July 2001 - MPlayer-dev-eng - lists.mplayerhq.hu

Re[3]: [Mplayer-users] 0.18pre4 released with native divx codec :)
by Felix Bünemann 07 Jul '01

07 Jul '01

Hello Arpi, Saturday, July 07, 2001, 6:49:24 PM, you wrote: A> Hi, >> On Sat, Jul 07, 2001 at 12:21:39PM +0200, Felix Buenemann wrote: >> > > that driver type. With your patch it is not possible to give that option >> > > into config file to TRY to force driver type but not disturb formats which >> > > hasn't got codec inside that driver type set. Or something :) I hope I was >> > > clear enough but who knows ... >> > > >> > > > so you can do something like ./mplayer foo.avi -ac mp3acm -vc ffdivx >> > > >> > > Nice, but as I mentioned above you CAN'T insert these options to your >> > > config file!! I didn't want to FORCE a special codec BUT TRY to force >> > > a DRIVER TYPE (not codec). You option is usefull but somewhat differs from >> > > my goals so I think patches of both of us are usefull and for different >> > > things. (It would be usefull to force a dedicated audio/video codec for >> > > playing but it would be usefull to try to use codec from a driver family >> > > as well. These are not the same!!) >> > I know, these switches are of course useless for config file, but if you >> > always want to prefer ffmpeg codecs you can simply move them up in codecs.conf >> > but we may simply allow bothe methods, so have -ac and -vc and also have -afm >> > and -vfm how about that? >> >> No. You may want to test performance of ffmpeg codecs only and you don't want >> to edit codecs.conf each time. I wanted to avoid this constant editing with >> my new option. You may want to use to test several ffmpeg codecs for a while >> on different type of movies (different codecs within ffmpeg codecs "familiy"). >> So it's simple to put a vcodec=5 line into your config file and you needn't >> worry about playing movies can't handled by ffmpeg codecs because mplayer >> would fall back other type of codecs in this way and so on. >> >> Maybe I'm wrong but imho now there is NO option to select codec driver >> FAMILY (not an individual codec, a set of codecs belong to the same driver >> type). For testing purposes I will keep my modification (though I'm not sure >> to mix it with your patch and commit or not) because it's much more easy to >> test several things in this way (dshow/nodshow switches were similar to my >> option but those were able to select between dshow/nodshow codecs only. >> I want to extend this to any type with -vcodec). >> >> Well, this is not a flame thread though it begins to become that :) A> Ok. Then my opinion: A> We need both -ac/-vc options (to select a single driver by name, and fail A> if it can't be selected) and -afm/-vfm to select codec family (it was A> 'format' before, but 'fm' is goof for 'family' too :)) A> This -afm/-vfm stuff should fallback to autodetect, if no usable codec A> found in the selected driver (it will solve Nick's -afm 1 vs. AC3 problem too) ->> it means that -afm should be fixed too (it will force driver, A> doesn't matter that if works...) and -vfm added again. Yea, exactly what I thought (even the stuff with format like family :), I'll code it this night and commit in the morning. A> If we are alredy here, then the status field of codec descriptors should A> be handled too. For exampel, simply skip drivers with 'crashing' status... A> (skip if there are other working codecs for this format, and print A> warning message if no better codec found) hmm, but currently there is no predefined stuff yet for status field so that should be defined first before adding that. A> and we should move to -dev-eng list... just done. A> A'rpi / Astral & ESP-team btw. sorry I mail with windooze but my linux mailprog is currently fucked up and here I have the whole maildatabase, too. :( -- Best regards, Felix mailto:atmosfear@users.sourceforge.net _______________________________________________ Mplayer-dev-eng mailing list Mplayer-dev-eng(a)lists.sourceforge.net http://lists.sourceforge.net/lists/listinfo/mplayer-dev-eng

