crashes on non-UTF-8 file identifiers

[jwilk]

Issue reported by GStager at Bitbucket:

stager@stager-laptop:~/massocr$ ~/ocrodjvu-0.7.15/ocrodjvu --in-place -e tesseract -t words --html5 --clear-text -lrus+eng 9af500e27db4351d7391f463c0e3f017.djvu
Processing '9af500e27db4351d7391f463c0e3f017.djvu':
Intermediate files were left in the '/tmp/ocrodjvu.z1KiDA' directory.
Traceback (most recent call last):
  File "/home/stager1/ocrodjvu-0.7.15/ocrodjvu", line 7, in <module>
    _.main(sys.argv)
  File "/home/stager1/ocrodjvu-0.7.15/lib/cli/ocrodjvu.py", line 533, in main
    context.process(options.path, options.pages)
  File "/home/stager1/ocrodjvu-0.7.15/lib/cli/ocrodjvu.py", line 515, in process
    self._process(*args, **kwargs)
  File "/home/stager1/ocrodjvu-0.7.15/lib/cli/ocrodjvu.py", line 471, in _process
    file_id = page.file.id.encode(system_encoding)
  File "decode.pyx", line 840, in djvu.decode.File.id.__get__ (djvu/decode.c:7605)
  File "common.pxi", line 128, in djvu.decode.decode_utf8 (djvu/decode.c:2802)
UnicodeDecodeError: 'utf8' codec can't decode byte 0xf2 in position 0: invalid continuation byte

Original file

[jwilk]

Thanks for the bug report.

This is really a problem with the DjVu file in question. Its pages identifies are not in UTF-8, but in a some legacy 8-bit encoding instead.

I'll add a work-around in ocrodjvu for this, but you should fix the DjVu file. You can do that by converting it to indirect (with djvm), and then perhaps back to bundled.

[jwilk]

Fixed in 9a11f6a.

[jwilk]

Fixed in 0.7.16.