file command reports incorrect mime type for mkv files

[David D Miller]

Binary package hint: file

libmagic1-4.26-2ubuntu4

running "file -i" on an mkv file (matroska video) results in "application/octet-stream" for the mime type. While not technically inaccurate, the generally accepted type for this filetype at the moment is "video/x-matroska" which can be evidenced by looking in /etc/mime.types, where it's already defined this way.

[Austriaco (lanieves)]

There's one line missing in /usr/share/misc/magic which specifies the mime type. The addition can be made in /etc/magic for the time being, until the fix is adopted in libmagic1

$ file -bi /media/video/my.Video.mkv
video/x-matroska; charset=binary

[Adam Buchbinder (adam-buchbinder)]

This is fixed upstream as of version 5.04; it'll be in the development version of Ubuntu as soon as the next Debian sync.

[Austriaco (lanieves)]

As of Maverick (10.10), this has not been fixed and now the solution I provided above doesn't work either

[Austriaco (lanieves)]

Not all MKV files are wrongly detected after upgrading file and libmagic1 to version 5.04. Only files produced by the new mkvmerge from mkvtoolnix-4.0.0-0ubuntu3. Matroska files produced by the previous mkvmerge are detected alright:

Mime Type Muxing application
application/octet-stream libebml v1.0.0 + libmatroska v1.0.0
video/x-matroska libebml v0.7.7 + libmatroska v0.8.1

An inspection of the mis-identified files shows that they have a little more info at the beginning:

$ hexdump -C new-matroska-file.mkv | head -7
00000000 1a 45 df a3 a3 42 86 81 01 42 f7 81 01 42 f2 81 |.E...B...B...B..|
00000010 04 42 f3 81 08 42 82 88 6d 61 74 72 6f 73 6b 61 |.B...B..matroska|
00000020 42 87 81 02 42 85 81 02 18 53 80 67 01 00 00 00 |B...B....S.g....|
00000030 76 a4 26 1f 11 4d 9b 74 be 4d bb 8c 53 ab 84 15 |v.&..M.t.M..S...|
00000040 49 a9 66 53 ac 82 10 03 4d bb 8c 53 ab 84 16 54 |I.fS....M..S...T|
00000050 ae 6b 53 ac 82 10 b7 4d bb 8e 53 ab 84 1c 53 bb |.kS....M..S...S.|
00000060 6b 53 ac 84 76 a3 bb 1e 4d bb 8c 53 ab 84 10 43 |kS..v...M..S...C|

While the old ones are like this:

$hexdump -C old-matroska-file.mkv | head -7
00000000 1a 45 df a3 93 42 82 88 6d 61 74 72 6f 73 6b 61 |.E...B..matroska|
00000010 42 87 81 02 42 85 81 02 18 53 80 67 01 00 00 00 |B...B....S.g....|
00000020 3e 45 2b 2a 11 4d 9b 74 c0 4d bb 8c 53 ab 84 15 |>E+*.M.t.M..S...|
00000030 49 a9 66 53 ac 82 10 03 4d bb 8c 53 ab 84 16 54 |I.fS....M..S...T|
00000040 ae 6b 53 ac 82 10 b0 4d bb 8e 53 ab 84 11 4d 9b |.kS....M..S...M.|
00000050 74 53 ac 84 3e 44 36 01 4d bb 8e 53 ab 84 1c 53 |tS..>D6.M..S...S|
00000060 bb 6b 53 ac 84 3e 43 cd 9b ec 4f bb 00 00 00 00 |.kS..>C...O.....|

[Austriaco (lanieves)]

With this file in ~/.magic file detects correctly both versions of matroska

[Ken Sharp (kennybobs)]

Still present, not fixed in 5.09.

[Pauli Virtanen (pauli-virtanen)]

Bug is still present in 5.11-2ubuntu4

[Ken Sharp (kennybobs)]

Fixed in Trusty+