1385893 - Crash in nsCOMPtr_base::~nsCOMPtr_base | nsJARURI::~nsJARURI

Status: RESOLVED WONTFIX

(Core :: Networking: JAR, defect, P3)

Description

[Marcia Knous [:marcia]]

This bug was filed from the Socorro interface and is 
report bp-ca4b0ca7-aa0b-4214-9946-88cda0170728.
=============================================================

This one needs some help bucketing. 

Seen while looking at nightly crash stats - crashes started using 20170727100347: http://bit.ly/2tW8D6y

Possible regression range based on Build ID: https://hg.mozilla.org/mozilla-central/pushloghtml?fromchange=e8400551c2e39f24c75a009ebed496c7acd7bf47&tochange=e5693cea1ec944ca077c7a46c5f127c828a90f1b

Comment 1

[Andrew McCreight [:mccr8]]

Haik, could bug 1376496 have caused this crash? It is in the regression range from comment 0, and seems to involve JARs.

Comment 2

[Haik Aftandilian [:haik]]

(In reply to Andrew McCreight [:mccr8] from comment #1)
> Haik, could bug 1376496 have caused this crash? It is in the regression
> range from comment 0, and seems to involve JARs.

I can't rule it out, but the JAR-related changes in bug 1376496 should only be triggered in the child process and (if I'm reading the crash reports correctly) all the reports are crashes in the parent process. The changes are in nsIJARChannel::EnsureCached() which is called from ExtensionProtocolHandler::SubstituteRemoteJarChannel() which is only executed in the child.

Leaving the needinfo for myself to investigate this more.

Comment 3

[Haik Aftandilian [:haik]]

I looked at this a bit more. The majority of these crashes are in the nsJARChannel destructor in the parent process. Bug 1376496  added a new member (of type nsCOMPtr<nsIZipReader>) to nsJARChannel which will have altered the size of the class, but shouldn't ever be used in the parent process. If bug 1376496 is related, I don't know of how it could trigger these crashes, apart from how it changed the size of nsJARChannel which might be exposing an existing bug.

Some of the crashes don't involve nsJARChannel. An example is https://crash-stats.mozilla.com/report/index/0ae6b147-16e6-44d0-a4bf-d1d880170802

And there haven't been any reports with a build ID later than 20170731100325.

Comment 4

[u408661]

Low-volume (except for the one spike at the end of July), seems to go back a while.

Comment 5

[Firefox Bug Husbandry Bot]

Bulk change to priority: https://bugzilla.mozilla.org/show_bug.cgi?id=1399258

Comment 6

[Firefox Bug Husbandry Bot]

Bulk change to priority: https://bugzilla.mozilla.org/show_bug.cgi?id=1399258

Comment 7

[Sylvestre Ledru [:Sylvestre]]

Closing because no crash reported since 12 weeks.