1 0

Re: [Mplayer-dev-eng] MMS crops up again... ;-)
by Mayuresh A Kathe 06 Jul '01

06 Jul '01

> For now I don't have much time to focus on it, but I will try > at least to fix the bug I have with asf streaming as soon as > I change the crappy gcc 2.96 :( > But sure the buffering/cache layer will help to improve > the streaming. Cool... By when can we expect _that_ to happen? Awaiting your reply. ~Mayuresh _______________________________________________ Mplayer-dev-eng mailing list Mplayer-dev-eng(a)lists.sourceforge.net http://lists.sourceforge.net/lists/listinfo/mplayer-dev-eng

1 0

Qt/Emb Interface to MPlayer FB?
by Mayuresh A Kathe 06 Jul '01

06 Jul '01

Hi, There was some work going on at creating a Qt/Emb Interface to MPlayerFB isn't it? Any ideas where I can lay my hands on it? BTW, I haven't been able to run MPlayerFB on my system, though a couple of other FB programs have run, MPlayer refuses to run. Any ideas what might be the problem? Awaiting your reply. ~Mayuresh _______________________________________________ Mplayer-dev-eng mailing list Mplayer-dev-eng(a)lists.sourceforge.net http://lists.sourceforge.net/lists/listinfo/mplayer-dev-eng

1 0

MMS crops up again... ;-)
by Mayuresh A Kathe 05 Jul '01

05 Jul '01

Hi, A while ago there was a whole furore of activity on the list regarding MMS support, the solution provided by Tibu and others worked absolutely well. But parallely there was some talk going on about MPlayer having built in MMS support, is there anything happening over there? Awaiting your reply. ~Mayuresh _______________________________________________ Mplayer-dev-eng mailing list Mplayer-dev-eng(a)lists.sourceforge.net http://lists.sourceforge.net/lists/listinfo/mplayer-dev-eng

3 4

RE: [Mplayer-dev-eng] Re: Experimental patch for optimized "wine" build
by Juergen Keil 05 Jul '01

05 Jul '01

> > Has anyone tested/played with the patch? Did it break anything? > It breaks syncing with avifile. You should send this patch & description to > the avifile authors (Eugene Kuznetsov), so they can accept it and apply > to their cvs. Then i can merge again and we'll have this fixed :) I've done that already, both Eugene Kuznetsov and Zdenek Kabelac have a copy of that patch for avifile-0.6-cvs, but upto now, nothing appeared in their CVS tree. (Zdenek may be on vacation now(?), no checkins from him in last few days... ;-) -- Jürgen Keil jk(a)tools.de Tools GmbH +49 (228) 9858011 _______________________________________________ Mplayer-dev-eng mailing list Mplayer-dev-eng(a)lists.sourceforge.net http://lists.sourceforge.net/lists/listinfo/mplayer-dev-eng

