[FFmpeg-trac] #8676(undetermined:new): Concat demuxer prefixing the first input file path with "i"

[FFmpeg]

#8676: Concat demuxer prefixing the first input file path with "i"
-------------------------------------+-------------------------------------
             Reporter:  kuba-orlik   |                     Type:  defect
               Status:  new          |                 Priority:  normal
            Component:               |                  Version:  git-
  undetermined                       |  master
             Keywords:               |               Blocked By:
             Blocking:               |  Reproduced by developer:  0
Analyzed by developer:  0            |
-------------------------------------+-------------------------------------
 Summary of the bug:

 I was trying to create a simple slideshow with the concat demuxer. I've
 created an input.txt file. The first file in the input file was named
 "a.png". Fffmpeg threw an error saying "[concat @ 0x55e0599ec6c0]
 Impossible to open 'ia.png'". I've checked twice and there's no "ia.png"
 in the input.txt file. Renaming the a.png file to ia.png solves this
 issue.

 This bug seems to apply only to the first file in the input.txt.

 How to reproduce:
 {{{
 % cat input.txt
 ffconcat version 1.0
 file a.png
 duration 10
 file b.png
 duration 10
 file b.png

 % ffmpeg -f concat -i input.txt output.mp4
 ffmpeg version git-2020-05-16-1ead7ed Copyright (c) 2000-2020 the FFmpeg
 developers
   built with gcc 9.3.0 (Arch Linux 9.3.0-1)
   configuration:
   libavutil      56. 45.100 / 56. 45.100
   libavcodec     58. 84.100 / 58. 84.100
   libavformat    58. 43.100 / 58. 43.100
   libavdevice    58.  9.103 / 58.  9.103
   libavfilter     7. 81.100 /  7. 81.100
   libswscale      5.  6.101 /  5.  6.101
   libswresample   3.  6.100 /  3.  6.100
 [concat @ 0x55d48fb866c0] Impossible to open 'ia.png'
 input.txt: No such file or directory
 }}}

 {{{{
 % ffmpeg -v 9 -loglevel 99 -i input.txt
 ffmpeg version git-2020-05-16-1ead7ed Copyright (c) 2000-2020 the FFmpeg
 developers
   built with gcc 9.3.0 (Arch Linux 9.3.0-1)
   configuration:
   libavutil      56. 45.100 / 56. 45.100
   libavcodec     58. 84.100 / 58. 84.100
   libavformat    58. 43.100 / 58. 43.100
   libavdevice    58.  9.103 / 58.  9.103
   libavfilter     7. 81.100 /  7. 81.100
   libswscale      5.  6.101 /  5.  6.101
   libswresample   3.  6.100 /  3.  6.100
 Splitting the commandline.
 Reading option '-v' ... matched as option 'v' (set logging level) with
 argument '9'.
 Reading option '-loglevel' ... matched as option 'loglevel' (set logging
 level) with argument '99'.
 Reading option '-i' ... matched as input url with argument 'input.txt'.
 Finished splitting the commandline.
 Parsing a group of options: global .
 Applying option v (set logging level) with argument 9.
 Successfully parsed a group of options.
 Parsing a group of options: input url input.txt.
 Successfully parsed a group of options.
 Opening an input file: input.txt.
 [NULL @ 0x55c35d7745c0] Opening 'input.txt' for reading
 [file @ 0x55c35d775100] Setting default whitelist 'file,crypto,data'
 [AVIOContext @ 0x55c35d77d4c0] Statistics: 57 bytes read, 0 seeks
 input.txt: Invalid data found when processing input
 }}}}

 I ran git bisect and found that the first commit to have this behavior is
 648051f07cffd0d91c89dc6706e3d0d6a286de43 (avformat/url: check url root
 node when rel include double dot and trim double dot)

--
Ticket URL: <https://trac.ffmpeg.org/ticket/8676>
FFmpeg <https://ffmpeg.org>
FFmpeg issue tracker