FWIW: NZB with .SFV gives problem, without .SFV all OK

[sander]

FWIW:

I have no idea if this good behaviour and as expected, or something to look at:

I created a NZB on binsearch including a .SFV file. Trying to download that, SAB says " PostProcessing was aborted (see logfile) ". See logfile snippet below, with error message "UnicodeDecodeError: 'utf8' codec can't decode byte 0xba in position 88: invalid start byte"
I created a new NZB on binsearch, now without the .SFV file, and the download went OK.

So: is this is problem in the post/upload itself, or in SABnzbd?

Both NZBs sent to bugs @ ...

Code: Select all

2011-10-15 10:02:20,919::INFO::[__init__:838] Saving data for queue9.sab in /home/ubuntu/.sabnzbd/admin/queue9.sab
2011-10-15 10:02:20,920::INFO::[postproc:83] Saving postproc queue
2011-10-15 10:02:20,920::INFO::[__init__:838] Saving data for postproc1.sab in /home/ubuntu/.sabnzbd/admin/postproc1.sab
2011-10-15 10:02:20,962::INFO::[postproc:255] Starting PostProcessing on Paris - Hi-Fi => Repair:True, Unpack:True, Delete:True, Script:None, Cat:*
2011-10-15 10:02:20,962::INFO::[postproc:498] Par2 check starting on Paris - Hi-Fi
2011-10-15 10:02:20,963::ERROR::[postproc:443] Post Processing Failed for Paris - Hi-Fi ()
2011-10-15 10:02:20,964::INFO::[postproc:445] Traceback: 
Traceback (most recent call last):
  File "/home/ubuntu/SABnzbd-0.6.10RC1/sabnzbd/postproc.py", line 259, in process_job
    par_error, re_add = parring(nzo, workdir)
  File "/home/ubuntu/SABnzbd-0.6.10RC1/sabnzbd/postproc.py", line 534, in parring
    par_error = par_error or not sfv_check(sfv)
  File "/home/ubuntu/SABnzbd-0.6.10RC1/sabnzbd/newsunpack.py", line 1349, in sfv_check
    logging.warning('File %s mssing in SFV check', latin1(path))
  File "/home/ubuntu/SABnzbd-0.6.10RC1/sabnzbd/encoding.py", line 152, in latin1
    return txt.decode('utf-8').encode('latin-1', 'replace').replace('?', '_')
  File "/usr/lib/python2.7/encodings/utf_8.py", line 16, in decode
    return codecs.utf_8_decode(input, errors, True)
UnicodeDecodeError: 'utf8' codec can't decode byte 0xba in position 88: invalid start byte
2011-10-15 10:02:20,982::INFO::[postproc:472] Cleaning up Paris - Hi-Fi (keep_basic=True)
2011-10-15 10:02:20,985::INFO::[misc:1168] Cannot remove folder /home/ubuntu/Downloads/incomplete/Paris - Hi-Fi/__ADMIN__
2011-10-15 10:02:20,987::INFO::[misc:1168] Cannot remove folder /home/ubuntu/Downloads/incomplete/Paris - Hi-Fi/__ADMIN__
2011-10-15 10:02:20,987::INFO::[postproc:83] Saving postproc queue

[sander]

PS:

The resulting file names contain non-US-ASCII characters. Maybe that's the cause of the problems?
When I run cksfv -g on the .sfv file, it reports "Errors Occured"

Still I think SAB should not bail out; I was able to download the full download with nzbget.

[shypike]

Good catch!
The problem is triggered by the content of the SFV file, the file names in it are coded in Latin-1 (as usual for Windows).
So on Windows it will work fine, but not on Linux and OSX.
The issue is that SABnzbd does not convert the file names to an encoding suitable for the platform.
This causes two problems: some files cannot verified and the subsequent logs fail too because of an unexpected encoding.
This will be fixed in 0.6.10

[sander]

I'm now running 0.6.10, and the same NZB (with SFV) goes without problems.

Nice work, Shy!