July 2012 - MPlayer-dev-eng - lists.mplayerhq.hu

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

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

[PATCH] Read normal duration when demuxing subtitles through lavf (v2)
by Philip Langdale 04 Aug '12

04 Aug '12

Hi all, Currently, mplayer only looks at convergence_duration when using lavf to demux subtitles. Some containers use the normal duration field for subtitles (eg: mov_text), so we should read that if the convergence_duration is not used. Thanks, --phil

4 10

[PATCH]Include stdlib from sdl_common.c
by Carl Eugen Hoyos 29 Jul '12

29 Jul '12

Hi! sdl_common.c uses setenv() which needs stdlib.h. I will apply if nobody objects, Carl Eugen

3 8

[PATCH] [3/3] vaapi support for mplayer 1.1
by VoJcEK 29 Jul '12

29 Jul '12

And the one I mentioned that has been reported working and probably most awaited for by Intel and Ati graphic cards users. Regards, rz

3 3

[PATCH] Read normal duration when demuxing subtitles through lavf
by Philip Langdale 27 Jul '12

27 Jul '12

Hi all, Currently, mplayer only looks at convergence_duration when using lavf to demux subtitles. Some containers use the normal duration field for subtitles (eg: mov_text), so we should read that if the convergence_duration is not used. Thanks, --phil

2 2

[PATCH] Play previously played entries in shuffled playlist
by Adam Buechler 27 Jul '12

27 Jul '12

Allows you to play previously played entries again in a shuffled playlist by stepping backward in the playtree. It does this by shuffling the playtree as you step through it, choosing a random song and swapping it with the next entry every time you're about to step onto an entry that hasn't been played before. It's kind of an ugly solution (can it even be called an iterator anymore if it modifies the data structure as it iterates through it?), and I have a feeling it might break anything that assumes that iterating through the playtree won't change the playtree, so I understand if you don't want to include the patch, but I've been using it for the past couple of weeks without problems and it was a feature I missed a lot, so I just wanted to share in case anyone else is interested. The patch is really small, so here it is: Index: playtree.c =================================================================== --- playtree.c (revision 35025) +++ playtree.c (working copy) @@ -574,9 +574,7 @@ iter->mode = PLAY_TREE_ITER_NORMAL; iter->file = -1; - if(iter->mode == PLAY_TREE_ITER_RND) - pt = play_tree_rnd_step(iter->tree); - else if( d > 0 ) { + if( d > 0 ) { int i; pt = iter->tree; for(i = d ; i > 0 && pt ; i--) @@ -591,6 +589,28 @@ } else pt = iter->tree; + if (iter->mode == PLAY_TREE_ITER_RND && (pt == NULL || !(pt->flags & PLAY_TREE_RND_PLAYED))) { + // If the non-random stepping code above hit the end of the playtree or + // hit an entry that hasn't been played before, try to choose a random entry. + pt = play_tree_rnd_step(iter->tree); + + if (pt != NULL && pt != iter->tree) { + // Move the random entry to be immediately before or after the current entry. + play_tree_remove(pt, 0, 0); + + if (d > 0) + // Insert after + play_tree_insert_entry(iter->tree, pt); + else { + // Insert before + if (iter->tree->prev) + play_tree_insert_entry(iter->tree->prev, pt); + else + play_tree_prepend_entry(iter->tree, pt); + } + } + } + if(pt == NULL) { // No next // Must we loop? if (iter->mode == PLAY_TREE_ITER_RND) {

3 6

[PATCH] Recognize audio format mp3 in wav files
by Ingo Brückl 26 Jul '12

26 Jul '12

Subject says it all. I happen to have a wav file that contains mp3 data. MPlayer's default demuxer can't play it, while -demuxer lavf plays it just fine. The patch makes MPlayer's default demuxer to play it as well. If there are no objections, I'd like to commit. Ingo

2 5