2 1

Fwd: Benchmarks -largefiles
by Arpi 04 Jul '01

04 Jul '01

Ok. LGB? --------- Forwarded message --------- From: Stephen Davies <steve(a)daviesfam.org> To: arpi(a)thot.banki.hu Subject: Benchmarks -largefiles Hi Arpi, To try to put your mind at rest regarding performance impact of the long-long's in the largefile patch, here are some benchmarks. I tested three mplayer versions against a 30 minute, 4Mb/sec MPEG1 file - first the standard source from cvs today: BENCHMARKs: V: 267.756s VO: 0.000s A: 0.000s Sys: 39.712s = 307.469s BENCHMARK%: V: 87.0841% VO: 0.0000% A: 0.0000% Sys: 12.9159% = 100.0000% BENCHMARKs: V: 267.308s VO: 0.000s A: 0.000s Sys: 39.710s = 307.018s BENCHMARK%: V: 87.0660% VO: 0.0000% A: 0.0000% Sys: 12.9340% = 100.0000% BENCHMARKs: V: 267.364s VO: 0.000s A: 0.000s Sys: 39.519s = 306.883s BENCHMARK%: V: 87.1224% VO: 0.0000% A: 0.0000% Sys: 12.8776% = 100.0000% 921.370 Then, I applied the largefiles patch, but compiled with largefiles disabled (consequently only code change being one or two conversions to long long datatype. Here are the results: BENCHMARKs: V: 267.451s VO: 0.000s A: 0.000s Sys: 39.810s = 307.261s BENCHMARK%: V: 87.0435% VO: 0.0000% A: 0.0000% Sys: 12.9565% = 100.0000% BENCHMARKs: V: 267.371s VO: 0.000s A: 0.000s Sys: 39.804s = 307.174s BENCHMARK%: V: 87.0420% VO: 0.0000% A: 0.0000% Sys: 12.9580% = 100.0000% BENCHMARKs: V: 267.856s VO: 0.000s A: 0.000s Sys: 39.494s = 307.351s BENCHMARK%: V: 87.1500% VO: 0.0000% A: 0.0000% Sys: 12.8500% = 100.0000% .046% diff 921.786 0.046% slower - I doubt that this is statistically siginificant even. I then compiled with 64bit file offsets enabled: BENCHMARKs: V: 267.376s VO: 0.000s A: 0.000s Sys: 39.753s = 307.128s BENCHMARK%: V: 87.0567% VO: 0.0000% A: 0.0000% Sys: 12.9433% = 100.0000% BENCHMARKs: V: 268.210s VO: 0.000s A: 0.000s Sys: 39.557s = 307.767s BENCHMARK%: V: 87.1470% VO: 0.0000% A: 0.0000% Sys: 12.8530% = 100.0000% BENCHMARKs: V: 267.135s VO: 0.000s A: 0.000s Sys: 39.894s = 307.029s BENCHMARK%: V: 87.0066% VO: 0.0000% A: 0.0000% Sys: 12.9934% = 100.0000% .060% diff 921.924 0.060% slower. Hope that that satisfies you with respect to performance? If you feel its still an issue I'll find you a 1% saving somewhere else - thats 20 times as much ;-) Regards, Steve _______________________________________________ Mplayer-dev-eng mailing list Mplayer-dev-eng(a)lists.sourceforge.net http://lists.sourceforge.net/lists/listinfo/mplayer-dev-eng

1 0

Re: Experimental patch for optimized "wine" build for mplayer
by Juergen Keil 04 Jul '01

