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

[PATCH] Test the appropriate header for the system
by Ingo Brückl 21 Feb '16

21 Feb '16

This is an old patch having lain around here. I found it recently when cleaning up some directories. This kind of header check is already done for _struct_addrinfo, _getaddrinfo, _struct_sockaddr_storage and _struct_sockaddr_sa_len, so no need to do it differently and in a kind of brute force way for three other checks. Ingo

1 0

[PATCH] codecs.conf: add pixel formats for ffvp9.
by Nicolas George 21 Feb '16

21 Feb '16

The following pixel formats are present in the source too but have no MPlayer name: AV_PIX_FMT_YUV440P10 AV_PIX_FMT_YUV440P12 AV_PIX_FMT_GBRP10 AV_PIX_FMT_GBRP12 --- etc/codecs.conf | 6 +++++- 1 file changed, 5 insertions(+), 1 deletion(-) Tested (and necessary) with 444P, other formats taken from the source code (read_colorspace_details() in vp9.c). Also tested with GBR24P: it starts working instead of flooding with error messages but the colors are wrong; colors are wrong the same way with ffplay, so I guess it is a bug in lavc, I will investigate if I have time. Wrong colors are still better than a flood of error messages. Will commit shortly if nobody objects, and even faster if someone approves. diff --git a/etc/codecs.conf b/etc/codecs.conf index b18829a..d8010eb 100644 --- a/etc/codecs.conf +++ b/etc/codecs.conf @@ -3015,7 +3015,11 @@ videocodec ffvp9 fourcc VP90 driver ffmpeg dll "vp9" - out YV12 + out YV12,420P10,420P12 + out 422P,422P10,422P12 + out 440P + out 444P,444P10,444P12 + out GBR24P videocodec fflibvpx info "FFmpeg wrapper for libvpx/VP8" -- 2.7.0

4 4

[PATCH] Fix broken build by checking for used structure definition
by Erik Auerswald 20 Feb '16

20 Feb '16

Hi, the attached patch fixes a build error if the header linux/videodev2.h defines struct v4l2_ext_controls, but not struct v4l2_query_ext_ctrl. This is the case on Ubuntu 14.04.4 LTS, for example. Build error: ---8<------8<------8<------8<------8<------8<------8<------8<--- stream/stream_pvr.c: In function 'add_v4l2_ext_control': stream/stream_pvr.c:1099:10: error: variable 'qctrl' has initializer but incomplete type struct v4l2_query_ext_ctrl qctrl = { .id = id }; ^ stream/stream_pvr.c:1099:10: error: unknown field 'id' specified in initializer stream/stream_pvr.c:1099:10: warning: excess elements in struct initializer [enabled by default] stream/stream_pvr.c:1099:10: warning: (near initialization for 'qctrl') [enabled by default] stream/stream_pvr.c:1099:30: error: storage size of 'qctrl' isn't known struct v4l2_query_ext_ctrl qctrl = { .id = id }; ^ stream/stream_pvr.c:1102:27: error: 'VIDIOC_QUERY_EXT_CTRL' undeclared (first use in this function) if (ioctl (pvr->dev_fd, VIDIOC_QUERY_EXT_CTRL, &qctrl) < 0) ^ stream/stream_pvr.c:1102:27: note: each undeclared identifier is reported only once for each function it appears in stream/stream_pvr.c:1099:30: warning: unused variable 'qctrl' [-Wunused-variable] struct v4l2_query_ext_ctrl qctrl = { .id = id }; ^ make: *** [stream/stream_pvr.o] Error 1 make: *** Waiting for unfinished jobs.... ---8<------8<------8<------8<------8<------8<------8<------8<--- Thanks, Erik -- http://www.unix-ag.uni-kl.de/~auerswal/

2 1

[PATCH] fix libass fails with -subcp enca: and external subtitles try 2
by Basin Ilya 20 Feb '16

