It is called by asyncqueue if remove flag is true, to let the newly added command check if it wants its predecessors to be removed.

Default is to always allow. Only implemented in CmdPlaytreeUpdate: only remove previous commands if they are about the same item.

-> Limit excess updates to playtree

(There are some debug messages left in this commit, will remove later)

* modules/gui/wxwidgets: wxwidgets 2.4.2 compilation fixes as suggested by Charlie Ledogar (charlie10 at gmail dot com)

The previous commit and this one globally have a rather negative impact, as we notify playlist waaayy too much while adding info. Should either be implemented in interface to reduce the number of updates, or as a flag to add_info if we have some more to do afterwards

     consolidated all dvdnav/dvdread patches into a single patch

   This is based on code from the current playlist dialog but this dialog should eventually make use of the widget as well (when enough features have been implemented).

   control, and fixed a few details

-Cette ligne, et les suivantes ci-dessous, seront ignorées--

M    src/Makefile

* Support little endian streams on Big endian output. In theory at least, the Mendelssohn_2.wav sample never worked for me. I think the dtssys demuxer has an endianness issue (probably in the header code or something). In my opinion the dtssys demux should only output big endian streams, because all DTS is generally big endian. It's just that .wav often isn't.

And just crash

Fix the a52 spdif filter, add big endian support to the DTS filter and add big endian support to the coreaudio module. Now if you use the coreaudio module, you can output encoded audio on the G5 and other systems. PLEASE check that your SPDIF support is not broken after this.

* Introduce 'spdb' the big endian variant of 'spdi' output. This will be needed for the encoded audio output on the PowerMac G5. Only A52 is supported so far. DTS support will follow later. (As well as the actual OSX output changes)

linux-dvb. This currently mean Twinhan cards and clones.

There are three compile time switches currently available to people who 
want to test this:

RESET_CAM_SLOTS can be defined to 0 to disable the CA_RESET call at 
startup. This greatly improves startup time for my cam at least. 

HLCI_WAIT_CAM_READY can be defined to 1 to enable a loop that waits 
untill the cam module is ready. This may take a long time (at least if 
RESET_CAM_SLOTS is set), and does not happen at all under some error 
conditions.

CAM_PROG_MAX is a limit on the number of programs the code will send 
CAPMT messages for. This can be used to work around buggy cam/card 
combos that have issues when too many capmts are added.

Hopefully this doesn't break the low level CAM support

* modules/codec/ffmpeg/ffmpeg.c: add BMP codec support for when ffmpeg increases their libavcodec version number.

   (http://forum.videolan.org/viewtopic.php?p=44548)

http://forum.videolan.org/viewtopic.php?t=14251