September 2012 - MPlayer-dev-eng

Homepage skin download issue
by Ingo Brückl 09 Sep '15

09 Sep '15

Yesterday, I updated hwswskin to 1.3, but it still isn't downloadable. Is the generation of the skin download tarballs working? Ingo

3 8

Homepage DOCS needs update
by Ingo Brückl 24 Apr '14

24 Apr '14

If it's already on the to-do list ignore, but the documentation on the homepage doesn't get updated. Ingo

3 12

[PATCH] xvmc: handle shm xvimage allocation error
by Marcin Slusarz 06 Jan '14

06 Jan '14

- handle xvimage/shm allocation failure and fallback to non-shm xv - assert xvimage->data_size!=0 because 0 size will lead to memory corruptions (I hit it because of bug in Nouveau's early xvmc implementation) Index: libvo/vo_xvmc.c =================================================================== --- libvo/vo_xvmc.c (wersja 34098) +++ libvo/vo_xvmc.c (kopia robocza) @@ -181,6 +181,15 @@ return cur_render->state || (cur_render->mpi && cur_render->mpi->usage_count); } +static void allocate_xvimage_non_shm(int xvimage_width, int xvimage_height, int xv_format) +{ + xvimage = (XvImage *) XvCreateImage(mDisplay, xv_port, xv_format, NULL, + xvimage_width, xvimage_height); + assert(xvimage->data_size); + xvimage->data = malloc(xvimage->data_size); + XSync(mDisplay,False); +} + static void allocate_xvimage(int xvimage_width,int xvimage_height,int xv_format) { /* @@ -188,7 +197,8 @@ * mit-shm this will bomb... trzing to fix ::atmos */ #ifdef HAVE_SHM - if ( mLocalDisplay && XShmQueryExtension( mDisplay ) ) Shmem_Flag = 1; + if ( mLocalDisplay && XShmQueryExtension( mDisplay ) ) + Shmem_Flag = 0; else { Shmem_Flag = 0; @@ -198,25 +208,41 @@ { xvimage = (XvImage *) XvShmCreateImage(mDisplay, xv_port, xv_format, NULL, xvimage_width, xvimage_height, &Shminfo); + if (!xvimage) + goto noshmimage; + assert(xvimage->data_size); Shminfo.shmid = shmget(IPC_PRIVATE, xvimage->data_size, IPC_CREAT | 0777); + if (Shminfo.shmid == -1) + goto shmgetfail; + Shminfo.shmaddr = (char *) shmat(Shminfo.shmid, 0, 0); + if ((long)Shminfo.shmaddr == -1) + goto shmatfail; + Shminfo.readOnly = False; xvimage->data = Shminfo.shmaddr; - XShmAttach(mDisplay, &Shminfo); + if (!XShmAttach(mDisplay, &Shminfo)) + goto shmattachfail; + XSync(mDisplay, False); shmctl(Shminfo.shmid, IPC_RMID, 0); } else #endif - { - xvimage = (XvImage *) XvCreateImage(mDisplay, xv_port, xv_format, NULL, xvimage_width, xvimage_height); - xvimage->data = malloc(xvimage->data_size); - XSync(mDisplay,False); - } -// memset(xvimage->data,128,xvimage->data_size); + allocate_xvimage_non_shm(xvimage_width, xvimage_height, xv_format); return; + +shmattachfail: + shmdt(Shminfo.shmaddr); +shmatfail: + shmctl(Shminfo.shmid, IPC_RMID, 0); +shmgetfail: + XFree(xvimage); +noshmimage: + Shmem_Flag = 0; + allocate_xvimage_non_shm(xvimage_width, xvimage_height, xv_format); } static void deallocate_xvimage(void)

4 6

Re: [MPlayer-dev-eng] MPlayer fails to build with live-2012.08.20
by Ross Finlayson 22 Apr '13

22 Apr '13

> Does adding > #define RTSPCLIENT_SYNCHRONOUS_INTERFACE 1 > to the configure test also fixes this problem? It probably will. I should warn you, though, that the synchronous "RTSPClient" interface is considered deprecated, and will eventually be removed. You should plan on updating your LIVE555 interface code to use the new asynchronous interface instead (as VLC did a couple of years ago). Ross Finlayson Live Networks, Inc. http://www.live555.com/

