bugGnash - The GNU Flash player - Bugs: bug #18236, Assertion failure in...

 
 

bug #18236: Assertion failure in sound_handler_sdl.cpp:931:

Submitted by:  Marti Raudsepp <intgr>
Submitted on:  Wed 08 Nov 2006 11:00:02 PM UTC  
 
Category: sound-sdlSeverity: 5 - Blocker
Release: NoneStatus: Fixed
Privacy: PublicAssigned to: Tomas Groth <tgc>
Open/Closed: Closed

Add a New Comment(Rich Markup)
   

You are not logged in

Please log in, so followups can be emailed to you.

 

(Jump to the original submission Jump to the original submission)

Tue 14 Nov 2006 08:05:30 PM UTC, comment #17:

That warning is ok. I'll commit the patch and close this bug for now.

Tomas Groth <tgc>In charge of this item.
Tue 14 Nov 2006 06:49:00 PM UTC, comment #16:

Yeah, forgot to mention, the sound corruption (bug #18244) still persists.

By the way, now that I've configured with --enable-debug, I can see this warning that might have something to do with it:

20:30:22: WARNING: Recoverable error while decoding MP3-stream, MAD error: bad main_data_begin pointer

Marti Raudsepp <intgr>
Tue 14 Nov 2006 06:39:45 PM UTC, comment #15:

Great! :)
I assume the sound corruption is still present?

Tomas Groth <tgc>In charge of this item.
Tue 14 Nov 2006 06:30:49 PM UTC, comment #14:

Indeed, the bug does not seem to occur any more with the latest patch.

Marti Raudsepp <intgr>
Tue 14 Nov 2006 06:03:56 PM UTC, comment #13:

New test patch attached... please test.

Tomas Groth <tgc>In charge of this item.
Mon 13 Nov 2006 08:57:42 AM UTC, comment #12:

Hmm, i guess this might be a blocker after all... I'll look into it tonight or tomorrow.

Tomas Groth <tgc>In charge of this item.
Fri 10 Nov 2006 06:09:45 PM UTC, comment #11:

> By the line number I can tell that you're using cvs-head, and not 0.7.2, so this is not a release blocker..


Sorry, I had indeed accidentally checked out HEAD this time (when originally reporting the bug, I did use the release_0_7_2 branch). I re-tested with 0.7.2 and similar assertion failures are still being fired.

With the patch (usually near the end of the animation):
lt-gnash: sound_handler_sdl.cpp:744: void sdl_audio_callback(void, Uint8, int): Assertion `sound->data_size > sound->position' failed.

And without the patch (when clicking 'PLAY'):
lt-gnash: sound_handler_sdl.cpp:881: void sdl_audio_callback(void, Uint8, int): Assertion `sound->raw_position + buffer_length - index < sound->raw_data_size' failed.

% ./gui/gnash --version
Gnash 0.7.2
Copyright (C) 2006 Free Software Foundation, Inc.
Gnash comes with NO WARRANTY, to the extent permitted by law.
You may redistribute copies of Gnash under the terms of the GNU General
Public License. For more information, see the file named COPYING.
Build options 0.7.2
Renderer: opengl GUI: gtk Sound handler: sdl Decoder: mad

Marti Raudsepp <intgr>
Fri 10 Nov 2006 04:08:29 PM UTC, comment #10:

> Sorry, I was wrong; the assertion failure does not appear to be completely
> fixed - it just happens much later in the animation now - and not always in
> the same place. (I also reported bug #18244 with the sound corruption issue)
>
> lt-gnash: sound_handler_sdl.cpp:794: void sdl_audio_callback(void, Uint8,
> int): Assertion `sound->data_size > sound->position' failed.
>
> Can you reproduce the bug? Is there anything I can do to help in debugging
> it?
>

By the line number I can tell that you're using cvs-head, and not 0.7.2, so this is not a release blocker...
The bug is related to what i explained in comment #6, and i doubt that the patch should go in 0.7.2.

Tomas Groth <tgc>In charge of this item.
Thu 09 Nov 2006 10:08:37 PM UTC, comment #9:

Sorry, I was wrong; the assertion failure does not appear to be completely fixed - it just happens much later in the animation now - and not always in the same place. (I also reported bug #18244 with the sound corruption issue)

lt-gnash: sound_handler_sdl.cpp:794: void sdl_audio_callback(void, Uint8, int): Assertion `sound->data_size > sound->position' failed.

Can you reproduce the bug? Is there anything I can do to help in debugging it?

Marti Raudsepp <intgr>
Thu 09 Nov 2006 08:39:03 PM UTC, comment #8:

The patch appears to fix the bug, but now the audio stream is just corrupt (just like it was with the latest win32 build). It seems that parts of the audio stream are consistently skipped.

Should I file another bug on this?

Marti Raudsepp <intgr>
Thu 09 Nov 2006 03:24:45 PM UTC, comment #7:

comment #1 says it is the release branch, it seems.

Sandro Santilli <strk>
Project Member
Thu 09 Nov 2006 03:21:15 PM UTC, comment #6:

The diff between the branch and head is the buffersize. The video stuff seems to need a bigger buffer (which seems a bit odd - i'll look into that at one point), so it has been made bigger. The patch makes sure enough data to fill the buffer is decoded, but it shouldn't be an issue at all in the release branch - though this bugreport might prove me wrong - unless it is the head branch.

