Comment 15 for bug 513044

Okay, in seg1-clean (the first crash) there is indeed a hang, but by the time the core was taken, most of the hung threads had been killed by the thread-killer:

16:07:26.423: 1101130064: loggerhead: Starting to process http://bazaar.launchpad.net/%7Evcs-imports/busybox/main/changes
16:12:55.338: Thread 1101130064 hung (working on task for 327 seconds)

There are several other hung threads that are in _parse_record_unchecked in bzrlib.knit. Each thread is at a different line in the same function, though:

File "/home/pqm/for_rollouts/production/eggs/bzr-2.1.0-py2.5-linux-x86_64.egg/bzrlib/knit.py", line 1915, in _parse_record_unchecked
File "/home/pqm/for_rollouts/production/eggs/bzr-2.1.0-py2.5-linux-x86_64.egg/bzrlib/knit.py", line 1905, in _parse_record_unchecked
File "/home/pqm/for_rollouts/production/eggs/bzr-2.1.0-py2.5-linux-x86_64.egg/bzrlib/knit.py", line 1898, in _parse_record_unchecked

The other threads, when killed are in:

  File "/usr/lib/python2.5/site-packages/pygments/lexer.py", line 163, in streamer
     yield t, v

  File "/usr/lib/python2.5/site-packages/pygments/formatters/html.py", line 565, in _format_lines
     if line and parts[-1]:

  File "/home/pqm/for_rollouts/production/eggs/bzr-2.1.0-py2.5-linux-x86_64.egg/bzrlib/pack.py", line 200, in readline

  File "/home/pqm/for_rollouts/production/eggs/bzr-2.1.0-py2.5-linux-x86_64.egg/bzrlib/btree_index.py", line 612, in __init__