04 Jul '01

Hi, A few day ago, I wrote: > Are there perhaps some .avi / .asf files that work now, that did not work > without the patch? Now I've found the first avi file (http://static.hugi.is/movies/405.avi ~8MB) that works with the optimized and fixed wine loader, and that crashes with the unpatched mplayer-cvs of today: ________________________________________ With today's cvs version, I get % ./mplayer -ao null /tmp/405.avi MPlayer 0.17cvs-010704-12:44(C) 2000-2001 Arpad Gereoffy <arpi(a)thot.banki.hu> Reading /home/jk/.mplayer/codecs.conf: 15 audio & 28 video codecs Font /home/jk/.mplayer/font/font.desc loaded successfully! (140 chars) Detected AVI file format! ==> Found video stream: 0 ==> Found audio stream: 1 VIDEO: [DIV3] 356x240 24bpp 30.00 fps 335.5 kbps (41.0 kbyte/s) [V] filefmt:3 fourcc:0x33564944 size:356x240 fps:30.00 ftime:=0.0333 Found audio codec: [divx] drv:4 (DivX audio (WMA)) Loading DLL: /usr/lib/win32/divxa32.acm OK Segmentation fault % adb mplayer core core file = core -- program ``mplayer'' on platform 'i86pc' SIGSEGV: Segmentation Fault $c acmDriverOpen(0x8046d60,0x81dd218,0x0,0x8087a27,0x0,0x66) + b7 acmStreamOpen(0x81dde9c,0x0,0x81dd1f0,0x81dde88,0x0,0x0) + 1fb init_acm_audio_codec(0x2,0x804709c,0x80470a8) + a2 _start() + 57 $ no process SIGSEGV: Segmentation Fault gs 0x10f ecx 0x0 fs 0x37 eax 0x81e0aa0 _END_+0x4474 es 0x1f trapno 0xe ds 0x1f err 0x7 edi 0x0 eip 0x80875cf acmDriverOpen+0xb7 esi 0x81dd218 _END_+0xbec cs 0x17 ebp 0x8046d18 efl 0x10246 kesp 0xef8bbfe4 esp 0x8046cc4 ebx 0x80872d3 DrvOpen+0xc3 ss 0x1f edx 0x8246a68 acmDriverOpen+0xb7: movl %eax,0x4(%ebx) [-,4] <ebx/X data address not found <ebx=X 80872d3 <-- That's some return address pointing into mplayer's .text area (DrvOpen+0xc3). Of cause we can't write there :-) ^D With my wine/loader patch installed, this avi file displays without a problem: ... VIDEO: [DIV3] 356x240 24bpp 30.00 fps 335.5 kbps (41.0 kbyte/s) [V] filefmt:3 fourcc:0x33564944 size:356x240 fps:30.00 ftime:=0.0333 Found audio codec: [divx] drv:4 (DivX audio (WMA)) Loading DLL: /usr/lib/win32/divxa32.acm OK dec_audio: Allocating 16384 + 65536 = 81920 bytes for output buffer AUDIO: srate=44100 chans=2 bps=2 sfmt=0x10 ratio: 8010->176400 Found video codec: [divxds] drv:4 (DivX ;-) (MS MPEG-4 v3)) vo: X11 color mask: FFFFFF (R:FF0000 G:FF00 B:FF) vo: X11 running at 1600x1200 depth: 32 ("elise:0" => remote display) Loading DLL: /usr/lib/win32/divx_c32.ax OK Using DirectShow codec: divx_c32.ax DivX setting result = 0 VO: [x11] 356x240 => 356x240 BGR 32-bit VO: Description: X11 ( XImage/Shm ) VO: Author: Aaron Holtzman <aholtzma(a)ess.engr.uvic.ca> Shared memory not supported Reverting to normal Xlib AO: [null] 44100Hz Stereo Signed 16-bit (Little-Endian) AO: Description: Null audio output AO: Author: A'rpi Initial frame delay A: 1 V: 0 AVI Initial frame delay: 0.033 Start playing... ... ________________________________________ I've placed an updated version of my wine/loader patch here: http://www.tools.de/solaris/mplayer/mplayer-wineopt-patch Has anyone tested/played with the patch? Did it break anything? Any objections if I just check in this patch into the current CVS tree? -- Jürgen Keil jk(a)tools.de Tools GmbH +49 (228) 9858011 _______________________________________________ Mplayer-dev-eng mailing list Mplayer-dev-eng(a)lists.sourceforge.net http://lists.sourceforge.net/lists/listinfo/mplayer-dev-eng