20 Feb '16

Hi. Here's a proposed patch for https://trac.mplayerhq.hu/ticket/2281 --- mplayer.c | 3 ++- sub/subreader.c | 8 ++++---- sub/subreader.h | 1 + 3 files changed, 7 insertions(+), 5 deletions(-) diff --git a/mplayer.c b/mplayer.c index aa5ca33..017e149 100644 --- a/mplayer.c +++ b/mplayer.c @@ -1085,11 +1085,12 @@ void add_subtitles(char *filename, float fps, int noerr) if (filename == NULL || mpctx->set_of_sub_size >= MAX_SUBTITLE_FILES) return; + enca_sub_cp = NULL; subd = sub_read_file(filename, fps); #ifdef CONFIG_ASS if (ass_enabled) #ifdef CONFIG_ICONV - asst = ass_read_stream(ass_library, filename, sub_cp); + asst = ass_read_stream(ass_library, filename, (enca_sub_cp ? enca_sub_cp : sub_cp)); #else asst = ass_read_stream(ass_library, filename, 0); #endif diff --git a/sub/subreader.c b/sub/subreader.c index ecc5ff5..510c43a 100644 --- a/sub/subreader.c +++ b/sub/subreader.c @@ -54,6 +54,7 @@ #include <iconv.h> #endif char *sub_cp=NULL; +char *enca_sub_cp=NULL; #ifdef CONFIG_FRIBIDI #include <fribidi/fribidi.h> char *fribidi_charset = NULL; ///character set that will be passed to FriBiDi @@ -1212,21 +1213,20 @@ void subcp_open (stream_t *st) char *tocp = "UTF-8"; if (sub_cp){ - const char *cp_tmp = sub_cp; #ifdef CONFIG_ENCA char enca_lang[3], enca_fallback[100]; if (sscanf(sub_cp, "enca:%2s:%99s", enca_lang, enca_fallback) == 2 || sscanf(sub_cp, "ENCA:%2s:%99s", enca_lang, enca_fallback) == 2) { if (st && st->flags & MP_STREAM_SEEK ) { - cp_tmp = guess_cp(st, enca_lang, enca_fallback); + enca_sub_cp = guess_cp(st, enca_lang, enca_fallback); } else { - cp_tmp = enca_fallback; + enca_sub_cp = enca_fallback; if (st) mp_msg(MSGT_SUBREADER,MSGL_WARN,"SUB: enca failed, stream must be seekable.\n"); } } #endif - if ((icdsc = iconv_open (tocp, cp_tmp)) != (iconv_t)(-1)){ + if ((icdsc = iconv_open (tocp, enca_sub_cp)) != (iconv_t)(-1)){ mp_msg(MSGT_SUBREADER,MSGL_V,"SUB: opened iconv descriptor.\n"); sub_utf8 = 2; } else diff --git a/sub/subreader.h b/sub/subreader.h index 0760081..554dbd7 100644 --- a/sub/subreader.h +++ b/sub/subreader.h @@ -33,6 +33,7 @@ extern int sub_no_text_pp; // disable text post-processing extern int sub_match_fuzziness; extern int sub_format; extern char *sub_cp; +extern char *enca_sub_cp; // subtitle formats #define SUB_INVALID -1 -- 2.7.0

2 3

[PATCH] gui/menu: optimize language table.
by Reimar Döffinger 20 Feb '16

20 Feb '16

Store fixed size 4 entry directly instead of via pointer. --- gui/dialog/menu.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/gui/dialog/menu.c b/gui/dialog/menu.c index de6f45f..fcb56fb 100644 --- a/gui/dialog/menu.c +++ b/gui/dialog/menu.c @@ -230,7 +230,7 @@ GtkWidget * AddSeparator( GtkWidget * Menu ) typedef struct { int id; - const char * id2; + const char id2[4]; const char * name; } Languages_t; -- 2.7.0

