Age | Commit message (Collapse) | Author |
|
2) Simplify libfaad init
CVS patchset: 6588
CVS date: 2004/05/24 22:17:57
|
|
1 - When a first frame is still referenced, poll until there is only one reference. (with a limit, see FIRST_FRAME_MAX_POLL). This way the frame is displayed as soon as possible.
2 - When a "flush" occured, don't wait the full duration of the frame and exit the waing loop as soon as possible.
The problems have been identified and fixed using a toxine script to stress seeking. It's available here (requires toxine cvs) :
http://xine.sourceforge.net/tmattern/seek_test_1.toxine
It performs approx. 500 seeks. I have tested avi, asf, mov, mpg, mkv streams.
The engine can do a bit more than 10 seeks per second here (athlon 2.6 GHz), with all kind of streams. ;)
toxine has been improved to support a such script (env vars, and the "repeat" command), thanks to Daniel.
CVS patchset: 6585
CVS date: 2004/05/23 21:30:26
|
|
CVS patchset: 6584
CVS date: 2004/05/23 18:47:02
|
|
Make it easier to debug using XINE_PROFILE.
CVS patchset: 6583
CVS date: 2004/05/23 18:41:57
|
|
CVS patchset: 6581
CVS date: 2004/05/23 16:20:56
|
|
Problem solved using a toxine script to stress seeking.
CVS patchset: 6574
CVS date: 2004/05/21 13:41:02
|
|
rewiring a closed port won't cause the new destination to be opened.
CVS patchset: 6564
CVS date: 2004/05/18 03:16:12
|
|
(intercepting the audio_out port).
calling xine_post_wire_audio_port( post output, audio_out )
would close the audio_out and then reopen it with rate=0, bits=0, mode=0.
(see post_audio_rewire() in post.c)
CVS patchset: 6563
CVS date: 2004/05/18 02:01:39
|
|
CVS patchset: 6561
CVS date: 2004/05/17 21:28:06
|
|
The fix stopps dropping properly received data.
I realized the problem when switching channels in VDR. From time to time, it happend that switching took more than 50 ms (the timeout value used in _x_read_abort()), which caused the select() to timeout.
In the case where 'stream->demux_action_pending' was set, the already received data was dropped due to returning 0. This finally resulted in your demuxer to fail with DEMUX_END, as it couldn't read 6 bytes from the PES header.
So for the case, that there is some data available, it now takes twice the timeout (i. e. 100 ms), before a return 0 happens.
Some more information: as in rare cases switching channels could take even longer than 100 ms, I've also added a loop around _x_read_about() in my VDR input plugin, which returns 0 just after this state lasts for 5 seconds.
CVS patchset: 6557
CVS date: 2004/05/16 21:45:24
|
|
Under certain circumstances, the result caused an overflow in 'num_frames', so that it could get negative.
CVS patchset: 6556
CVS date: 2004/05/16 21:39:55
|
|
CVS patchset: 6552
CVS date: 2004/05/16 17:58:16
|
|
CVS patchset: 6545
CVS date: 2004/05/15 23:44:25
|
|
CVS patchset: 6544
CVS date: 2004/05/15 20:27:50
|
|
allows the code for handling this special case to actually be used
CVS patchset: 6532
CVS date: 2004/05/13 21:38:49
|
|
something outside the delivered image.
fixes border calculation for zoom < 100%.
CVS patchset: 6517
CVS date: 2004/05/11 02:21:16
|
|
CVS patchset: 6508
CVS date: 2004/05/09 22:05:34
|
|
CVS patchset: 6504
CVS date: 2004/05/09 19:21:22
|
|
no replacement, so it's used, but with configure check.
