Jonathan Vaughn | 1 Jun 2003 01:51

Re: Web site / Some suggestions

> On Sat, May 31, 2003 at 04:09:05PM -0500, Jonathan Vaughn wrote:
>> >> BTW in ffdshow/ffvfw CVS I have Visual Sudio projects for building
>> >> libavcodec using Visual C++ and Intel C++ Compliler. It did required
>> >> some modifications to libavcodec source code mainly removing GCC
>> >> specific extensions, too.
>> >
>> > I prefer not to favour non free compilers in ffmpeg, so I won't
>> integrate
>> > such
>> > patches anytime soon (at least for Visual C++).
>> >
>> > Fabrice.
>>
>> I think VC++ might have the ability to specify an external / custom
>> compiler, in which case it would be possible to use either cygwin or
>> mingw32 and specify in the project to use whichever compiler. This would
>
> Why?? Just throw out MSVC entirely. mingw can generate DLLs that are
> linkable to MSVC programs just fine, and if you don't insist on using
> msvc for your own stuff, you can just use mingw entirely.
>
> Rich

I don't care much one way or the other for the compiler, but the debugging
facilities are rather nice and it makes a good IDE to work in. I'm open to
suggestions for IDEs that could replace it both for debugging (stepping
through code / setting breakpoints / watching variables / etc) and for
ease of use as far as editing code and syntax highlighting and automatic
completion of members etc...

(Continue reading)

D Richard Felker III | 1 Jun 2003 02:18

Re: Web site / Some suggestions

On Sat, May 31, 2003 at 06:51:06PM -0500, Jonathan Vaughn wrote:
> > On Sat, May 31, 2003 at 04:09:05PM -0500, Jonathan Vaughn wrote:
> >> >> BTW in ffdshow/ffvfw CVS I have Visual Sudio projects for building
> >> >> libavcodec using Visual C++ and Intel C++ Compliler. It did required
> >> >> some modifications to libavcodec source code mainly removing GCC
> >> >> specific extensions, too.
> >> >
> >> > I prefer not to favour non free compilers in ffmpeg, so I won't
> >> integrate
> >> > such
> >> > patches anytime soon (at least for Visual C++).
> >> >
> >> > Fabrice.
> >>
> >> I think VC++ might have the ability to specify an external / custom
> >> compiler, in which case it would be possible to use either cygwin or
> >> mingw32 and specify in the project to use whichever compiler. This would
> >
> > Why?? Just throw out MSVC entirely. mingw can generate DLLs that are
> > linkable to MSVC programs just fine, and if you don't insist on using
> > msvc for your own stuff, you can just use mingw entirely.
> >
> > Rich
> 
> I don't care much one way or the other for the compiler, but the debugging
> facilities are rather nice and it makes a good IDE to work in. I'm open to
> suggestions for IDEs that could replace it both for debugging (stepping
> through code / setting breakpoints / watching variables / etc) and for
> ease of use as far as editing code and syntax highlighting and automatic
> completion of members etc...
(Continue reading)

Michael Niedermayer | 1 Jun 2003 02:56
Picon

CVS: ffmpeg/libavcodec 4xm.c,NONE,1.1 Makefile,1.94,1.95 allcodecs.c,1.25,1.26 avcodec.h,1.179,1.180

Update of /cvsroot/ffmpeg/ffmpeg/libavcodec
In directory sc8-pr-cvs1:/tmp/cvs-serv5419

Modified Files:
	Makefile allcodecs.c avcodec.h 
Added Files:
	4xm.c 
Log Message:
4xm codec

--- NEW FILE: 4xm.c ---
/*
 * 4XM codec
 * Copyright (c) 2003 Michael Niedermayer
 *
 * This library is free software; you can redistribute it and/or
 * modify it under the terms of the GNU Lesser General Public
 * License as published by the Free Software Foundation; either
 * version 2 of the License, or (at your option) any later version.
 *
 * This library is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
 * Lesser General Public License for more details.
 *
 * You should have received a copy of the GNU Lesser General Public
 * License along with this library; if not, write to the Free Software
 * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
 */

(Continue reading)

