[vdr] [ANNOUNCE] vdr-xine-0.7.3 plugin
rnissl at gmx.de
Mon Apr 11 23:33:49 CEST 2005
I'm pleased to announce release 0.7.3:
2005-04-11: Version 0.7.3
- Started detection of AFD header in xine to automatically crop out the
interesting part of the image later.
- Adopted implementation of cXineDevice::SetDigitialAudioDevice() to
different calling order in VDR 1.3.23.
- Improved cXineDevice::SetDigitalAudioDevice() when replaying
- Added setup option to automatically make vdr-xine the primary device
while xine is connected to vdr-xine (requested by Der_Olli at
- Added setup option to consider all semi transparent colors as opaque
to make the menu more readable.
- Added commandline option '-s' to switch to skin 'curses' while xine is
not connected to vdr-xine (requested by Rantanen Teemu).
- Added commandline option '-q' to suppress debug messages (useful in
combination with option '-s').
- Moved disconnect to cXineDevice::Stop() to get the new options to
- Fixed all (?) deadlock situations in RPC command processing (e. g.
stopping replay while switching a channel).
- Fixed deadlocks in vdr-xine's xread(). A possible drawback is that now
a disconnect might happen in such a case.
- Fixed VDR's I-frame processing which caused disconnects e. g. while
moving cut marks in HDTV recordings. vdr-1.2.23-dvbplayer3.patch is
highly recommended for proper operation of vdr-xine.
- Improved cXineDevice::StillImage() implementation to immediate display
the frame (improves moving cut marks).
- Fixed cXineDevice::StillImage() to work properly in combination with
the plugin vdr-radio.
BUG: xine's driver xxmc shows just a black screen on my EPIA
- Reintroduced usleeps() in input_vdr.c for flush, OSD flush and reset
audio. sched_yield() simply caused to much CPU load while waiting
about 40 ms to reach the expected state. Improves number of dropped
frames when switching channels.
- Optimized OSD display: VDR's channel display repeatedly sends a dirty
OSD which doesn't differ from the previous one. Improves number of
dropped frames while switching channel.
BUG: it's still unclear whether this causes some OSD artefacts.
- Fixed demux_mpeg_pes' discontinuity detection. Previously, when a PTS
wrap happend, xine stopped replay for about 26.5 hours.
- vdr-xine now nolonger set's xine's metronom directly but tells it's
demuxer to do the job. Improves switching channels.
- Optimized implementation of cXineDevice::Clear() in input_vdr.c.
BUG: it may happen that xine's audio driver ALSA might get into a
state of "silence" where it doesn't recover from until you stop
replaying the recording. I still didn't find a way to reproduce this
but it has to do with trickspeed, pause, play, and probably cut marks.
- Fixed post_vdr.c to detect MRL changes for discovering streams sent
from VDR, e. g. when xine is not started with the MRL specified on
it's command line.
BUG: It's possible that xine crashes due to this detection. xine
doesn't allocate a different stream for a different MRL, but maybe
other players do. I'm not sure whether I managed to increase the
streams usage counter properly (by allocating an event queue) until I
detect the new stream respectively MRL.
BUG: post_vdr doesn't operate when xine's driver xxmc is used due to
some limitation/incomplete implementation in xine's plugin interface.
- Fixed xine's deinterlacer interface to take care of cropping.
Previously the OSD was resizing like mad e. g. between 1920x1080 and
- Added support for VDR's new AUDIO key in xine (thanks to Darren Salt
for reporting this issue).
For this release I suggest the following xine sources:
Highly recommended is the following patch:
Dipl.-Inform. (FH) Reinhard Nissl
mailto:rnissl at gmx.de
More information about the vdr