Comment 4 for bug 1310598

Revision history for this message
Zakhar (alainb06) wrote :

After some months where it didn't seem to crash -or at least apport didn't report anything-, now this is crashing again with the same stack trace about reading my UTF-8/non-ascii mounts.

I'm sending new apports automatic crash reports.

Although, Christian, I would find it normal that it crashes with
LANG=C aa-status

... because that defaults to english/ASCII, and then you should NOT have non-ascii mounts.

If I do
aa-status (implied with my locale fr_FR.UTF-8) it works fine and says:
apparmor module is loaded.
22 profiles are loaded.
22 profiles are in enforce mode.
   /sbin/dhclient
   /usr/bin/evince
   /usr/bin/evince-previewer
   /usr/bin/evince-previewer//sanitized_helper
   /usr/bin/evince-thumbnailer
   /usr/bin/evince-thumbnailer//sanitized_helper
   /usr/bin/evince//sanitized_helper
   /usr/bin/vidalia
   /usr/lib/NetworkManager/nm-dhcp-client.action
   /usr/lib/connman/scripts/dhclient-script
   /usr/lib/cups/backend/cups-pdf
   /usr/lib/lightdm/lightdm-guest-session
   /usr/lib/lightdm/lightdm-guest-session//chromium
   /usr/lib/telepathy/mission-control-5
   /usr/lib/telepathy/telepathy-*
   /usr/lib/telepathy/telepathy-*//pxgsettings
   /usr/lib/telepathy/telepathy-*//sanitized_helper
   /usr/lib/telepathy/telepathy-ofono
   /usr/sbin/cups-browsed
   /usr/sbin/cupsd
   /usr/sbin/tcpdump
   system_tor
0 profiles are in complain mode.
4 processes have profiles defined.
4 processes are in enforce mode.
   /sbin/dhclient (1375)
   /usr/lib/telepathy/mission-control-5 (2551)
   /usr/sbin/cups-browsed (1371)
   /usr/sbin/cupsd (2606)
0 processes are in complain mode.
0 processes are unconfined but have a profile defined.

Thus, although there is a crash reported at startup, it seems that I still have app-armor up an running.

The bug might then be at startup. Could be that when aa is run at startup the language is not yet defined, thus it is running with default LANG=C and failing (which is then kind of "normal situation" as the mounts can use any encoding they see fit!).

I'll try to investigate a little bit more around that idea.