MacOSX/Framework/VLCMediaListAspect.m: Use the cachedMedia array when giving back the aspect items. That really boost performances by the way.

on Glen Gray patch.
(The code is commented until it is confirmed to compil/work.)

Based on a patch by Glen Gray.

DEMUX_CAN_PAUSE and DEMUX_CAN_CONTROL_PACE can now fail.

MacOSX/Framework/VLCMedia.m: Don't forget to init the metaDictionary, when initing from a media_descriptor object.

MacOSX/Framework/VLCMedia.m: Separate the different part by some nice comments. Remove uneeded and complicated code in meta information retrieval. Use stringWithUT8String instead of stringWithCString.

MacOSX/Framework/VLCMedia.m: Use mediaWithURL:(NSURL*) instead of mediaWithURL:(NSString*). Add a mediaWithPath:(NSString*) constructor.

MacOSX/Framework/VLC.xcodeproj: Various updates, such as removal of the SDK, gc support, and QuartzCore linking, and use NATIVE_ARCH as the only target.

Qt4 - Rewrite of the VLM Dialog. It features a list on the left of the VLM object, and a right pane to see them in details. Creation and modification is doable. This is not yet a VCR, but it could be :D
The code is way not finished and uncleaned, but feel free to improve.
Only Brodcast elements have a start of code.

It should fix some issues with Samples with Video + Cook.
If you think this is wrong, please revert.

* fixed a corner case on PowerPC-based Macs since MACOSX_DEPLOYMENT_TARGET is set to 10.1 on these devices, although we need at least 10.3 and prefer 10.4

* disable the usage of the OSX SDKs by default due to a known incompatibility between the libraries used on 10.4.8 (or higher) and the existing 10.4u SDK. This solves the current compilation issues with the AUHAL and Screen.Mac modules and closes #1399

* disable Screen Capturing on OSX until it is fixed because it blocks all builds on most 10.4 setups (refs #1399)

qt4: use preprocessor flags with moc, so it finds the headers in the right place

Fix hang in live555 when sending RTSP OPTIONS by letting readSocket() timeout after 5 seconds. Added reconnect case to modules/demux/live555.cpp (without backing off scenario).

(Modules that would still depend on module_config_t are not going to like this, but it's not like you weren't warned a long time ahead)