1289183 - Crashes for signature _preextrapolate_helper disappeared

Status: RESOLVED INVALID

(Socorro :: Webapp, task)

Description

[Nils Ohlmeier [:drno]]

End of June users reported a WebRTC related crash back in bug 1283069.
In comment #1 Byron shared this link to crash-stats:

https://crash-stats.mozilla.com/signature/?signature=_preextrapolate_helper&_columns=date&_columns=product&_columns=version&_columns=build_id&_columns=platform&_columns=reason&_columns=address&page=1#reports

I'm 100% sure that when I commented on the bug on 2016-6-30 that query for the crash signature '_preextrapolate_helper' showed something like 10 crashes or so with the same signature.
When I click on that link today I see zero reports.

The second strange part, and potentially related, is that the bug report has two direct links to crash reports in it:
https://crash-stats.mozilla.com/report/index/fa016cb2-8558-4765-8ce4-397602160629
https://crash-stats.mozilla.com/report/index/16b37840-95f1-4eeb-b94f-606b52160701

which in fact have the same crash signature in '_preextrapolate_helper', but still crash-stats claims that there are no crashes with the same signature?!

The strange or not very trustworthy thing about these reports is that they claim that the crashing thread had only a single function call on the stack. Has there maybe been some feature been added to hide unrealistic signatures with crashing stacks with less then X frames from the signature search?

Comment 1

[[DEACTIVATED] Adrian Gaudebert]

The Signature report page, as well as the Super Search page, have a date filter set by default to "the last 7 days". So that link you pasted will only show recent results. It appears there has not been any crash reports with that signature in the past week. However, if you set a date filter, you can see older results, like this: https://crash-stats.mozilla.com/signature/?date=%3E2016-06-20&signature=_preextrapolate_helper&_columns=date&_columns=product&_columns=version&_columns=build_id&_columns=platform&_columns=reason&_columns=address&_sort=-date&page=1#reports (the date filter can be seen by clicking the "Show" link in the top right corner).

(In reply to Nils Ohlmeier [:drno] from comment #0)
> The strange or not very trustworthy thing about these reports is that they
> claim that the crashing thread had only a single function call on the stack.
> Has there maybe been some feature been added to hide unrealistic signatures
> with crashing stacks with less then X frames from the signature search?

There has been no such change recently. It is possible that this crash is not happening anymore, or that the client now sends a different stack trace for it.

Comment 2

[Peter Bengtsson [:peterbe]]

Follow-up bug about making that a lot clearer on the user interface? Even with the Search filter thing expanded it's not clear that it's limited to the last 7 days http://jmp.sh/Ubgf5iK

Comment 3

[Nils Ohlmeier [:drno]]

Thanks. I was not aware of the implicit 7 day filter. Glad there is an easy workaround.