apport crashed with error in report.py in anonymize(): nothing to repeat - if GECOS field contains regular expression metacharacters

Status changed to 'Confirmed' because the bug affects multiple users.

I'm stil having this problem in one particular system, only with my normal user (not with the guest user or running apport with "sudo"). The user name is not strange, and I use it in a lot of other systems and apport does not produce this error.

I found a workaround by commenting out lines 1296 and 1297 of /usr/lib/python2.7/dist-packages/apport/report.py. This seems to be some sort of problem with a regular expression in this line, similar to what is being discussed here:

http://stackoverflow.com/questions/3675144/regex-error-nothing-to-repeat

But what I don't understand is why I only get this error in one particular system.

I'm still having this error on the same system. I have update from 11.10 to 12.04 and now to 12.10 and the error persists. Whenever I use apport I get this error (nothing to repeat). I don't get it if I run apport with "sudo", nor in the guest account. What configuration files of apport may be causing this?

There are several workarounds suggested:

    Commenting out lines 1296 and 1297 of /usr/lib/python2.7/dist-packages/apport/report.py.
    Changing user name or creating a new user account.

Two related bugs in python:

    http://bugs.python.org/issue2537
    http://bugs.python.org/issue214033

12.10 unity-compiz

same problem, different python, tried commenting out 1296 and 1297 they look the same in python3.

                    # avoid ':' in ELF paths, we use that as separator
                    stack.append(offset.replace(':', '..'))

since one is already commented im guessing these are not the correct lines to comment out. Would appreciate a working workaround. thanks.

Traceback (most recent call last):
  File "/usr/share/apport/apport-gtk", line 588, in <module>
    app.run_argv()
  File "/usr/lib/python3/dist-packages/apport/ui.py", line 637, in run_argv
    return self.run_report_bug()
  File "/usr/lib/python3/dist-packages/apport/ui.py", line 439, in run_report_bug
    self.collect_info(symptom_script)
  File "/usr/lib/python3/dist-packages/apport/ui.py", line 1041, in collect_info
    anonymize_thread.exc_raise()
  File "/usr/lib/python3/dist-packages/apport/REThread.py", line 64, in exc_raise
    raise self._exception[1].with_traceback(self._exception[2])
  File "/usr/lib/python3/dist-packages/apport/REThread.py", line 34, in run
    self._retval = self.__target(*self.__args, **self.__kwargs)
  File "/usr/lib/python3/dist-packages/apport/report.py", line 1340, in anonymize
    replacements.append((re.compile('\\b%s\\b' % s), 'User Name'))
  File "/usr/lib/python3.2/re.py", line 214, in compile
    return _compile(pattern, flags)
  File "/usr/lib/python3.2/re.py", line 263, in _compile
    return _compile_typed(type(pattern), pattern, flags)
  File "/usr/lib/python3.2/functools.py", line 184, in wrapper
    result = user_function(*args, **kwds)
  File "/usr/lib/python3.2/re.py", line 275, in _compile_typed
    return sre_compile.compile(pattern, flags)
  File "/usr/lib/python3.2/sre_compile.py", line 491, in compile
    p = sre_parse.parse(p, flags)
  File "/usr/lib/python3.2/sre_parse.py", line 692, in parse
    p = _parse_sub(source, pattern, 0)
  File "/usr/lib/python3.2/sre_parse.py", line 315, in _parse_sub
    itemsappend(_parse(source, state))
  File "/usr/lib/python3.2/sre_parse.py", line 520, in _parse
    raise error("nothing to repeat")
sre_constants.error: nothing to repeat

I'm still having this issue in the same system now updated to raring. My workaround is to comment out the "if" statement at the line number mentioned at the end of the traceback reported. In my last case I get

File "/usr/lib/python3.3/sre_parse.py", line 557, in _parse
    raise error("nothing to repeat")
sre_constants.error: nothing to repeat

as the last three lines before apport stopped. So I edited "/usr/lib/python3.3/sre_parse.py" and commented out lines 556 and 557, which now look as follows

# if not item or (_len(item) == 1 and item[0][0] == AT):
# raise error("nothing to repeat")

Still curious to know what causes this issue, I've tried purging apport, crash reports, but to no avail, I always get the above error as soon as apport or python get some update.

This is because the GECOS field contains a string starting with a '+' (for example a phone number) and the string that is replaced in the regex to anonymize user name is not escaped.

The code below is an excerpt of apport.report.anonymize(self) that reproduces the issue:
==========
import re
p='My User-Name,12.123,+123456789,+123456789,'
replacements=[]
for s in p.split(','):
    s = s.strip()
    if len(s) > 2:
        print(s)
        replacements.append((re.compile('\\b%s\\b' % s), 'User Name'))
==========

The GECOS field should be escaped as it contains general information and forbidden characters in regexp.

Fixed in trunk r2622

This bug was fixed in the package apport - 2.10-0ubuntu1

---------------
apport (2.10-0ubuntu1) saucy; urgency=low

  [ Martin Pitt ]
  * New upstream release. Changes since our previous snapshot:
    - Support retracing foreign powerpc reports, thanks Steve Langasek.
    - apport/report.py: Generate a crash signature for suspend/resume failures.
    - hookutils.py: Add attach_upstart_logs(), thanks Steve Langasek.
    - hookutils.py, in_session_of_problem(): Port from ConsoleKit to logind.
    - apport/report.py: Handle the case where the user has been removed from
      the system, but one of its still-running binaries crashes (LP: #1163453).
    - Fix anonymization of user/GECOS fields with regexp control characters
      like '+'. (LP: #985049)
    - Run tests under LC_CTYPE=C and unset LANG as well, to discover
      assumptions about UTF-8 locales. Fix the two failing tests.
    - Fix UnicodeDecodeError when apport encounters non-ASCII environment
      variables. (LP: #1172638)
  * debian/control: Update Vcs-* for saucy branch.
  * data/general-hooks/cloud_archive.py: Fix PEP-8 error.

  [ Steve Langasek ]
  * data/general-hooks/ubuntu.py: Call attach_upstart_logs(), to attach the
    logs for any upstart user jobs shipped by the binary package.
 -- Martin Pitt <email address hidden> Tue, 30 Apr 2013 14:49:42 -0700

This is still not fixed in Raring 13.04. I just linked a bugreport created some minutes ago on my fully up-to-date 13.04. See #1220267