[tor-bugs] #8049 [Stem]: Stem's DescriptorReader misses 10% of descriptors in tarballs

[Tor Bug Tracker & Wiki]

#8049: Stem's DescriptorReader misses 10% of descriptors in tarballs
---------------------+------------------------------------------------------
 Reporter:  karsten  |          Owner:  atagar  
     Type:  defect   |         Status:  assigned
 Priority:  normal   |      Milestone:          
Component:  Stem     |        Version:          
 Keywords:           |         Parent:          
   Points:           |   Actualpoints:          
---------------------+------------------------------------------------------

Comment(by atagar):

 Ah ha! After a week of banging my head on the wall I got it...

 {{{
 atagar at morrigan:~/Desktop/stem$ python example.py
 279042 descriptors in tarball.
 279042 descriptors in extracted directory.
 }}}

 The difference is that I dropped a single descriptor:

 server-descriptors-2012-12/5/6/564778a0a0aea0d6c6c0ee4bbce9c9233a0d70cd

 Something is seriously screwy about its contact line. In vim it appears
 as...

 ^Tþ^Xcontact root@[ipaddress]

 ... but while parsing it causes an exception because it can't be converted
 to ASCII. When I tried (... repeatedly) to log output the logging
 exception with those characters corrupted the output (causing the log file
 to be truncated to 6k of binary, and peg my cpu when viewed in vim or
 less).

 So... er, yea. Fun. The descriptor file is malformed (since the characters
 appear before the contact keyword) but stem certainly shouldn't choke this
 way on... whatever that is. I'll add this descriptor to our integ tests
 and make it fail more gracefully.

-- 
Ticket URL: <https://trac.torproject.org/projects/tor/ticket/8049#comment:8>
Tor Bug Tracker & Wiki <https://trac.torproject.org/>
The Tor Project: anonymity online