2 1

Large file support patch
by Stephen Davies 03 Jul '01

03 Jul '01

Hi, Here's my patch to support large files in mplayer. I've added an --enable-largefiles switch to configure to enable it; by default it's turned off. The patch essentially makes sure that all variables and structures holding file offsets are type off_t, and uses the _LARGEFILE_SOURCE and _FILE_OFFSET_BITS defines to enable glibc's 64 bit offset support. With these, off_t is 64bits, without them it is 32bit. Its tested and works nicely for mpegs. I haven't tested ASF files though I think if made the right changes (anyone have a large asf?). I haven't changed the AVI seeking code - if I'm not mistaken they can't be larger than 2GB anyway. My system in Mandrake 8 originally; glibc 2.1.3. Let me know if this patch is of interest, any issues with it please let me know. Regards, Steve diff -U3 -rN main.b4changes/configure main/configure --- main.b4changes/configure Fri Jun 29 11:45:00 2001 +++ main/configure Tue Jul 3 13:17:55 2001 @@ -6,6 +6,9 @@ # # Changes in reversed order: # +# 2001/07/03 by Steve Davies +# - added --enable-largefiles +# # 2001/06/05 by Pontscho # - added alsa and esd detection # @@ -101,6 +104,7 @@ --prefix=DIR use this prefix for installing mplayer [/usr/local] --enable-debug[=1-3] compile debugging information into mplayer [disable] --enable-profile compile profiling information into mplayer [disable] + --enable-largefiles build with support for files >2^32 bytes long [disable] --enable-mmx build with mmx support [autodetect] --enable-mmx2 build with mmx2 support (PIII, Athlon) [autodetect] --enable-3dnow build with 3dnow! support [autodetect] @@ -764,6 +768,9 @@ --enable-debug) _debug='-g' ;; + --enable-largefiles) + _largefiles=yes + ;; --enable-debug=*) _debug=`echo $_echo_n '-g'$_echo_c; echo $ac_option | cut -d '=' -f 2` ;; @@ -1282,6 +1289,11 @@ # for MT applications: if [ "$system_name" = "FreeBSD" ]; then CFLAGS="$CFLAGS -D_THREAD_SAFE" +fi + +# 64 bit file offsets? +if [ "$_largefiles" = "yes" ]; then +CFLAGS="$CFLAGS -D_LARGEFILE_SOURCE -D_FILE_OFFSET_BITS=64" fi echo diff -U3 -rN main.b4changes/demuxer.c main/demuxer.c --- main.b4changes/demuxer.c Mon Jun 4 00:34:59 2001 +++ main/demuxer.c Tue Jul 3 13:27:33 2001 @@ -1,5 +1,8 @@ //=================== DEMUXER v2.5 ========================= +#include <sys/types.h> +#include <sys/stat.h> + #include <stdio.h> #include <stdlib.h> @@ -65,12 +68,12 @@ ds->first=ds->last=dp; } if(verbose>=2) - printf("DEMUX: Append packet to %s, len=%d pts=%5.3f pos=%d [packs: A=%d V=%d]\n", + printf("DEMUX: Append packet to %s, len=%d pts=%5.3f pos=%qd [packs: A=%d V=%d]\n", (ds==ds->demuxer->audio)?"d_audio":"d_video", - dp->len,dp->pts,dp->pos,ds->demuxer->audio->packs,ds->demuxer->video->packs); + dp->len,dp->pts,(long long)dp->pos,ds->demuxer->audio->packs,ds->demuxer->video->packs); } -void ds_read_packet(demux_stream_t *ds,stream_t *stream,int len,float pts,int pos,int flags){ +void ds_read_packet(demux_stream_t *ds,stream_t *stream,int len,float pts,off_t pos,int flags){ demux_packet_t* dp=new_demux_packet(len); stream_read(stream,dp->buffer,len); dp->pts=pts; //(float)pts/90000.0f; @@ -245,4 +248,3 @@ return len; } } - diff -U3 -rN main.b4changes/demuxer.h main/demuxer.h --- main.b4changes/demuxer.h Mon Jun 4 00:34:59 2001 +++ main/demuxer.h Tue Jul 3 13:27:41 2001 @@ -16,10 +16,11 @@ #define DEMUXER_TIME_BPS 3 +// Holds one packet/frame/whatever typedef struct demux_packet_st { int len; float pts; - int pos; // pozicio indexben (AVI) ill. fileban (MPG) + off_t pos; // pozicio indexben (AVI) ill. fileban (MPG) unsigned char* buffer; int flags; // keyframe, etc struct demux_packet_st* next; @@ -32,8 +33,8 @@ float pts; // current buffer's pts int pts_bytes; // number of bytes read after last pts stamp int eof; // end of demuxed