[MPlayer-dev-eng] [PATCHES] vobsub back to the future

[Reimar Döffinger]

On Sun, Jun 12, 2011 at 12:10:50PM +0200, Clément Bœsch wrote:
> On Sun, Jun 12, 2011 at 11:37:04AM +0200, Reimar Döffinger wrote:
> > On Sat, Jun 11, 2011 at 10:08:35PM +0200, Clément Bœsch wrote:
> > > On Sat, Jun 11, 2011 at 06:50:56PM +0200, Reimar Döffinger wrote:
> > > > On Sat, Jun 11, 2011 at 06:40:38PM +0200, Clément Bœsch wrote:
> > > > > Two patches that were lying in a forgotten directory attached. I remember
> > > > > sending them a while ago, but I can't remember the outcome :)
> > > > 
> > > > Hm, I think the second one changes the parsing from unsigned to signed.
> > > > Or rather, strtoul behaves rather strangely for negative input.
> > > > I suspect it doesn't really make things worse, but I don't know.
> > > 
> > > Fixed locally. It doesn't matter much thought, numbers are small in that
> > > case.
> > > 
> > > > Error detection also changes slightly.
> > > > E.g. previously a line ending with
> > > > filepos: z
> > > > would have failed, now it would be interpreted as filepos 0 without
> > > > causing an error.
> > > 
> > > No, it fails:
> > > 
> > > ERROR in timestamp: 00:02:13:600, filepos: z
> > > 
> > > > Admittedly it's not much worse, because before
> > > > filepos: 0z
> > > > would still have gone through...
> > > 
> > > Still "works".
> > > 
> > > I'll apply those patches soon if you don't mind.
> > 
> > I don't mind. But could you also apply the patch from
> > Mike Castle that fixes the comment for that code
> > (comment says format is HH:MM:SS.mmm but it is
> > HH:MM:SS:mmm).
> 
> I removed it since the sscanf is clear enough and appears quickly on top
> of the function.

Yes, that's probably fine, too.