Negative time format contains too many minuses: -577014:-32:-22.-77

[q345t7n3o67ym23oy8939t]

Summary of the bug: time format for negative time contains too many minuses: "-577014:-32:-22.-77", but should be "-577014:32:22.77".
ffmpeg version: N-60087-g94a5241 from ​http://ffmpeg.zeranoe.com/builds/
How to reproduce:

% "ffmpeg.exe" -loglevel "debug" -i "http://93.95.99.231:8200/" -f "segment" -codec "copy" -map "0" -segment_format "mpegts" -segment_list "target file.csv" -segment_list_size "0" -segment_list_type "csv" -segment_time "906" -segment_start_number "0" -reset_timestamps "1" "target file %d.ts"

… and wait negative time.
My ffmpeg log contains:

frame=89905 fps= 25 q=-1.0 size=N/A time=00:59:57.34 bitrate=N/A    
frame=89921 fps= 25 q=-1.0 size=N/A time=00:59:57.97 bitrate=N/A    
frame=89936 fps= 25 q=-1.0 size=N/A time=00:59:58.57 bitrate=N/A    
frame=89952 fps= 25 q=-1.0 size=N/A time=-577014:-32:-22.-77 bitrate=N/A    
frame=89952 fps= 25 q=-1.0 Lsize=N/A time=00:59:59.15 bitrate=N/A    

[Carl Eugen Hoyos]

Please provide your command line together with the complete, uncut console output to make this a valid ticket. If the output is really enormous, please provide at least the first 50 and the last 50 lines.

[q345t7n3o67ym23oy8939t]

Archive of logs

[q345t7n3o67ym23oy8939t]

Attacment ffmpeg.log.7z​ contains another log with "frame= 1238 fps= 16 q=-1.0 size=N/A time=-577014:-32:-22.-77 bitrate=N/A " on 3458 line. Again with time -577014:-32:-22.-77.

[Michael Niedermayer]

can this only be reproduced with network input ?
only with segment output ?

[q345t7n3o67ym23oy8939t]

Replying to michael:

can this only be reproduced with network input ?
only with segment output ?

I can not reproduce without segment output.

[Carl Eugen Hoyos]

Two patches were sent; since we all cannot reproduce, can you test them?
​http://thread.gmane.org/gmane.comp.video.ffmpeg.devel/175346
​http://thread.gmane.org/gmane.comp.video.ffmpeg.devel/175415

[q345t7n3o67ym23oy8939t]

Replying to cehoyos:

Two patches were sent; since we all cannot reproduce, can you test them?
​http://thread.gmane.org/gmane.comp.video.ffmpeg.devel/175346
​http://thread.gmane.org/gmane.comp.video.ffmpeg.devel/175415

I test on ffmpeg version N-61286-gdbc3e11 built on Mar 11 2014 22:06:56 with gcc 4.8.2 (GCC) without this patches, and reprodused error. But it took 10 hours. Error began to appear less often.
Where can I get a compiled version with the patch? I find it hard to compile it yourself.

[Carl Eugen Hoyos]

Should be fixed in 4d17179b - please test current FFmpeg git head!

[q345t7n3o67ym23oy8939t]

Replying to cehoyos:

Should be fixed in 4d17179b - please test current FFmpeg git head!

​http://ffmpeg.zeranoe.com/builds/ on until the old version. Wait for tomorrow, please.

[q345t7n3o67ym23oy8939t]

Replying to cehoyos:

Should be fixed in 4d17179b - please test current FFmpeg git head!

I can not compile ffmpeg, and ​http://ffmpeg.zeranoe.com/builds/ does not contains this patch. I can not test the patch.

[Carl Eugen Hoyos]

Replying to q345t7n3o67ym23oy8939t:

Replying to cehoyos:

Should be fixed in 4d17179b - please test current FFmpeg git head!

I can not compile ffmpeg

Why not? It is supposed to work out of the box both on common and less common operating systems.

and ​http://ffmpeg.zeranoe.com/builds/ does not contains this patch.

This sounds impossible.

I can not test the patch.

Just test a current Zeranoe version.

[q345t7n3o67ym23oy8939t]

Just test a current Zeranoe version.

Negative time no longer appears. The test lasted 24 hours.

[Carl Eugen Hoyos]

I can still reproduce this problem:

$ ffmpeg -i fate-suite/utvideo/utvideo_rgba_left.avi -f null -
ffmpeg version N-63346-g2491579 Copyright (c) 2000-2014 the FFmpeg developers
  built on May 20 2014 20:00:07 with gcc 4.7 (SUSE Linux)
  configuration: --enable-gpl
  libavutil      52. 86.100 / 52. 86.100
  libavcodec     55. 63.100 / 55. 63.100
  libavformat    55. 39.100 / 55. 39.100
  libavdevice    55. 13.101 / 55. 13.101
  libavfilter     4.  5.100 /  4.  5.100
  libswscale      2.  6.100 /  2.  6.100
  libswresample   0. 19.100 /  0. 19.100
  libpostproc    52.  3.100 / 52.  3.100
Input #0, avi, from 'fate-suite/utvideo/utvideo_rgba_left.avi':
  Duration: 00:00:00.17, start: 0.000000, bitrate: 41978 kb/s
    Stream #0:0: Video: utvideo (ULRA / 0x41524C55), rgba, 640x480, 29.97 fps, 29.97 tbr, 29.97 tbn, 29.97 tbc
Output #0, null, to 'pipe:':
  Metadata:
    encoder         : Lavf55.39.100
    Stream #0:0: Video: rawvideo (RGBA / 0x41424752), rgba, 640x480, q=2-31, 200 kb/s, 29.97 fps, 90k tbn, 29.97 tbc
    Metadata:
      encoder         : Lavc55.63.100 rawvideo
Stream mapping:
  Stream #0:0 -> #0:0 (utvideo -> rawvideo)
Press [q] to stop, [?] for help
[null @ 0x28952e0] Encoder did not produce proper pts, making some up.
frame=    5 fps=0.0 q=0.0 Lsize=N/A time=-577014:-32:-22.-77 bitrate=N/A
video:0kB audio:0kB subtitle:0kB other streams:0kB global headers:0kB muxing overhead: 3843071590396854272.000000%

[Michael Niedermayer]

i get a
frame= 5 fps=0.0 q=0.0 Lsize=N/A time=00:00:00.16 bitrate=N/A
so i guess this must have been fixed

[Carl Eugen Hoyos]

I will open a new ticket if I find another example.