GNOME Bugzilla – Bug 164234
[swfdec] segmentation fault playing a flash v6 that swf_play plays ok
Last modified: 2006-02-23 23:10:44 UTC
Steps to reproduce: trying to transcode a flash animation into an ogg theora file gst-launche will crash, here the comandline used for the transcoding: gst-launch-0.8 -vvv { filesrc location=file.swf ! decodebin name=decode } { decode. ! queue max-size-buffers =500 max-size-time=0 max-size-bytes=0 ! theoraenc ! oggmux name=mux ! filesink location=file.ogg decode. ! queue max-size-buffers=500 max-size-bytes=0 max-size-time=0 ! audioconvert ! rawvorbisenc ! mux. } this will result in: RUNNING pipeline ... /pipeline0/thread0/filesrc0.src: caps = application/octet-stream /pipeline0/thread0/decode/typefind.sink: caps = application/octet-stream /pipeline0/thread0/filesrc0.src: active = TRUE /pipeline0/thread0/decode/typefind.sink: active = TRUE /pipeline0/thread0/decode/typefind.src: active = TRUE /pipeline0/thread0/decode/swfdec0.audio_00: active = TRUE /pipeline0/thread0/decode/swfdec0.audio_00: caps = audio/x-raw-int, rate=(int)44100, channels=(int)2, endianness=(int)1234, width=(int)16, depth=(int)16, signed=(boolean)true /pipeline0/thread1/queue1.sink: caps = audio/x-raw-int, rate=(int)44100, channels=(int)2, endianness=(int)1234, width=(int)16, depth=(int)16, signed=(boolean)true /pipeline0/thread1/audioconvert0.sink: active = TRUE /pipeline0/thread1/audioconvert0.src: active = TRUE /pipeline0/thread1/queue1.src: caps = audio/x-raw-int, rate=(int)44100, channels=(int)2, endianness=(int)1234, width=(int)16, depth=(int)16, signed=(boolean)true /pipeline0/thread1/audioconvert0.sink: caps = audio/x-raw-int, rate=(int)44100, channels=(int)2, endianness=(int)1234, width=(int)16, depth=(int)16, signed=(boolean)true /pipeline0/thread1/vorbisenc0: last-message = "encoding at quality level 0.30" /pipeline0/thread1/audioconvert0.src: caps = audio/x-raw-float, rate=(int)44100, buffer-frames=(int)0, channels=(int)2, endianness=(int)1234, width=(int)32 /pipeline0/thread1/vorbisenc0.sink: caps = audio/x-raw-float, rate=(int)44100, buffer-frames=(int)0, channels=(int)2, endianness=(int)1234, width=(int)32 /pipeline0/thread1/vorbisenc0.sink: active = TRUE /pipeline0/thread1/vorbisenc0.src: active = TRUE /pipeline0/thread1/vorbisenc0.src: caps = audio/x-vorbis /pipeline0/thread1/mux.sink_419128573: caps = audio/x-vorbis /pipeline0/thread1/mux.src: active = TRUE /pipeline0/thread1/mux.sink_419128573: active = TRUE /pipeline0/thread1/mux.sink_540980785: active = TRUE /pipeline0/thread1/mux.src: caps = application/ogg /pipeline0/thread1/filesink0.sink: caps = application/ogg /pipeline0/thread1/filesink0.sink: active = TRUE /pipeline0/thread0/decode/swfdec0.video_00: active = TRUE /pipeline0/thread0/decode/swfdec0.sink: active = TRUE /pipeline0/thread0/decode/typefind.src: caps = application/x-shockwave-flash /pipeline0/thread0/decode/swfdec0.sink: caps = application/x-shockwave-flash Segmentation fault Stack trace:
+ Trace 54564
Other information:
looks like a crash in swfdec itself but be happy, swfdec author is also a gst dev (ds/david schleef). he's also liboil author. can you provide the file (attached or a link to it)
file anim1.swf anim1.swf: Macromedia Flash data (compressed), version 6 http://www.3dweb.no/galleri/stuestolbm/bilder/anim1.swf
/usr/bin/swf_play (from swfdec 0.3.2) can play it here (with a lot of "video_buffer == NULL" warnings on the console) but totem segfaults on it
i can play it on the command line : gst-launch-0.8 filesrc location=anim1.swf ! swfdec ! ffmpegcolorspace ! xvimagesink but i have lots of other warnings : ERROR (0x80508e8 - 307239:55:21.800541000) default(14085) gstswfdec.c(602):art_warn: caught art_warn repeated ad vitam eternam
I can't even play it on the command line: gst-launch-0.8 filesrc location=anim1.swf ! swfdec ! ffmpegcolorspace ! xvimagesink segfaults here bt seems clean enough:
+ Trace 54640
+ Trace 54641
p b->ptr[0] $1 = 0 '\0' p b->ptr[1] Cannot access memory at address 0x82b4000
This is a bug in swfdec, which has been fixed in CVS.
*** Bug 300379 has been marked as a duplicate of this bug. ***
*** Bug 332369 has been marked as a duplicate of this bug. ***