stream? (true if all buffer empty) - int pos; // position in the input stream (file) - int dpos; // position in the demuxed stream + off_t pos; // position in the input stream (file) + off_t dpos; // position in the demuxed stream int pack_no; // serial number of packet int flags; // flags of current packet (keyframe etc) //--------------- @@ -54,12 +55,12 @@ typedef struct demuxer_st { stream_t *stream; int synced; // stream synced (used by mpeg) - int filepos; // input stream current pos. + off_t filepos; // input stream current pos. // int endpos; // input stream end pos. (return EOF fi filepos>endpos) int type; // mpeg system stream, mpeg elementary s., avi raw, avi indexed // int time_src;// time source (pts/file/bps) - unsigned int movi_start; - unsigned int movi_end; + off_t movi_start; + off_t movi_end; // demux_stream_t *audio; demux_stream_t *video; @@ -67,6 +68,7 @@ // index: // AVIINDEXENTRY* idx; +// FIXME: off_t??? void* idx; int idx_size; int idx_pos; @@ -102,12 +104,12 @@ demuxer_t* new_demuxer(stream_t *stream,int type,int a_id,int v_id,int s_id); void ds_add_packet(demux_stream_t *ds,demux_packet_t* dp); -void ds_read_packet(demux_stream_t *ds,stream_t *stream,int len,float pts,int pos,int flags); +void ds_read_packet(demux_stream_t *ds,stream_t *stream,int len,float pts,off_t pos,int flags); int demux_fill_buffer(demuxer_t *demux,demux_stream_t *ds); int ds_fill_buffer(demux_stream_t *ds); -inline static int ds_tell(demux_stream_t *ds){ +inline static off_t ds_tell(demux_stream_t *ds){ return (ds->dpos-ds->buffer_size)+ds->buffer_pos; } diff -U3 -rN main.b4changes/mplayer.c main/mplayer.c --- main.b4changes/mplayer.c Sat Jun 30 23:17:14 2001 +++ main/mplayer.c Tue Jul 3 13:40:37 2001 @@ -1,24 +1,25 @@ // AVI & MPEG Player v0.18 (C) 2000-2001. by A'rpi/ESP-team -#include <stdio.h> -#include <stdlib.h> -#include <string.h> - -#include <signal.h> +#include "version.h" +#include "config.h" #include <sys/ioctl.h> -#include <unistd.h> -#include <time.h> #include <sys/mman.h> - #include <sys/types.h> #include <sys/wait.h> #include <sys/time.h> #include <sys/stat.h> + +#include <signal.h> + +#include <time.h> + #include <fcntl.h> +#include <unistd.h> -#include "version.h" -#include "config.h" +#include <stdio.h> +#include <stdlib.h> +#include <string.h> #if defined(sun) #define DEFAULT_CDROM_DEVICE "/vol/dev/aliases/cdrom0" @@ -312,7 +313,7 @@ int osd_level=2; int divx_quality=0; char *seek_to_sec=NULL; -int seek_to_byte=0; +off_t seek_to_byte=0; int has_audio=1; //int has_video=1; int audio_format=0; // override @@ -684,7 +685,7 @@ stream->end_pos=ret2; } else { //============ Open plain FILE ============ - int len; + off_t len; if(!strcmp(filename,"-")){ // read from stdin printf("Reading from stdin...\n"); @@ -699,6 +700,10 @@ f=open(filename,O_RDONLY); if(f<0){ fprintf(stderr,"File not found: '%s'\n",filename);return 1; } len=lseek(f,0,SEEK_END); lseek(f,0,SEEK_SET); + if (len == -1) + perror("Error: lseek failed to obtain video file size"); + else + fprintf(stderr, "File size is %qd bytes\n", (long long)len); stream=new_stream(f,STREAMTYPE_FILE); stream->end_pos=len; #ifdef STREAMING @@ -2202,6 +2207,7 @@ switch(file_format){ + //FIXME: OFF_T - Didn't check AVI case yet (avi files can't be >2G anyway?) case DEMUXER_TYPE_AVI: { //================= seek in AVI ========================== int rel_seek_frames=rel_seek_secs*sh_video->fps; @@ -2324,12 +2330,14 @@ } break; + //FIXME: OFF_T - didn't test ASF case yet (don't have a large asf...) + //FIXME: reports good or bad to steve(a)daviesfam.org please case DEMUXER_TYPE_ASF: { //================= seek in ASF ========================== float p_rate=10; // packets / sec - int rel_seek_packs=rel_seek_secs*p_rate; - int rel_seek_bytes=rel_seek_packs*asf_packetsize; - int newpos; + off_t rel_seek_packs=rel_seek_secs*p_rate; // FIXME: