[PATCH] MSU Codec and vd_vfw change fixing seg fault
-Added info in codecs.conf to support MSU Screen Capture Lossless Codec -Changed the call to ICDecompress video for windows API to pass output buffer even if frame drop is true. Certain codecs like MSU seem to write to the output buffer even if told to frame skip and this was causing a memory write to 0x00000000.
On Fri, 11 Apr 2008 06:09:27 -0400, AsSlowAsHell wrote:
-Added info in codecs.conf to support MSU Screen Capture Lossless Codec -Changed the call to ICDecompress video for windows API to pass output buffer even if frame drop is true. Certain codecs like MSU seem to write to the output buffer even if told to frame skip and this was causing a memory write to 0x00000000.
there is also another win32 loader patch pending here: http://bugzilla.mplayerhq.hu/show_bug.cgi?id=963 still trying to locate sirenacm.dll and a sample i think. -compn
On Fri, Apr 11, 2008 at 09:09:04AM -0400, Compn wrote:
On Fri, 11 Apr 2008 06:09:27 -0400, AsSlowAsHell wrote:
-Added info in codecs.conf to support MSU Screen Capture Lossless Codec -Changed the call to ICDecompress video for windows API to pass output buffer even if frame drop is true. Certain codecs like MSU seem to write to the output buffer even if told to frame skip and this was causing a memory write to 0x00000000.
there is also another win32 loader patch pending here: http://bugzilla.mplayerhq.hu/show_bug.cgi?id=963 still trying to locate sirenacm.dll and a sample i think.
I've uploaded sirenacm.dll to /incoming/sirenacm I got it by decompressing then decompressing then decompressing :D the msi of latest windows live messenger Regards -- Gianluigi Tiesi <sherpya@netfarm.it> EDP Project Leader Netfarm S.r.l. - http://www.netfarm.it/ Free Software: http://oss.netfarm.it/
On Fri, 18 Apr 2008 06:41:19 +0200, Gianluigi Tiesi wrote:
On Fri, Apr 11, 2008 at 09:09:04AM -0400, Compn wrote:
On Fri, 11 Apr 2008 06:09:27 -0400, AsSlowAsHell wrote:
-Added info in codecs.conf to support MSU Screen Capture Lossless Codec -Changed the call to ICDecompress video for windows API to pass output buffer even if frame drop is true. Certain codecs like MSU seem to write to the output buffer even if told to frame skip and this was causing a memory write to 0x00000000.
there is also another win32 loader patch pending here: http://bugzilla.mplayerhq.hu/show_bug.cgi?id=963 still trying to locate sirenacm.dll and a sample i think.
I've uploaded sirenacm.dll to /incoming/sirenacm I got it by decompressing then decompressing then decompressing :D the msi of latest windows live messenger
Regards
ping rxt , i will apply both patches in a week or two... -compn
On Sat, 19 Apr 2008 18:05:59 -0400 Compn <tempn@twmi.rr.com> wrote:
On Fri, 18 Apr 2008 06:41:19 +0200, Gianluigi Tiesi wrote:
On Fri, Apr 11, 2008 at 09:09:04AM -0400, Compn wrote:
On Fri, 11 Apr 2008 06:09:27 -0400, AsSlowAsHell wrote:
-Added info in codecs.conf to support MSU Screen Capture Lossless Codec -Changed the call to ICDecompress video for windows API to pass output buffer even if frame drop is true. Certain codecs like MSU seem to write to the output buffer even if told to frame skip and this was causing a memory write to 0x00000000.
there is also another win32 loader patch pending here: http://bugzilla.mplayerhq.hu/show_bug.cgi?id=963 still trying to locate sirenacm.dll and a sample i think.
I've uploaded sirenacm.dll to /incoming/sirenacm I got it by decompressing then decompressing then decompressing :D the msi of latest windows live messenger
Regards
ping rxt , i will apply both patches in a week or two...
The sirenacm patch is applied with some extensions. The dll for new codecs not yet included in the code packages can be found at http://www1.mplayerhq.hu/~rtogni/newcodecs/ Ciao, Roberto -- Better is the enemy of good enough.
On Fri, 11 Apr 2008 06:09:27 -0400 AsSlowAsHell <asslowashell@gmail.com> wrote:
-Added info in codecs.conf to support MSU Screen Capture Lossless Codec -Changed the call to ICDecompress video for windows API to pass output buffer even if frame drop is true. Certain codecs like MSU seem to write to the output buffer even if told to frame skip and this was causing a memory write to 0x00000000.
Applied the codecs.conf part of the patch, and added two dummy functions to win32.c to remove some calls to unk_ functions (they are some constructor/destructor for c++ stuff). From the name (std::_Lockit) and by looking on the net it's probably related to memory locking in multithread dll for ms c++, so it would be nice if somebody with a multi-core cpu can test it (on 32-bit os that uses the loader, in windows should be ok). With the only sample I have http://samples.mplayerhq.hu/V-codecs/SCLS/somewins.avi your other patch to vd_vfw.c is not needed. Can you provide a sample that need this fix? Ciao, Roberto -- Better is the enemy of good enough.
On Sun, Apr 20, 2008 at 11:28:19PM +0200, Roberto Togni wrote:
On Fri, 11 Apr 2008 06:09:27 -0400 AsSlowAsHell <asslowashell@gmail.com> wrote:
-Added info in codecs.conf to support MSU Screen Capture Lossless Codec -Changed the call to ICDecompress video for windows API to pass output buffer even if frame drop is true. Certain codecs like MSU seem to write to the output buffer even if told to frame skip and this was causing a memory write to 0x00000000.
Applied the codecs.conf part of the patch, and added two dummy functions to win32.c to remove some calls to unk_ functions (they are some constructor/destructor for c++ stuff). From the name (std::_Lockit) and by looking on the net it's probably related to memory locking in multithread dll for ms c++, so it would be nice if somebody with a multi-core cpu can test it (on 32-bit os that uses the loader, in windows should be ok).
While I agree with testing, our loader always returns 1 as number of CPUs/cores, which should make it less likely to be a problem. Greetings, Reimar Döffinger
participants (5)
-
AsSlowAsHell -
Compn -
Gianluigi Tiesi -
Reimar Döffinger -
Roberto Togni