Mike Melanson | 1 Jun 2003 04:29
Favicon

New 4xm Decoder

Hi Michael,
	4xm video decoder. Very nifty. But (never quite sure how to phrase
this), is is supposed to work yet? I get a segfault straight out of the
gate as it tells me "error dc run != 0". Here is the backtrace:

Program received signal SIGSEGV, Segmentation fault.
decode_i_block
(f=0x82b59c0, block=0x82b6530) at common.h:501
501             return be2me_32( unaligned32(v)); //original
(gdb) bt
#0  decode_i_block (f=0x82b59c0, block=0x82b6530) at common.h:501
#1  0x0813ce01 in decode_i_frame (f=0x82b59c0, buf=0x82b72cc "\220\n",
    length=4828) at 4xm.c:453
#2  0x0813d406 in decode_frame (avctx=0x8257d48, data=0xbffff0e0,
    data_size=0xbffdf07c, buf=0x82b72cc "\220\n", buf_size=4840) at
4xm.c:668
#3  0x08076c68 in avcodec_decode_video (avctx=0x8257d48,
picture=0xbffff0e0,
    got_picture_ptr=0xbffdf07c, buf=0x82b72c0 "ifrm\022", buf_size=4840)
    at utils.c:362
#4  0x0805293d in av_encode (output_files=0x8202e60, nb_output_files=1,
    input_files=0x8202e00, nb_input_files=1, stream_maps=0x8202ec0,
    nb_stream_maps=0) at ffmpeg.c:1267
#5  0x08054c55 in main (argc=4, argv=0xbffff5a4) at ffmpeg.c:2886
#6  0x4005e566 in __libc_start_main (main=0x80549d4 <main>, argc=4,
    ubp_av=0xbffff5a4, init=0x805031c <_init>, fini=0x81b89e0 <_fini>,
    rtld_fini=0x4000a5e0 <_dl_fini>, stack_end=0xbffff59c)
    at ../sysdeps/generic/libc-start.c:129

Same with any of the MPlayer FTP samples.
(Continue reading)

gabucino | 1 Jun 2003 08:04
Picon

Re: New 4xm Decoder

Mike Melanson wrote:
> (never quite sure how to phrase this), is is supposed to work yet?
                                            ^^
Absolutely not this way. :>

--

-- 
Gabucino
MPlayer Core Team
Michael Niedermayer | 1 Jun 2003 12:13
Picon

CVS: ffmpeg/libavcodec 4xm.c,1.1,1.2

Update of /cvsroot/ffmpeg/ffmpeg/libavcodec
In directory sc8-pr-cvs1:/tmp/cvs-serv5347/libavcodec

Modified Files:
	4xm.c 
Log Message:
10l (array[-1] ...)

Index: 4xm.c
===================================================================
RCS file: /cvsroot/ffmpeg/ffmpeg/libavcodec/4xm.c,v
retrieving revision 1.1
retrieving revision 1.2
diff -u -d -r1.1 -r1.2
--- 4xm.c	1 Jun 2003 00:56:04 -0000	1.1
+++ 4xm.c	1 Jun 2003 10:13:29 -0000	1.2
 <at>  <at>  -492,22 +492,23  <at>  <at> 
 //        printf("%2X", ptr[j]);

     for(j=257; j<512; j++){
-        int smallest[2]= {-1,-1};
+        int min_freq[2]= {256*256, 256*256};
+        int smallest[2]= {0, 0};
         int i;
         for(i=0; i<j; i++){
             if(frequency[i] == 0) continue;
-            if(frequency[i] < frequency[ smallest[1] ]){
-                if(frequency[i] < frequency[ smallest[0] ]){
-                    smallest[1]= smallest[0];
-                    smallest[0]= i;
(Continue reading)

Michael Niedermayer | 1 Jun 2003 12:20
Picon
Picon

Re: New 4xm Decoder

Hi

On Sunday 01 June 2003 04:29, Mike Melanson wrote:
> Hi Michael,
> 	4xm video decoder. Very nifty. But (never quite sure how to phrase
> this), is is supposed to work yet? I get a segfault straight out of the
yes