2 2

GUI and vo_x11 functions
by Ingo Brückl 02 Mar '13

02 Mar '13

I'm about to commit the vo_x11_ewmh_fullscreen() patch which adds a window parameter to the function, but I'm not quite sure whether this really is ok, although I got permission. The question is whether vo_x11_ewmh_fullscreen() and all the other vo_x11 functions should only operate on the global vo variables (like vo_window in that case) or whether they should be reusable functions, offering services for whoever wants to call them (the GUI in my case). The reason why I raise this issue is that the GUI does call three other vo_x11 functions (with relative success), namely vo_x11_decoration() vo_x11_sizehint() vo_x11_setlayer() While vo_x11_decoration() seems to be problem-free (it has some useful parameters, but uses global variables as well), vo_x11_sizehint() mostly uses global variables (which causes some X errors with the GUI then if called at a time when these variables aren't yet set). vo_x11_setlayer() is pointlessly called by the GUI, because it doesn't do anything in case of WinID >= 0. So, what is the right way to continue? Setting all necessary global vo variables in the GUI - even prior to the vo driver (vo_window is known to become WinID later anyway, for example), trying to modestly patch these four functions so that they can be reused or copy the code and letting the GUI use its own copies? All these four functions are window specific and window handling should be GUI's task although commonly used code is convenient. Pros and cons should be obvious. Ingo

3 16

how to enable hardware acceleration video decoding based on stagefright of android ?
by lean fu 14 Jan '13

14 Jan '13

hi list: ffmpeg 0.10 or later has "libstagefright.cpp" providing hardware acceleration on android , and there's an building script at " ffmpeg/tool/build_libstagefright" for compiling this feature , but unfortunate, 'configure' of mplayer don't support "--enable-libstagefright-h264" , i have to modify config.h and config.mak to add some stagefright related macro manually. and change the 'name' member of ff_libstagefright_h264_decoder from "libstagefright_h264" to "h264". after these changes, when play a file contains h264, procedure will call into 'Stagefright_decode_frame'. but the picture looks abnormal , sometimes appears full blue(or green) screen,for some 1080p files,even throw fatal errors. had somebody get the stagefright features works successfully? and is there some more simple way that i can integrate the stagefright feature into my android port of mplayer ? thanks !

3 3

[PATCH] make clean should remove all files generated by make
by Reimar Döffinger 14 Nov '12

14 Nov '12

Hello, I don't know what the idea behind the current split between "clean" and "distclean" is. But IMHO "make clean" should remove all files that "make" can regenerate, only leaving stuff e.g. generated by configure. Do you agree? Index: Makefile =================================================================== --- Makefile (revision 35213) +++ Makefile (working copy) @@ -972,17 +972,17 @@ clean: -$(MAKE) -C ffmpeg $@ -rm -rf tests/res - -rm -f $(call ADD_ALL_DIRS,/*.o /*.a /*.ho /*~) + -rm -f $(call ADD_ALL_DIRS,/*.o /*.d /*.a /*.ho /*~) -rm -f $(call ADD_ALL_EXESUFS,mplayer mencoder) + -rm -f $(VIDIX_PCI_FILES) + -rm -f $(call ADD_ALL_EXESUFS,codec-cfg cpuinfo) + -rm -f codecs.conf.h help_mp.h version.h + -rm -f DOCS/xml/html-chunk.xsl DOCS/xml/html-single.xsl -rm -rf DOCS/tech/doxygen DOCS/HTML distclean: clean testsclean toolsclean driversclean dhahelperclean -$(MAKE) -C ffmpeg $@ - -rm -f DOCS/xml/html-chunk.xsl DOCS/xml/html-single.xsl - -rm -f $(call ADD_ALL_DIRS,/*.d) - -rm -f config.* codecs.conf.h help_mp.h version.h TAGS tags - -rm -f $(VIDIX_PCI_FILES) - -rm -f $(call ADD_ALL_EXESUFS,codec-cfg cpuinfo) + -rm -f config.* TAGS tags doxygen: doxygen DOCS/tech/Doxyfile

4 4

[PATCH] vd_null to output black frames
by Xidorn Quan 10 Nov '12

10 Nov '12

Hi, This patch makes vd_null return black frames for various image format instead of only NULL. Regards, Xidorn Quan

4 26

[PATCH] really remove mp3lib now
by Reimar Döffinger 21 Oct '12

21 Oct '12

Hello, this time my intention is to let nothing but an angry mob stop me, mpg123 seems to be really, really, _really_ close and maintained, if someone isn't satisfied with its performance it would be better to spend time improving it than hanging onto our copy IMHO. Patch attached in case someone sees any specific issues with it. I won't give it more than a few days.

5 6

[PATCH] fix direct.h io.h
by Stephen Sheldon 15 Oct '12

15 Oct '12

Something changed in ffmpeg. The attached patch made mplayer compile for me on mingw64.

2 5