces at videolan.org [mailto:vlc-devel-
> bounces at videolan.org] De la part de Laurent Aimar
> Envoyé : lundi 2 février 2009 20:47
> À : Mailing list for VLC media player developers
> Objet : Re: [vlc-devel] RE : [PATCH] -- fullscreen enhancements
> On Mon, Feb 02, 2009, brezhoneg1 wrote:
> > > > The second patch (patch 7) corrects transition issues (vout
> > > > especially in fullscreen mode.
> > > I really don't feel at ease with that one. Don't ask me to apply
> > >
> > This patch was originally aimed at proving width/height restrictions
> > should not be seen as a big problem. This patch can certainly be
> > improved. Ideally, setting (the first time) and resetting (when
> > vout) should be mutualized. Yet, it is a working patch, and
> > have not made wrong assumptions.
> > My point is:
> > - vlc1.0 will be a landmark in vlc history.
> > - vlc1.0 has raised expectations that the transition problems will
> > solved (see numerous posts and answers on the forum)
> > For the sake of vlc, I really think working out this height/width
> > limitation would avoid users complaining bitterly for what can
> > be considered a bug. (With vlc1.0, this will even be more annoying
> > users don't have in mind this limitation and will see it as an
> > behavior)
> I do agree that this functionnality/fix is really welcome, and I
> for your patch.
> > Maybe, if someone with good knowledge of vout could react to it
> > from what I was said ?).
> The only problem is that I do not feel confident enough on the vout
> to be sure it won't create any problem.
> If we were not about to issue the 1.0 release I would have applied
> But for 1.0 I think it is safer to not use it (or at least I cannot
> if applying it is safe), so yes it will be a bit less user friendly
> prefer to avoid regression.
> As soon as 1.0 is released (and in its own branch) I will apply this
> (bug me if I don't) and if tested stable it could (if agreed) make it
> stabilization branch later on.
Well, let's do it your way.
I too am not sure 100% there is no side effect whatsoever. The one thing
I know is that I have been using this patch for a while on many
different formats as a rather basic user (no fancy vout-filter and so
on), and it works fine.
I also think the vout architecture is well designed once you get to know
how it is done (only missing thing is a good doc to start with). And my
conviction is that width and height are never a problem (proof is
resizing is done at the vout level for x11 for example). Chroma has not
been a problem either in my tests (chroma conversion is also managed on
the fly when required), but I am less certain of all the implications.
More information about the vlc-devel