CVS patchset: 6503
CVS date: 2004/05/09 17:42:23
|
|
CVS patchset: 6497
CVS date: 2004/05/07 22:49:24
|
|
strings
-> read and update them as strings
* remove the obsolete CONFIG_TYPE_* constants and replace them with
XINE_CONFIG_TYPE_* from xine.h
CVS patchset: 6496
CVS date: 2004/05/07 14:38:14
|
|
CVS patchset: 6483
CVS date: 2004/05/04 21:49:00
|
|
CVS patchset: 6464
CVS date: 2004/04/30 09:38:52
|
|
CVS patchset: 6452
CVS date: 2004/04/27 20:46:26
|
|
please everybody watch out for spelling (and other) errors
CVS patchset: 6442
CVS date: 2004/04/26 17:50:06
|
|
CVS patchset: 6433
CVS date: 2004/04/25 14:50:32
|
|
Thanks Michael ;)
CVS patchset: 6429
CVS date: 2004/04/24 11:28:00
|
|
This is usefull when the format does'nt contain a frame rate / duration info like the ASF format.
See this thread for more details:
http://thread.gmane.org/gmane.comp.video.xine.devel/8387
CVS patchset: 6426
CVS date: 2004/04/22 23:19:03
|
|
"down" is the usual datapath from decoder to output, "up" is the opposite way
-> this means swapping two function names
-> this means increasing the post plugin iface version
external plugin developers: just swap usage of
_x_post_frame_copy_up() and _x_post_frame_copy_down() functions
CVS patchset: 6412
CVS date: 2004/04/17 19:54:31
|
|
can receive error events before the end of xine_open()
CVS patchset: 6410
CVS date: 2004/04/16 16:34:22
|
|
CVS patchset: 6407
CVS date: 2004/04/15 19:36:57
|
|
CVS patchset: 6399
CVS date: 2004/04/13 07:59:22
|
|
they might call port->close(), which might finish a pending disposal of a
post plugin, which will try to decrement the reference count for that plugin,
which will try to grab the plugin catalog lock...
CVS patchset: 6393
CVS date: 2004/04/11 14:51:20
|
|
CVS patchset: 6385
CVS date: 2004/04/10 17:41:44
|
|
CVS patchset: 6372
CVS date: 2004/04/10 03:45:52
|
|
about buffer overruns.
CVS patchset: 6368
CVS date: 2004/04/09 22:12:09
|
|
CVS patchset: 6367
CVS date: 2004/04/09 21:46:54
|
|
explains
CVS patchset: 6363
CVS date: 2004/04/09 15:06:02
|
|
xine_init().
CVS patchset: 6357
CVS date: 2004/04/09 11:26:10
|
|
(use toxine luke... ;-) )
CVS patchset: 6351
CVS date: 2004/04/08 16:20:14
|
|
CVS patchset: 6350
CVS date: 2004/04/08 13:37:54
|
|
CVS patchset: 6348
CVS date: 2004/04/08 11:10:10
|
|
CVS patchset: 6342
CVS date: 2004/04/07 18:10:20
|
|
x = a / b + c
to
y = b + c
x = a / y
which is wrong and breaks the resampling sync method; changing to
y = a / b
x = y + c
should fix; thanks to Carsten Weinhold for the patch
CVS patchset: 6341
CVS date: 2004/04/07 18:07:25
|
|
firewall!!!).
New functions into io_helper:
_x_io_tcp_connect_finish ... wait for successful connect with timeout
_x_io_tcp_read_line ... read the string from the socket (improvements neded)
Use io_helper functions in input_cdda.
Don't try connect again to freedb server, if previous connecting was failed.
CVS patchset: 6337
CVS date: 2004/04/06 18:40:32
|
|
busy.
CVS patchset: 6333
CVS date: 2004/04/05 20:01:24
|
|
CVS patchset: 6329
CVS date: 2004/04/05 17:58:54
|
|
CVS patchset: 6318
CVS date: 2004/03/29 19:50:58
|
|
CVS patchset: 6313
CVS date: 2004/03/28 19:56:08
|
|
(buf->free_buffer() for example must not be called with a ticket held, since
it can block, when the demuxer triggered nbc, which will occupy the fifo lock)
CVS patchset: 6312
CVS date: 2004/03/28 19:51:56
|