2 3

MPlayer 1.3.0 release
by Roberto Togni 16 Feb '16

16 Feb '16

Hi, it looks like FFmpeg is planning to do the next release 3.0.0 next week, so that it's out before the next Ubuntu LTS (freeze see http://ffmpeg.org/pipermail/ffmpeg-devel/2016-February/189321.html ). Since MPlayer 1.2.x will not work with FFmpeg 3.0, we need a new release. I have no idea if Ubuntu 16.04 will pick up FFmpeg 3.0, and then MPlayer 1.3.0 (Debian hasn't upgraded to 1.2.1 yet); I just want to be ready in case it happens. Surely other distributions will get FFmpeg 3.0 anyway, and thy will need an upgrade path from 1.2.1. How to proceed? If we want to reach the Ubuntu milestone we should make the 1.3 branch on Monday and the release on Thursday (or later if FFmpeg 3.0 is not out yet). It's a very short time frame. Also I won't be available on Wednesday and Tuesday, so if FFmpeg is late somebody else will have to do it. Comments and suggestions? Should we try to do it, of just forget about it and release 1.3.0 in the next weeks? (over the next weekend or the one after that). Regarding the release itself: - Pending patches: I think the BSD build regression fix from Alexander should be committed; I'm not aware of other critical fixes. If you have anything safe that you want to see in the release please commit it now. - FFmpeg: 1.3.0 will include unpatched FFmpeg 3.0 release; all the bugs known to me with the new sws filter are fixed. - Release notes: volunteers? Else I'll write e a few sentences. Main points: compatibility with 3.0, now requires libswresample, usual buzzwords about new features. - Release name: if no suggestions, it will be "worksforme" Ciao, Roberto

2 5

Re: [MPlayer-dev-eng] [MPlayer-users] Build regression on FreeBSD
by Alexander Strasser 15 Feb '16

15 Feb '16

Hi Thomas, AFAICT there are multiple problems at work, that make a quick solution rather involved or hackish. That's why I have Cc'ed MPlayer development mailing list. On 2015-12-19 20:21 +0100, Thomas Zander wrote: > porting an up-to-date snapshot to FreeBSD I noticed two build > regressions and want to kindly ask for resolving those upstream: > > 1) Makefile (at root dir) > codec-cfg.c now includes <vdpau/vdpau.h> > On FreeBSD this file ends up in LOCALBASE (by default > /usr/local/include) which is not included in the default compiler > search path, so in its build target > codec-cfg$(EXESUF) .... the compile command > $(HOST_CC) $(HOSTCFLAGS) -o $@ $< > is not sufficient. The include path for localbase needs to be added. I guess r37454 introduced this problem, by including vdpau.h in our img_format.h header. It cannot be reverted because that would probably break the build with future FFmpeg versions. The problem is if we want to use external stuff, which cannot be forward declared, in img_format.h we need to add additional flags to HOSTCFLAGS as well. Furthermore it is interesing that vdpau gets detected at all on your system. I suspect this is because of manual addition to CFLAGS or because of modification by previous configure tests. I might be missing a few things, but all in all I do not see why we need the structure vdpau_frame_data in codec-cfg.c source. So maybe the structure should be defined in a different header? All in all I see no really easy way to set HOSTCFLAGS correctly. Additionally I am not sure if the work is justified, because if I am not mistaken above it is not needed in codec-cfg. [...] Alexander

