Re: [xine-devel] Re: [xine-cvs] CVS: xine-lib/src/xine-engineaudio_decoder.c,1.28,1.29 metronom.c,1.18,1.19 metronom.h,1.9,1.10xine.c,1.47,1.48 xine_internal.h,1.40,1.41

[Daniel Caujolle-B. <seg...@cl...>]

Guenter Bartsch wrote:
> 
> Hi Daniel,
> 
> > >         audio_decoder.c metronom.c metronom.h xine.c xine_internal.h
> > > Log Message:
> > > fixed seeking (back to the old method)
> >
> >       old method is WRONG, it broke some *features*. So, seek is currently
> > broken
> 
> erm, sorry, but it's the other way round, seeking with the new method had
> a lot of problems, lots of races and metronom as well as the video decoder
> were not reset properly - if you still have the older version before my
> latest commits, just try to hit some number keys rapidly or try pgdown to
> jump from one mpg stream in the filelist to the next - you'll deadlocks,
> no audio (since metronom is confused), mpeg block artefacts (because
> libmpeg2 is not reset properly) ... so there'd be a lot work to do before
> the new seeking method would work.

	Lot of race of what ???? err, remember, the only thing is the
seek call ip->stop() instead of ip->close, so tell me what's wrong here
!

> 
> so, now that I've implemented a true pause function I thought we can get
> back to the old stable seeking implementation - I never had any trouble
> with it - I don't know about the dvdnav plugin, IMHO that polugin
> shouldn't be seekable in nav mode any way - since seeking the cirtual
> machine is a hack.

	Not sure too, seek just manipulate blocks (i locally implemented
that, and that's work, i hate to not be able to seek in nav mode) !

-- 
73's de Daniel, F1RMB.

              -=- Daniel Caujolle-Bert -=- seg...@cl... -=-
                        -=- f1...@f1... (AMPR NET) -=-