> gate as it tells me "error dc run != 0". Here is the backtrace:
fixed, i hope

[...]
--

-- 
Michael
level[i]= get_vlc(); i+=get_vlc();		(violates patent EP0266049)
median(mv[y-1][x], mv[y][x-1], mv[y+1][x+1]);	(violates patent #5,905,535)
buf[i]= qp - buf[i-1];				(violates patent #?)
for more examples, see http://mplayerhq.hu/~michael/patent.html
stop it, see http://petition.eurolinux.org & http://petition.ffii.org/eubsa/en

-------------------------------------------------------
This SF.net email is sponsored by: eBay
Get office equipment for less on eBay!
http://adfarm.mediaplex.com/ad/ck/711-11697-6916-5
Michael Niedermayer | 1 Jun 2003 12:56
Picon

CVS: ffmpeg/libavcodec 4xm.c,1.2,1.3

Update of /cvsroot/ffmpeg/ffmpeg/libavcodec
In directory sc8-pr-cvs1:/tmp/cvs-serv16325

Modified Files:
	4xm.c 
Log Message:
fix 4xm yuv->rgb565 transform

Index: 4xm.c
===================================================================
RCS file: /cvsroot/ffmpeg/ffmpeg/libavcodec/4xm.c,v
retrieving revision 1.2
retrieving revision 1.3
diff -u -d -r1.2 -r1.3
--- 4xm.c	1 Jun 2003 10:13:29 -0000	1.2
+++ 4xm.c	1 Jun 2003 10:56:36 -0000	1.3
 <at>  <at>  -416,32 +416,38  <at>  <at> 
     int stride= f->current_picture.linesize[0]>>1;
     int i;
     uint16_t *dst = ((uint16_t*)f->current_picture.data[0]) + y * stride + x;
+    
+    for(i=0; i<4; i++){
+        block[i][0] += 0x80*8*8;
+        idct(block[i]);
+    }

-    for(i=0; i<6; i++) idct(block[i]);
+    if(!(f->avctx->flags&CODEC_FLAG_GRAY)){
+        for(i=4; i<6; i++) idct(block[i]);
+    }
(Continue reading)

Stefan Olsson | 1 Jun 2003 13:07

ffserver disappearing memory

Hi!

I have been running ffmpeg for a while now (live streaming for a local 
radio station) with great performance and sound quality (see 
ffserver.conf attached)!

However I do have a problem. Over time something (!) consumes all 
available memory - that is, it won't show up in free or other tools. 
System info: Linux Slackware 9.0 (kernel 2.4.20) running on a dual 
PII600, audiosource two SoundBlaster Live! cards with kernel emu10k1 driver.

The only idea that I have come up with so far is that it has to do with 
the way ffserver deals with the files in /tmp/ - but more specific than 
that I have no clue...

There must be others that have observed this - I have tested various 
versions of linux, soundcards etc to eliminate the possibility that it's 
a local config issue.

Any pointers on where to go from here?

Brgds

/Stefan

--

-- 
A real person has two reasons for doing anything... 
...a good reason and the real reason.

(Continue reading)

Mark Hills | 1 Jun 2003 13:57
Picon

Re: ffserver disappearing memory

On Sun, 1 Jun 2003, Stefan Olsson wrote:

> However I do have a problem. Over time something (!) consumes all 
> available memory - that is, it won't show up in free or other tools. 

I have had this problem with ffserver.

The first problem I traced it to was the use of the char *rc_eq in
AVCodecContext. This is a dynamically allocated string, but lots of things
don't deallocate it so essentially the memory was getting lost for each
connecting client. I have a patch which changes it to a static string and
avoids the problem (but obviously limits the length of the string). This
is fine on my audio streaming server.

However I don't think this is the overall problem. My ffserver frequently 
shows obscure crashes which don't make sense at all through a debugger. 
This can typically happen every couple of days which is incredibly 
frustrating.

Running ffserver through valgrind shows a hint of the problem. Although 
this isn't practical on my actual running server 'cos it's so slow. 
Valgrind shows errors occurring around the dyn_buf functions such as 
trying to deallocate a pointer which points to the middle of an allocated 
block, and certain occasions when buffers aren't deallocated. I've tried 
to trace the problem but I keep getting tied up in spaghetti code. Can 
anyone who knows a little more about the ffserver buffers and things offer 
a possible explanation?

Thanks, Mark
--

-- 
(Continue reading)


Gmane