Daisy

indicator-sound crashes frequently fail to retrace

Bug #1318034 reported by Brian Murray
8
This bug affects 1 person
Affects Status Importance Assigned to Milestone
Daisy
New
Undecided
Unassigned

Bug Description

This particular indicator-sound crash fails to retrace all the time:

https://errors.ubuntu.com/problem/0c26bcadfdcdb7aeff92fd904054b0c1eba9ab13

In an effort to find out why I increased the logging of retracer.py to return some information when there is a failure to retrace. It found the following for one of the crashes:

Apport did not return a crash_signature.
StacktraceTop:
g_variant_is_trusted (value=value@entry=0x0) at /build/buildd/glib2.0-2.40.0/./glib/gvariant-core.c:607
g_variant_builder_add_value (builder=0x7fff87ed0f10, value=0x0) at /build/buildd/glib2.0-2.40.0/./glib/gvariant.c:3430
?? ()
accounts_service_user_set_player ()
_start ()
Could not retrace.

So it looks to me like this one instance failed, based off my reading of apport's crash_signature function, because of the line with "?? ()" and the one with "_start ()".

See original description
Revision history for this message
Brian Murray (brian-murray) wrote :

The are actually 2 crashes with the same StacktraceTop that failed to retrace.

Revision history for this message
Brian Murray (brian-murray) wrote :

Martin - is my assessment here correct regarding apport's failure to generate a crash signature?

Revision history for this message
Martin Pitt (pitti) wrote :

The _start() doesn't matter, but indeed the unknown symbol in the third frame breaks it. Report.crash_signature() expects all frames in the StacktraceTop to have a name, otherwise it rejects it (i. e. returns None).

I guess in those cases we need to fall back to using the StacktraceAddressSignature?

Does the log also have a list of the missing/mismatching debug symbols?

description: updated
Revision history for this message
Brian Murray (brian-murray) wrote :

Yes, the log file (which can be found on snakefruit) does have a list of missing or mismatching and there were none. I think we do fall back to the StacktraceAddressSignature, but every single instances of this crash is failing to retrace so it doesn't help.

Revision history for this message
Brian Murray (brian-murray) wrote :

By the way I recently setup for daisy to try retracing new instances of a bucket that failed to retrace, rather than only retracing the first instance of a StacktraceAddressSignature.

To post a comment you must log in.
This report contains Public information  
Everyone can see this information.
Subscribing...

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.