[MPlayer-dev-eng] [PATCH] Support for MOV-Reference Files

[Fabian Franz]

Am Sonntag, 12. Januar 2003 23:03 schrieb Arpi:
> Hi,
>
> > > > If mime/type==video/quicktime and video is bigger than normal set
> > > > cache (40 kb), cache-size is set to Content-Length ...
> > >
> > > BAD!!!
> >
> > ok, ok ;)
> >
> > ... removed from patch ...
> >
> > But anyway how can a plugin for example know what kind of buffer it needs
> > to smoothly playback qt-files for example ...
> >
> > So it has to always send -cache 8192 or so :-(
>
> i don't understand this... i've already fixed the .mov streaming few hours
> ago, your sample url works with 40k cache... so why do you want to pump up
> cache to n*10 mb ?

hm, but I still get after some time: (2-3 secs of video & audio)

AO: [oss] 22050Hz 2ch Signed 16-bit (Little-Endian)
Building audio filter chain for 22050Hz/2ch/16bit -> 22050Hz/2ch/16bit...
Start playing...
NewGWorldFromPtr returned:65536
ImageCodecPreDecompress cres=0x0
supported csp: 0x79757675 uvuy
supported csp: 0x79757673 svuy
Stream not seekable!V:  0,106 ct:  0,017   50/ 50   9%  5% 14,9% 0 0 15%
Stream not seekable!
ImageCodecBandDecompress cres=0xFFFFFFCE (-0x32) -509%  5% 15,9% 1 0 15%
Stream not seekable!V:  0,678 ct:  0,033   52/ 52   9%  5% 16,0% 2 0 0%
Stream not seekable!
ImageCodecBandDecompress cres=0xFFFFFFCE (-0x32) -50
Stream not seekable!V:  0,967 ct:  0,042   53/ 53   8%  5% 18,0% 3 0 0%
Stream not seekable!
ImageCodecBandDecompress cres=0xFFFFFFCE (-0x32) -50
Stream not seekable!V:  1,255 ct:  0,050   54/ 54   8%  5% 22,5% 4 0 10%
Stream not seekable!
ImageCodecBandDecompress cres=0xFFFFFFCE (-0x32) -50
A:   6,0 V:   4,5 A-V:  1,543 ct:  0,058   55/ 55   8%  5% 26,9% 5 0 5%

MPlayer interrupted by signal 11 in module: decode_video
- MPlayer crashed by bad usage of CPU/FPU/RAM.
  Recompile MPlayer with --enable-debug and make a 'gdb' backtrace and
  disassembly. For details, see DOCS/bugreports.html section 5.b.
- MPlayer crashed. This shouldn't happen.
  It can be a bug in the MPlayer code _or_ in your drivers _or_ in your gcc
  version. If you think it's MPlayer's fault, please read DOCS/bugreports.html
  and follow instructions there. We can't and won't help unless you provide
  these informations when reporting a possible bug.
SoundConverterEndConversion:0
SoundConverterClose:0

I saw your cvs commit and it is in my tree, but it don't seems to work :-(

>
> anyway browsers can be configured to first download the file to tmp and
> then run the app on the local file, instead of passing the http url to the
> app.
>
> > But then I vote to fix it and add an -cacheopts option that you can set
> > the prefill-buffer to something different than 20% of cache ...
>
> agree, option for cache-prefill (20%) and cache-back (50%) should be added

hm, ok ...

can you do it ? /me feels slowly tired ...

>
> > > > PS: I hope it is clean and there are no 10l's in for example the code
> > > > with the basename-thingy (I really don't like string-handling, I
> > > > really don't ;) )
> > >
> > > ask .so for help :)))
> >
> > Ah, #mplayer-ppl told me what you meant :)
>
> what did they told? :))

pont so = Pontscho ;)

>
> > ok, changed to one line ;)
> >
> > file_format=demuxer->type...
> > As file_format is used in next switch and ifs .. :-/
>
> ok, it would be a 10l for me :)

:))
>
> > PS: Hope this time it is ok :)
>
> i'll test

ok :)

cu

Fabian
>
>
> A'rpi / Astral & ESP-team