Enca should probably exit 1 each time it prints 'unknown'

Question

Enca should probably exit 1 each time it prints 'unknown'

JelF opened this issue 8 years ago · comments

I have some string, which encoding i want to detect (attached)

When i'm calling enca -m, result is "unknown\n", but exit code is zero (same as enca's, where response is

Macintosh Cyrillic
  LF line terminators

)

Michal Čihař · Answer 1 · Thu Jul 28 2016 22:51:41 GMT+0800 (China Standard Time)

I somehow can not reproduce this:

$ enca -m /tmp/123.txt
unknown
$ echo $?
1

What enca version are you using?

Alexandr Smirnov · Answer 2 · Sat Jul 30 2016 00:38:03 GMT+0800 (China Standard Time)

@nijel

$ enca -v
enca 1.14

Features: +librecode-interface +iconv-interface +external-converter +language-detection +locale-alias +target-charset-auto +ENCAOPT

As far as i see, it is not latest. I'll build it from master and try to reproduce bug sometime later

Alexandr Smirnov · Answer 3 · Sat Jul 30 2016 02:15:42 GMT+0800 (China Standard Time)

➜  bin git:(master) ✗ ./enca --version
enca 1.19-dev

Features: +librecode-interface +iconv-interface +external-converter +language-detection +locale-alias +target-charset-auto +ENCAOPT 

Copyright (C) 2000-2005 David Necas (Yeti) (<yeti@physics.muni.cz>),
              2005 Zuxy Meng (<zuxy.meng@gmail.com>).

Enca is free software; it can be copied and/or modified under the terms of
version 2 of GNU General Public License, run `enca --license' to see the full
license text.  There is NO WARRANTY; not even for MERCHANTABILITY or FITNESS
FOR A PARTICULAR PURPOSE.
➜  bin git:(master) ✗ ./enca -m /tmp/123.txt 
unknown
➜  bin git:(master) ✗ echo $?
0

Michal Čihař · Answer 4 · Mon Aug 01 2016 18:03:44 GMT+0800 (China Standard Time)

Hmm, that's strange. On what platform is that? I've fixed some minor issues in the return codes handling, but I don't think these will fix it for you.

Alexandr Smirnov · Answer 5 · Mon Aug 01 2016 20:02:32 GMT+0800 (China Standard Time)

I am currently running gentoo linux (4.3.6-ck). Pulled current master, problem persists. I'll try make PR at weekend

Alexandr Smirnov · Answer 6 · Sat Aug 06 2016 15:09:10 GMT+0800 (China Standard Time)

sorry, no idea how to fix :(

Michal Čihař · Answer 7 · Mon Aug 08 2016 18:41:27 GMT+0800 (China Standard Time)

Can you use gdb to see process_file return value?

Alexandr Smirnov · Answer 8 · Mon Aug 08 2016 23:33:15 GMT+0800 (China Standard Time)

@nijel no, i am completly not familar with gdb and other c-related stuff :(

Frank Dana · Answer 9 · Wed May 02 2018 10:14:53 GMT+0800 (China Standard Time)

I've reproduced this. The issue appears to be, as @JelF hinted at above, that enca -m exits with the same exit value enca would, even when enca -m is outputting "unknown".

It's easiest to illustrate with an example. I'm in an en_US locale, so normally enca doesn't support my local language settings:

$ enca -m ~/Desktop/123.txt
enca: Cannot determine (or understand) your language preferences. [...etc...]
$ echo $?
1

$ echo `# When enca can't recognize the encoding at all`

$ enca -L none ~/Desktop/123.txt
Unrecognized encoding
$ echo $?
1

$ enca -L none -m ~/Desktop/123.txt
unknown
$ echo $?
1

$ echo `# When enca CAN identify the file encoding, but not the corresponding MIME type`

$ enca -L russian ~/Desktop/123.txt
Macintosh Cyrillic
  LF line terminators
$ echo $?
0

$ enca -L russian -m ~/Desktop/123.txt
unknown
$ echo $?
0