[Mplayer-dev-eng] release 0.18...

[Arpi]

Hi,

> Another problem.
> Your work with mplayer looks great except seeking (DVD playing).
> Of course 32 bit integer is not enough to hold the "global" offset
> value (of course you will see negative values and so on for sizes caused
> by integer overflows).
> Using 'long long' types would solve the problem but it's ugly
> and slower. I think we should move to sector base seeking, and offset
> storing and so in, since that is the smallest element can be read once
> from encrypted disks. So offset, size, etc 1 means the second sector
> (0 s the first of course). This would simplify code and makes it faster
> too (you don't have to convert between byte offset and sector offset
> values). And of course it fits into 32 bit integer without problem.
I agree.

> Is there any problem which does not allow to use ONLY sector based
> values for DVD playback?
I dunno. We should check the whole code, there are many reference to the
offset value.

But.
There are 2 position-related field in teh demuxer struct:
  int filepos; // input stream current pos.
  int idx_pos;

later is used for AVI files only, but it could keep the DVD sector position.

we must change mpeg seeking and progress bar code, other should work this
way.


A'rpi / Astral & ESP-team

--
mailto:arpi at thot.banki.hu
http://esp-team.scene.hu

_______________________________________________
Mplayer-dev-eng mailing list
Mplayer-dev-eng at lists.sourceforge.net
http://lists.sourceforge.net/lists/listinfo/mplayer-dev-eng