apport-gtk crashed with OSError in mark_report_seen()

[Matti Lindell]

Binary package hint: apport

I reckon apport didn't like that I manually removed files from /var/crash when it was displaying one of the earlier crash reports.

ProblemType: Crash
Architecture: i386
CrashReports: 600:1000:1000:25681:2007-06-26 20:05:25.000000000 +0300:2007-06-26 20:05:32.000000000 +0300:/var/crash/_usr_share_apport_apport-gtk.1000.crash

Date: Tue Jun 26 20:05:27 2007
DistroRelease: Ubuntu 7.10
ExecutablePath: /usr/share/apport/apport-gtk
InterpreterPath: /usr/bin/python2.5
Package: apport-gtk 0.85
PackageArchitecture: all
ProcCmdline: /usr/bin/python /usr/share/apport/apport-gtk
ProcCwd: /home/matti
ProcEnviron:
 PATH=/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games
 LANG=en_US.utf8
 SHELL=/bin/bash
PythonArgs: ['/usr/share/apport/apport-gtk']
SourcePackage: apport
Title: apport-gtk crashed with OSError in mark_report_seen()
Uname: Linux sandbox 2.6.22-7-generic #1 SMP Mon Jun 25 17:33:14 GMT 2007 i686 GNU/Linux

[Martin Pitt (pitti)]

Thank you for your bug report. I have all information I need now. I will fix it soon.

[Martin Pitt (pitti)]

apport (0.91) gutsy; urgency=low

  * bin/apport: Remove code that supported the Edgy kernel way of core dump
    passing. Also factorize the CORE_REAL_RLIM evaluation, since it is likely
    to change in the near future.
  * apport/crashdb_impl/launchpad.py, close_duplicate(): Delete some
    attachments, as specified in apport-crash-duplicates spec, and make the
    bug public afterwards.
  * apport/crashdb_impl/launchpad.py, close_duplicate(): If the master bug is
    already duped to yet another bug, mark the bug to that one instead of the
    master.
  * apport/crashdb.py: Split out duplicate_db_last_consolidation() for getting
    the date (or seconds since) the last consolidation, so that we can use it
    externally.
  * apport/crashdb.py: Add duplicate_db_change_master_id() to change the
    master ID of a crash. Add test case to apport/crashdb_impl/memory.py.
  * Add bin/dupdb-admin: Initial version of duplicate db CLI app; can dump the
    db, display consolidation state, and change master bug IDs for now. Ship
    it in apport-retrace.
  * apport/crashdb.py, duplicate_db_last_consolidation(): Fix timedelta
    seconds calculation to actually take the days into account, too.
  * bin/crash-digger: Fix dumping of dup db after consolidation.
  * apport/ui.py:
    - test_run_report_bug_package(): Add test case for calling the UI in bug
      filing mode with an invalid package name.
    - run_report_bug(): Do not crash on invalid package name, generate an
      error message instead. (LP: #123644)
  * apport/fileutils.py, mark_report_seen(): Do not crash if the file has
    already been deleted underneath us. (LP: #122347)
  * apport/ui.py, run_report_bug(): Do not crash if the target process runs as
    a different user. Print a proper error message instead. Add test case
    test_run_report_bug_noperm_pid(). (LP: #121121)
  * apport/fileutils.py, likely_packaged(): Ignore /var/lib/schroot. Add test
    case. (LP: #122859)
  * apport/ui.py, open_url(): Intercept weird race condition for os.close()
    trying to close an already invalidated fd. (LP: #123180)

  Merge the fedora branch, thanks to Will Woods <email address hidden>:

  * Add apport.init.fedora: Fedora specific init script.
  * Add apport.spec: RPM build recipe.
  * Add backends/packaging_rpm.py: Partial implementation of the packaging
    backend for RPM which applies to all RPM-based distros.
  * Add backends/packaging_fedora.py: Concrete packaging backend
    implementation for Fedora.
  * apport/elfcore.py: Classes for parsing general ELF files, and information
    from core dumps.
  * bin/apport: Fall back to reading signal number and PID directly from the
    core file (via elfcore.py) if CORE_SIGNAL and CORE_PID are not defined (i.
    e. when running on a non-Ubuntu kernel).
  * crashdb.conf: Add stanzas for Fedora and a 'debug' database which uses the
    'memory' crashdb implementation.

 -- Martin Pitt <email address hidden> Sat, 14 Jul 2007 15:08:35 +0200