to ignore information about the _actual_ buffer size returned by the
kernel and continued to use the requested buffer size (the old fix was
to request smaller buffers, but not ignoring information about the
actual buffer size is way better)
git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@11377 b3059339-0415-0410-9bf9-f77b7e298cf2
now filter chains with >8 scale filters initalize before u fall asleep
git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@11374 b3059339-0415-0410-9bf9-f77b7e298cf2
This one fixes a sig11 when the -vo can't be initialized and get_buffer fails
for libavcodec
git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@11373 b3059339-0415-0410-9bf9-f77b7e298cf2
Following patch adds ID of Radeon 9800 (non pro) (R350) to mplayer pci
database, thus enabling vidix for it. Vidix works great on it... Well,
not exactly great, just the same way it works on other radeons.
git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@11372 b3059339-0415-0410-9bf9-f77b7e298cf2
of course ffmpeg svq3 decoder will crash if you don't give it the ImageDesc!
now it works fine
git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@11371 b3059339-0415-0410-9bf9-f77b7e298cf2
at least they're better than before, where "progressive"/"telecine"
meant the opposite of what they should!
git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@11368 b3059339-0415-0410-9bf9-f77b7e298cf2
root by the '#' prompt, as suggested by Lukasz frogu Proszek <l_j_p@wp.pl>.
git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@11365 b3059339-0415-0410-9bf9-f77b7e298cf2
the HTML docs long ago. Somebody forgot to sync and should start drinking
his share of Coke soon. Prime suspect is a developer whose name starts
with the letter 'D'.
git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@11359 b3059339-0415-0410-9bf9-f77b7e298cf2
this patch fixes a recently discovered bug for which DVB-C users
couldn't tune
(wrong parsing of the config file and incorrect parameter passing to
tune_it())
and includes the still unapplied patch posted in date 6/9/2003:
- it works correctly with and without caches; in the former case it
doesn't take anymore a lot of time to empty the cache before changing channel;
the uninit_cache() function is called in mplayer.c just after
the new tuning operation
- initialized a variable identifying the tuner type, and exit if it
isn't supported
- doesn't crash anymore when
1) the channels file doesn't exists
2) the tuner is used by another application
3) in the menu, when trying to select a channel before the first
4) some mp_msg() called in case of error
git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@11353 b3059339-0415-0410-9bf9-f77b7e298cf2
original patch by Thomas Glanzmann <sithglan@stud.uni-erlangen.de>
git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@11352 b3059339-0415-0410-9bf9-f77b7e298cf2
stream with multibyte characters in the url.
There is a bug in string_utf16() of libmpdemux/asf_mmst_streaming.c:
the url encoding code is written for single-byte character only. It
just fill a zero byte between every two adjacent bytes of the url
string. This is wrong when the url string contains multi-byte
characters.
My patch is in the attachment, and has been tested working. It uses
iconv() to convert the url's encoding to utf-16 correctly. So can this
patch be accepted and commit into CVS?
patch by Wang WenRui <wangwr@mail.ustc.edu.cn>
git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@11351 b3059339-0415-0410-9bf9-f77b7e298cf2
