ffplay segmentation fault on oldabi branch when using --enable-small

[exobuzz]

Compiling oldabi branch HEAD as of 05/08/2011 (last commit 25c2f13d00fcfdb81c33a459795c33d29f8690e8) with --enable-small crashes when playing a video.

running gcc version 4.5.2 (Ubuntu/Linaro 4.5.2-8ubuntu4) on 64 bit,.

building with "./configure --enable-small"

I git bisected it to this commit:

44f669e7bc4f7f064e3f81d3596637a0e043b501 is the first bad commit
commit 44f669e7bc4f7f064e3f81d3596637a0e043b501
Author: Stefano Sabatini <stefano.sabatini-lala@poste.it>
Date:   Sat Jun 11 18:43:11 2011 +0200

    lavfi: add vsink_buffer, and use it in ff* tools
    
    Also add the public interface libavfilter/vsink_buffer.h.

:100644 100644 7745fa0938958ba7a2334be1a2ef9c33dc224d8c 80ed401ab9e99210665c40710d340ef4ca880c02 M	Changelog
:100644 100644 3b9cfaa3d97318d9bd34a8956f0c09998aca41a6 c3c5c0efa9e30452ae96dde9742da54809394171 M	cmdutils.c
:100644 100644 c0c3ce07d95ebf7310ff726a9d56fb11ab165248 e001ab9201487f90281655dc273e49b22e4bde66 M	cmdutils.h
:040000 040000 131e7beeba6126fbbf469ed87486ec0ca38efb61 5bcd802168bf5ed0adb736ba8d944234dcdd4473 M	doc
:100644 100644 2921a06ddfa5333ee08ae523e5f8cc697c932c90 0b1b1bee8db5246525ca629873cc14c88886f110 M	ffmpeg.c
:100644 100644 31a683216fc416455090f8472a52a7a113d122dd 548fda627a08c6260a3d5d691ab89072a84b9a31 M	ffplay.c
:040000 040000 4cedfd0c9064e7e6bb84c9917af1a7af6dd88b2b 2b1aa7a2d415b5109fd36012bf714768ccd17f19 M	libavfilter

backtrace didn't seem useful (Please feel free to give me some guidance if you can't reproduce).

(gdb) run
Starting program: /home/jools/Repos/Public/ffmpeg/ffplay afilm.mkv
[Thread debugging using libthread_db enabled]
ffplay version N-30958-g25c2f13, Copyright (c) 2003-2011 the FFmpeg developers
  built on Aug  5 2011 01:21:15 with gcc 4.5.2
  configuration: --enable-small
  libavutil    50. 43. 0 / 50. 43. 0
  libavcodec   52.122. 0 / 52.122. 0
  libavformat  52.110. 0 / 52.110. 0
  libavdevice  52.  5. 0 / 52.  5. 0
  libavfilter   1. 80. 0 /  1. 80. 0
  libswscale    0. 14. 1 /  0. 14. 1
[New Thread 0x7ffff26cc700 (LWP 16816)]
[New Thread 0x7fffedeca700 (LWP 16817)]
[Thread 0x7fffedeca700 (LWP 16817) exited]
[New Thread 0x7fffedeca700 (LWP 16818)]
[matroska,webm @ 0x1039740] max_analyze_duration 5000000 reached at 5002000
[matroska,webm @ 0x1039740] Estimating duration from bitrate, this may be inaccurate
Input #0, matroska,webm, from 'afilm.mkv':
  Metadata:
    title           : Dune (1984)
  Duration: 02:16:18.02, start: 0.000000, bitrate: 1536 kb/s
    Stream #0.0(eng): Video: h264, yuv420p, 1280x544, PAR 1:1 DAR 40:17, 24 fps, 24 tbr, 1k tbn, 48 tbc (default)
    Stream #0.1(eng): Audio: dca, 48000 Hz, 5.1, s16, 1536 kb/s (default)
    Stream #0.2(eng): Subtitle: ass (default)
    Stream #0.3(cze): Subtitle: ass
    Stream #0.4(fre): Subtitle: ass
    Stream #0.5(ger): Subtitle: ass
    Stream #0.6(hun): Subtitle: ass
    Stream #0.7(spa): Subtitle: ass
    Stream #0.8: Attachment: [0][0][0][0] / 0x0000
    Metadata:
      filename        : CronosPro-Bold.ttf
    Stream #0.9: Attachment: [0][0][0][0] / 0x0000
    Metadata:
      filename        : CronosPro-Semibold.ttf
    Stream #0.10: Attachment: [0][0][0][0] / 0x0000
    Metadata:
      filename        : CronosPro-SemiboldIt.ttf
[New Thread 0x7ffff1dfd700 (LWP 16819)]
[New Thread 0x7ffff15fc700 (LWP 16820)]
[New Thread 0x7ffff0dfb700 (LWP 16821)]

Program received signal SIGSEGV, Segmentation fault.
[Switching to Thread 0x7ffff15fc700 (LWP 16820)]
0x00007ffff6af72a4 in ?? () from /lib/x86_64-linux-gnu/libc.so.6
(gdb) bt
#0  0x00007ffff6af72a4 in ?? () from /lib/x86_64-linux-gnu/libc.so.6
#1  0x000000000043965d in ?? ()
#2  0x0000000000431c7a in ?? ()
#3  0x00007ffff7640875 in ?? () from /usr/lib/libSDL-1.2.so.0
#4  0x00007ffff7683049 in ?? () from /usr/lib/libSDL-1.2.so.0
#5  0x00007ffff6d6cd8c in start_thread () from /lib/x86_64-linux-gnu/libpthread.so.0
#6  0x00007ffff6ab804d in clone () from /lib/x86_64-linux-gnu/libc.so.6
#7  0x0000000000000000 in ?? ()

[exobuzz]

Could be a compiler bug of course, but if so, I would need some guidance to report it "upstream". thanks for listening.

[Carl Eugen Hoyos]

Could you test again?
Many changes were pushed today.

[exobuzz]

will do. I wish I had gone to sleep last night instead of branching oldabi and merging in a bunch of stuff that I needed for the project im working on, as I saw today, I could have saved myself a lot of time. that will teach me not to be patient ;-)

[exobuzz]

Seems ok now with the latest updates. Thanks.

[Carl Eugen Hoyos]

Thank you for testing!