bsync only supports filenames encoded in the current locale

[ceremcem]

Problem can not be clearly understood:

ceremcem@cca-erik:/media/ceremcem/zfs_zencefil$ bsync ./-.-\ cca-home+/ /media/ceremcem/7c32ca7d-fb9e-4c0b-8e89-6de320908b8b/cca-sync/-.-\ cca-home+/
Loading filelists...
Old filelist not found. Starting with empty history.
Traceback (most recent call last):
  File "/home/ceremcem/bin/bsync", line 779, in <module>
    dir2 = load_dir(ssh2, dir2name, ignores)
  File "/home/ceremcem/bin/bsync", line 403, in load_dir
    record = read_file_record(gen)
  File "/home/ceremcem/bin/bsync", line 307, in read_file_record
    i,p,t,d,s,perms = next(gen),next(gen),next(gen),next(gen),next(gen),next(gen)
  File "/home/ceremcem/bin/bsync", line 294, in fileLineIter
    charsJustRead = inputFile.read(readSize)
  File "/usr/lib/python3.3/codecs.py", line 313, in decode
    (result, consumed) = self._buffer_decode(data, self.errors, final)
UnicodeDecodeError: 'utf-8' codec can't decode byte 0x8d in position 1277: invalid start byte

[dooblem]

Nice catch ceremcem,
This is clearly a bug. I've been able to reproduce it by creating non utf8 filenames on my filesystem.
I'll take a look at it.

I will try to avoid converting filenames to strings (using the current locale) and force the filenames to be kept as binary data.

The patch is not simple though.

[dooblem]

I fixed it and pushed it in a new branch.
This is the raw version: https://raw.githubusercontent.com/dooblem/bsync/exoticfiles/bsync
Can you try it and tell me if it's ok ?

[ceremcem]

This branch solved the problem. I have synchronized 500 GB of local files and there is no issue with the filenames. Thanks!

[dooblem]

Thanks. I just merged it to master.