Original commit message from CVS:
I checked lame packages and source code and they seem to want lame.h in
prefix/include/lame.h
so I fixed stuff accordingly.
Do any systems have lame in include/lame/lame.h ?
If so, mail me and we'll work it out.
Original commit message from CVS:
- my mistake... mpeg2dec libs does not depend on X11 libraries
(vo library distributed with mpeg2dec depends on X11 libs).
Add configure option so user can specify where mpeg2dec library
is installed. This way we do not guess where is mpeg2dec.
If it is not in standard location (prefix==/usr), then leave it
to the user.
Changes inspired by AM_PATH_LIBMIKMOD.
Original commit message from CVS:
- use MPEG2DEC_CFLAGS and MPEG2DEC_LIBS which are set with
GST_CHECK_MPEG2DEC macro
- gcc -O3 option removed - it is set by the user if he wants such
optimization
Original commit message from CVS:
- GST_CHECK_MPEG2DEC macro - checks for mpeg2dec in /usr/lib.
If not found then checks for mpeg2dec in /usr/X11R6/lib.
This macro is needed because mpeg2dec depends on X11R6 libraries
and as FHS states, it should be placed in /usr/X11R6 tree but
some distros do not care and put mpeg2dec into /usr/lib.
It depends on CHECK_LIBHEADER macro.
Original commit message from CVS:
- CHECK_LIBHEADER macro, which is based on GST_CHECK_LIBHEADER
the main difference between them is that you can specify cpp flags with
CHECK_LIBHEADER
Original commit message from CVS:
please don't try to change what is going to be build in Makefile.am files
everything is done using a USE_ conditional, so set it to no in
configure.ac if you don't want it. This one is a good example.
Don't know if qcam needs nasm though : is checking for nasm a check for
what hardware platform ? If so we need to change this.
In any case, let's not "fix" "problems" like this because this is how
cruft comes in. Keep it in a local tree and bring it up so we can fix it
right ;)
Original commit message from CVS:
... and cvs commits should not be allowed for people who spent six hours
on the phone talking to hardware vendors and struggling with failing raid
systems
Original commit message from CVS:
how do people expect sdl to build if make has no way of getting to the dir
and we don't point it in the right direction ?
Original commit message from CVS:
* remove win32 path stuff (not used anymore?)
* check for core 0.3.1 not 0.3.0
* update Xv check for Xv_pic systems
include text explaining it so check is not removed (again)
* remove GLIB_ vars since we don't set those vars
they are assumed to be in GST_ vars
* fix festival check - lib not used, just net connection to a server
* better FLAC test
* remove some core configure.ac leftovers
