[FFmpeg-devel] [PATCH] Swscale YUV2RGB table generator

[Diego Biurrun]

On Mon, Feb 09, 2009 at 09:50:44PM +0100, Michael Niedermayer wrote:
> On Mon, Feb 09, 2009 at 07:21:29PM +0100, Diego Biurrun wrote:
> > On Mon, Feb 09, 2009 at 07:11:11PM +0100, Michael Niedermayer wrote:
> > > On Mon, Feb 09, 2009 at 05:16:29PM +0100, Diego Biurrun wrote:
> > > > On Mon, Feb 09, 2009 at 05:38:24PM +0200, Kostya wrote:
> > > > 
> > > > Nits aside, this looks like a possibly disruptive change.
> > > > I have no idea if you will finish it before the release,
> > > > but I suggest that you commit it after the release.
> > > 
> > > id like to drop the old scaler before the release and that needs
> > > this.
> > > and IMHO if this patch is tested the risk is low
> > 
> > I can understand your intentions, but I would prefer not to do it
> > before the release.  While the risk may be low, not dropping it keeps
> > the risk even lower.
> 
> i like to repeat, that kostyas patch should go in, if you ask for the
> old scaler droping to be postponed i can understand this but this code
> is not risky if it is tested, its also one file replacing another, this
> is easy to revert if it causes a problem.
> 
> if you want swscale completely frozen so that no bugfixes nor features
> can be commited then you should make a fork of it now IMHO
> 
> and before you say ohh no, bugfixes are ok, i will not do bugfixes twice,
> now on walkens GPL code and then in 2 weeks on kostyas LGPL code. Nor will
> i debug or test code that i know will be droped in 2 weeks, OTOH i of
> course wont stop you or anyone else from doing work twice.

I don't see you having to do work twice, nor anybody else really.

As I said, let's not waste time discussing hypotheses.  So far,
Kostya is not finished.  If he does finish, we can talk.

> PS: I think it should be the maintainers and patch authors decission if
> a patch is too risky for the release or not.

I'd like to err on the side of caution.

Diego