int may be enough? + off_t rel_seek_bytes=rel_seek_packs*asf_packetsize; + off_t newpos; //printf("ASF: packs: %d duration: %d \n",(int)fileh.packets,*((int*)&fileh.duration)); // printf("ASF_seek: %d secs -> %d packs -> %d bytes \n", // rel_seek_secs,rel_seek_packs,rel_seek_bytes); @@ -2359,14 +2367,14 @@ case DEMUXER_TYPE_MPEG_ES: case DEMUXER_TYPE_MPEG_PS: { //================= seek in MPEG ========================== - int newpos; + off_t newpos; if(picture->bitrate==0x3FFFF) // unspecified? newpos=demuxer->filepos+2324*75*rel_seek_secs; // 174.3 kbyte/sec else newpos=demuxer->filepos+(picture->bitrate*1000/16)*rel_seek_secs; if(newpos<seek_to_byte) newpos=seek_to_byte; - newpos&=~(STREAM_BUFFER_SIZE-1); /* sector boundary */ + newpos&=~((long long)STREAM_BUFFER_SIZE-1); /* sector boundary */ stream_seek(demuxer->stream,newpos); // re-sync video: videobuf_code_len=0; // reset ES stream buffer diff -U3 -rN main.b4changes/stream.c main/stream.c --- main.b4changes/stream.c Thu Jun 21 01:06:40 2001 +++ main/stream.c Mon Jul 2 14:39:12 2001 @@ -1,10 +1,15 @@ -#include <stdio.h> -#include <stdlib.h> +#include "config.h" +#include <sys/types.h> +#include <sys/stat.h> #include <sys/ioctl.h> +#include <fcntl.h> #include <unistd.h> +#include <stdio.h> +#include <stdlib.h> + #include "stream.h" extern int verbose; // defined in mplayer.c @@ -40,14 +45,14 @@ return len; } -int stream_seek_long(stream_t *s,unsigned int pos){ -unsigned int newpos; +int stream_seek_long(stream_t *s,off_t pos){ +off_t newpos; -// if(verbose>=3) printf("seek to 0x%X\n",pos); +// if(verbose>=3) printf("seek to 0x%qX\n",(long long)pos); if(verbose>=3){ - printf("s->pos=%X newpos=%X new_bufpos=%X buflen=%X \n", - (unsigned int)s->pos,newpos,pos,s->buf_len); + printf("s->pos=%qX newpos=%qX new_bufpos=%qX buflen=%X \n", + (long long)s->pos,(long long)newpos,(long long)pos,s->buf_len); } s->buf_pos=s->buf_len=0; @@ -55,7 +60,7 @@ switch(s->type){ case STREAMTYPE_FILE: case STREAMTYPE_STREAM: - newpos=pos&(~(STREAM_BUFFER_SIZE-1));break; + newpos=pos&(~((long long)STREAM_BUFFER_SIZE-1));break; case STREAMTYPE_VCD: newpos=(pos/VCD_SECTOR_DATA)*VCD_SECTOR_DATA;break; } @@ -97,7 +102,7 @@ s->buf_pos=pos; // byte position in sector return 1; } - if(verbose) printf("stream_seek: WARNING! Can't seek to 0x%X !\n",pos+newpos); + if(verbose) printf("stream_seek: WARNING! Can't seek to 0x%qX !\n",(long long)(pos+newpos)); return 0; } diff -U3 -rN main.b4changes/stream.h main/stream.h --- main.b4changes/stream.h Mon Jun 4 18:51:17 2001 +++ main/stream.h Mon Jul 2 15:06:12 2001 @@ -20,16 +20,17 @@ typedef struct { int fd; - long pos; + off_t pos; int eof; int type; // 0=file 1=VCD unsigned int buf_pos,buf_len; - long start_pos,end_pos; + off_t start_pos,end_pos; unsigned char buffer[STREAM_BUFFER_SIZE>VCD_SECTOR_SIZE?STREAM_BUFFER_SIZE:VCD_SECTOR_SIZE]; } stream_t; int stream_fill_buffer(stream_t *s); -int stream_seek_long(stream_t *s,unsigned int pos); + +int stream_seek_long(stream_t *s,off_t pos); inline static int stream_read_char(stream_t *s){ return (s->buf_pos<s->buf_len)?s->buffer[s->buf_pos++]: @@ -91,16 +92,16 @@ return s->eof; } -inline static int stream_tell(stream_t *s){ +inline static off_t stream_tell(stream_t *s){ return s->pos+s->buf_pos-s->buf_len; } -inline static int stream_seek(stream_t *s,unsigned int pos){ +inline static int stream_seek(stream_t *s,off_t pos){ -// if(verbose>=3) printf("seek to 0x%X\n",pos); +// if(verbose>=3) printf("seek to 0x%qX\n",(long long)pos); if(pos<s->pos){ - int x=pos-(s->pos-s->buf_len); + off_t x=pos-(s->pos-s->buf_len); if(x>=0){ s->buf_pos=x; // putchar('*');fflush(stdout); _______________________________________________ Mplayer-dev-eng mailing list Mplayer-dev-eng(a)lists.sourceforge.net http://lists.sourceforge.net/lists/listinfo/mplayer-dev-eng

