VideoLAN authored Mar 08, 2008 and Rafaël Carré committed Mar 08, 2008

Signed-off-by: Rafaël Carré <funman@videolan.org>

Once again, very nicely handled...

Windows libgcrypt.a went from over 470KB to 317KB...

Revert "Fix some issue on Compiling portaudio for Windows."

This reverts commit 12be1dfd.

How come noone noticed that one before ?

Disable some ciphers in libgcrypt. 20% size reduction on gcrypt for windows. This can still be improved I think. Not done for mac

Fix cross compilation from 64bits host

Signed-off-by: Jean-Paul Saman <jpsaman@videolan.org>

contribs/extras: Compensate for shortfalls in ffmpeg's svn repository.  Will submit shortfalls as bugs + patch back to ffmpeg dev team.

contrib/src/Makefile: Checkout CMake project into 'cmake' subdirectory, set up code for downloading cmake directly from archive for future use.

causing segfault on Windows platform

* extras/contrib/src/Patches/x264-svn-win32.patch: revert [25113] and [25140] log x264 thread count from win32.patch and put it in a generic patch for all platforms (threads=0 autodetect for x264 is reportedly failing on win32 and linux too and always reports threads=1, though x264 cli works fine so it needs more debugging..)

* extras/contrib/src/Patches/x264-svn-win32.patch: log x264 thread count as normal info instead of needing debug (extreme logging needing verbose option)

* extras/contrib/src/x264-svn-win32.patch: add threads count debug info to check how many threads x264 will use (threads 0 is default and will make libx264 automagically set threads = cores * 1.5) and whether win32-pthread works as expected (libx264 can report not compiled with pthread and forces threads = 1)

* fix pthread support for x264 on win32
  - Not sure at all the configure.ac part is right, anyone more knowledgable, please check.

contribs: really fix ffmpeg compilation on Darwin (by reverting most of [25028] and adapting its 2nd hunk to fit the upstream code)