Tomas Groth <tgc>In charge of this item.
Thu 09 Nov 2006 03:14:00 PM UTC, comment #5:

I get the same assertion with SimGirl.swf, but can NOT reproduce in the release branch, Thomas could you check the difference between the two branches ?

BTW, I checked your patch, and I confirm it fixes it for me
in head. I'm now trying to apply to release branch too....

Sandro Santilli <strk>
Project Member
Thu 09 Nov 2006 01:17:33 AM UTC, comment #4:

Just "for fun" I made a patch and attached it, please test.

Tomas Groth <tgc>In charge of this item.
Thu 09 Nov 2006 01:00:17 AM UTC, comment #3:

I'm playing it with sdl/mad just now, withouht problems (0.7.2)
except some warnings:

01:58:37: WARNING: Recoverable error while decoding MP3-stream, MAD error: bad main_data_begin pointer

what architecture are you on ?

Sandro Santilli <strk>
Project Member
Wed 08 Nov 2006 11:57:38 PM UTC, comment #2:

Also tested and getting the same result with --with-mp3-decoder=ffmpeg (version output says Decoder: ffmpeg)

Marti Raudsepp <intgr>
Wed 08 Nov 2006 11:27:14 PM UTC, comment #1:

% gnash --version
Gnash 0.7.2
[...]
Build options 0.7.2
Renderer: opengl GUI: gtk Sound handler: sdl Decoder: mad

Marti Raudsepp <intgr>
Wed 08 Nov 2006 11:00:02 PM UTC, original submission:

I've been experiencing the following assertion failure with the latest CVS HEAD as of Thu Nov 9 00:27:40 2006 (EET):

gnash: sound_handler_sdl.cpp:931: void sdl_audio_callback(void, Uint8, int): Assertion `sound->raw_position + buffer_length - index < sound->raw_data_size' failed.

I'm running Gentoo on the AMD64 arch, with GCC 4.1.1, alsa-lib 1.0.13, sdl-sound 1.0.1-r2, sdl-mixer 1.2.7, libsdl 1.2.11

The offending movie can be found at http://www.weebls-stuff.com/toons/On+The+Moon+ep.1/ (direct download: http://www.weebls-stuff.com/flash/798moon1.swf)

Click 'PLAY' to reproduce the bug.

Full stack trace:

Program received signal SIGABRT, Aborted.
---Type <return> to continue, or q <return> to quit---
[Switching to Thread 1090525504 (LWP 3441)]
0x00002b62d56ef7b5 in raise () from /lib/libc.so.6
(gdb) bt
#0 0x00002b62d56ef7b5 in raise () from /lib/libc.so.6
#1 0x00002b62d56f0a4e in abort () from /lib/libc.so.6
#2 0x00002b62d56e9156 in __assert_fail () from /lib/libc.so.6
#3 0x00002b62d0ba5e24 in SDL_sound_handler::create_sound ()
from /usr/lib/libgnashbackend.so.0
#4 0x00002b62d13ce0ea in SDL_AudioInit () from /usr/lib/libSDL-1.2.so.0
#5 0x00002b62d13d5cb7 in SDL_GetThreadID () from /usr/lib/libSDL-1.2.so.0
#6 0x00002b62d1414b79 in SDL_ThreadID () from /usr/lib/libSDL-1.2.so.0
#7 0x00002b62d26bf135 in start_thread () from /lib/libpthread.so.0
#8 0x00002b62d577d3dd in clone () from /lib/libc.so.6
#9 0x0000000000000000 in ?? ()

Marti Raudsepp <intgr>

 

(Note: upload size limit is set to 16384 kB, after insertion of the required escape characters.)

Attach File(s):
   
   
Comment:
   

Attached Files
file #11265:  fillbuf2.patch added by tgc (12KiB - text/x-patch - a new patch to fix the issue)
file #11204:  fillbuf.patch added by tgc (12KiB - text/x-patch - Just a little test patch)

 

Depends on the following items: None found

Items that depend on this one: None found

 

CC list is empty

 

Do you think this task is very important?
If so, you can click here to add your encouragement to it.
This task has 0 encouragements so far.

Only logged-in users can vote.

 

Please enter the title of George Orwell's famous dystopian book (it's a date):

 

 

Follow 11 latest changes.

Date Changed By Updated Field Previous Value => Replaced By
Tue 14 Nov 2006 08:05:30 PM UTCtgcStatusIn Progress=>Fixed
  Open/ClosedOpen=>Closed
Tue 14 Nov 2006 06:03:56 PM UTCtgcAttached File-=>Added fillbuf2.patch, #11265
Mon 13 Nov 2006 08:57:42 AM UTCtgcSeverity3 - Normal=>5 - Blocker
Fri 10 Nov 2006 04:08:29 PM UTCtgcSeverity5 - Blocker=>3 - Normal
Thu 09 Nov 2006 03:14:00 PM UTCstrkSummaryAssertion failure in sound_handler_sdl.cpp:931:=>Assertion failure in sound_handler_sdl.cpp:931:
Thu 09 Nov 2006 01:17:33 AM UTCtgcStatusNone=>In Progress
  Attached File-=>Added fillbuf.patch, #11204
Wed 08 Nov 2006 11:25:54 PM UTCstrkSeverity3 - Normal=>5 - Blocker
  Assigned toNone=>tgc
Wed 08 Nov 2006 11:00:02 PM UTCintgrCarbon-Copy-=>Added intgr

Back to the top


Powered by Savane 3.1-cleanup1