4 12

[PATCH] input.c: optimize mp_cmd list.
by Reimar Döffinger 14 Feb '16

14 Feb '16

Reduce its size and encode strings directly to avoid relocations. --- gui/interface.c | 2 +- gui/ui/actions.c | 2 +- input/input.c | 8 ++------ input/input.h | 4 ++-- 4 files changed, 6 insertions(+), 10 deletions(-) diff --git a/gui/interface.c b/gui/interface.c index 30a564f..3d98998 100644 --- a/gui/interface.c +++ b/gui/interface.c @@ -1262,7 +1262,7 @@ void mplayer(int what, float value, void *data) mp_cmd = calloc(1, sizeof(*mp_cmd)); mp_cmd->id = MP_CMD_PANSCAN; - mp_cmd->name = strdup("panscan"); + strcpy(mp_cmd->name, "panscan"); mp_cmd->args[0].v.f = value; mp_cmd->args[1].v.i = 1; mp_input_queue_cmd(mp_cmd); diff --git a/gui/ui/actions.c b/gui/ui/actions.c index 4304817..2728986 100644 --- a/gui/ui/actions.c +++ b/gui/ui/actions.c @@ -578,7 +578,7 @@ void uiPause(void) if (cmd) { cmd->id = MP_CMD_PAUSE; - cmd->name = strdup("pause"); + strcpy(cmd->name, "pause"); mp_input_queue_cmd(cmd); } } else diff --git a/input/input.c b/input/input.c index 4f80e68..1a52efe 100644 --- a/input/input.c +++ b/input/input.c @@ -212,7 +212,7 @@ static const mp_cmd_t mp_cmds[] = { { MP_CMD_GUI, "gui", 1, { {MP_CMD_ARG_STRING, {0}}, {-1,{0}} } }, - { 0, NULL, 0, {} } + { 0, "", 0, {} } }; /// The names of the keys as used in input.conf @@ -831,7 +831,7 @@ mp_input_parse_cmd(char* str) { cmd = calloc(1, sizeof(mp_cmd_t)); cmd->id = cmd_def->id; - cmd->name = strdup(cmd_def->name); + memcpy(cmd->name, cmd_def->name, sizeof(cmd->name)); if (pausing == -1) { switch (cmd->id) { case MP_CMD_KEYDOWN_EVENTS: @@ -1432,8 +1432,6 @@ mp_cmd_free(mp_cmd_t* cmd) { //#endif if ( !cmd ) return; - free(cmd->name); - for(i=0; i < MP_CMD_MAX_ARGS && cmd->args[i].type != -1; i++) { if(cmd->args[i].type == MP_CMD_ARG_STRING) free(cmd->args[i].v.s); @@ -1451,8 +1449,6 @@ mp_cmd_clone(mp_cmd_t* cmd) { ret = malloc(sizeof(mp_cmd_t)); memcpy(ret,cmd,sizeof(mp_cmd_t)); - if(cmd->name) - ret->name = strdup(cmd->name); for(i = 0; i < MP_CMD_MAX_ARGS && cmd->args[i].type != -1; i++) { if(cmd->args[i].type == MP_CMD_ARG_STRING && cmd->args[i].v.s != NULL) ret->args[i].v.s = strdup(cmd->args[i].v.s); diff --git a/input/input.h b/input/input.h index e723618..3e1b463 100644 --- a/input/input.h +++ b/input/input.h @@ -174,7 +174,7 @@ typedef enum { #define MP_CMD_ARG_VOID 3 #ifndef MP_CMD_MAX_ARGS -#define MP_CMD_MAX_ARGS 10 +#define MP_CMD_MAX_ARGS 6 #endif // Error codes for the drivers @@ -215,7 +215,7 @@ typedef struct mp_cmd_arg { typedef struct mp_cmd { int id; - char* name; + char name[24]; int nargs; mp_cmd_arg_t args[MP_CMD_MAX_ARGS]; int pausing; -- 2.7.0

2 7

[PATCH] vf_filmdint: make code slightly less confusing.
by Reimar Döffinger 14 Feb '16

14 Feb '16

The number of blocks should be the same for both, so it shouldn't matter which we use, but it seems more obvious to use the matching one, and it silences Coverity. --- libmpcodecs/vf_filmdint.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/libmpcodecs/vf_filmdint.c b/libmpcodecs/vf_filmdint.c index 2b4ee09..2e43bc1 100644 --- a/libmpcodecs/vf_filmdint.c +++ b/libmpcodecs/vf_filmdint.c @@ -1291,7 +1291,7 @@ static int put_image(struct vf_instance *vf, mp_image_t *mpi, double pts) (s->low.noise + s->interlaced_low < (s->num_blocks>>8) || s->sad.noise < 160)) || ((show_fields & 12) == 12 && - (ps->low.noise + ps->interlaced_low < (s->num_blocks>>8) || + (ps->low.noise + ps->interlaced_low < (ps->num_blocks>>8) || ps->sad.noise < 160))) { p->export_count++; dmpi = vf_get_image(vf->next, mpi->imgfmt, MP_IMGTYPE_EXPORT, -- 2.7.0

1 0

[PATCH] vf_divtc.c: Ensure shift is always valid.
by Reimar Döffinger 14 Feb '16

14 Feb '16

Not sure if this can actually happen, and if it might impact performance (on x86 the compiler should optimize the "and" away anyway). Just thinking it might be safer, and it avoids a Coverity warning. --- libmpcodecs/vf_divtc.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/libmpcodecs/vf_divtc.c b/libmpcodecs/vf_divtc.c index 5a23487..111e16f 100644 --- a/libmpcodecs/vf_divtc.c +++ b/libmpcodecs/vf_divtc.c @@ -168,7 +168,7 @@ static unsigned int checksum_plane(unsigned char *p, unsigned char *z, t=be2me_32(wsum); #endif - for(sum^=(t<<shift|t>>(32-shift)); p<e;) + for(sum^=(t<<shift|t>>((32-shift)&31)); p<e;) sum^=*p++<<(shift=(shift-8)&31); } -- 2.7.0

1 0