3 5

Re: RE: Re: [Mplayer-dev-eng] old binutils problem
by Juergen Keil 03 Jul '01

03 Jul '01

> > How about this change? Does it detect the correct version of 'as' used on > > your system? > > yes, it works! > please commit this patch to CVS! OK, it's in the CVS tree. _______________________________________________ Mplayer-dev-eng mailing list Mplayer-dev-eng(a)lists.sourceforge.net http://lists.sourceforge.net/lists/listinfo/mplayer-dev-eng

1 0

Re: [Mplayer-dev-eng] old binutils problem
by Juergen Keil 03 Jul '01

03 Jul '01

> is there any method to check 'as' version through gcc? How about this change? Does it detect the correct version of 'as' used on your system? Index: configure =================================================================== RCS file: /cvsroot/mplayer/main/configure,v retrieving revision 1.94 diff -u -B -r1.94 configure --- configure 29 Jun 2001 10:45:00 -0000 1.94 +++ configure 3 Jul 2001 11:58:58 -0000 @@ -203,7 +203,10 @@ # --- Check for C compiler: _cc=gcc -_as=as +_as=`gcc -print-prog-name=as` +if [ x"_$as" = x ]; then + _as=as +fi _x11=auto _x11libdir= @@ -986,7 +993,7 @@ # Checking assembler (_as) compatibility... as_version=`$_as --version 2>&1 | sed -n 's/^.*assembler $[0-9.]*$.*$/\1/p'` -echo $_echo_n "Checking assembler (as) ... $as_version, $_echo_c" +echo $_echo_n "Checking assembler ($_as) ... $as_version, $_echo_c" _pref_as_version='2.9.1' ### this test disabled, see _binutils test above! --A'rpi # cat > astest.S <<EOF Jürgen Keil jk(a)tools.de Tools GmbH +49 (228) 9858011 _______________________________________________ Mplayer-dev-eng mailing list Mplayer-dev-eng(a)lists.sourceforge.net http://lists.sourceforge.net/lists/listinfo/mplayer-dev-eng

2 1