[MPlayer-dev-eng] won't compile on FreeBSD8.1 (dvdio.h)

[Diego Biurrun]

On Sun, Sep 05, 2010 at 08:37:00PM +0200, Reimar Döffinger wrote:
> On Sat, Sep 04, 2010 at 03:52:35PM +0200, Diego Biurrun wrote:
> > > Cleanup is good, but I think you are rather optimistic as to how
> > > much crap is necessary vs. how much is actually there for a very
> > > good reason.
> > 
> > I can assure you that most of the crap in the build system was not
> > there for a good reason.
> 
> I had not doubts about that.
> 
> > > Admittedly it not being documented is a good reason to
> > > try, but still it would be nicer to do it a bit more carefully
> > > for platforms you can't/won't test yourself.
> > 
> > It would be nice if the fringe platforms like *BSD got their act together
> > and fixed their stuff instead of complaining.  I doubt we shall see the
> > day though.
> > 
> > In the meantime I consider breaking broken fringe platforms every once
> > in a while a necessary cost of life.  Fixing them is OK and I will do
> > it when issues are pointed out, but I don't think we should let them
> > hold us back while doing development.
> 
> The purpose is not to hold back, just if you can easily do it
> it would be nice to have e.g. two commits with the first one
> saying "..., tested on ...", and the next one
> "..., untested, might break on obscure and broken platforms".

For the record: I have access to Debian on x86_32 and Ubuntu on x86_64,
plus OS X 10.5 on PPC. I usually don't test on all of them before
committing if I feel confident that the change is likely safe. Breakage
on any of those I notice quickly. I do not test on other platforms.

> Particularly if the second is smaller more people have an idea
> what might have gone wrong and reply quicker to bug reports.
> I don't intend to declare that a must, no worries, just
> saying that might be one way to make things a bit smoother.

Sure, I'll keep it in mind.

Diego