984379 - crashes [NS_ABORT_OOM(unsigned int) | nsAString_internal::Assign(wchar_t const*, unsigned int] because of loading streaming videos

Status: VERIFIED FIXED

(Core :: Internationalization, defect, critical)

Description

[Suburp]

User Agent: Mozilla/5.0 (Windows NT 5.1; rv:27.0) Gecko/20100101 Firefox/27.0 (Beta/Release)
Build ID: 20140212131424

Steps to reproduce:

I loaded two parallel youtube videos


Actual results:

firefox crashed, crash reporter appeared (https://crash-stats.mozilla.com/report/index/cba10729-8d3d-4fde-a57f-9e6ec2140302)


Expected results:

loaded normally

Comment 1

[Makoto Kato [:m_kato]]

nsScriptableUnicodeConverter should use fallible_t.

Comment 2

[Makoto Kato [:m_kato]]

Attachment: Fix

Comment 3

[Makoto Kato [:m_kato]]

https://hg.mozilla.org/integration/mozilla-inbound/rev/c8ecf1ba4139

Comment 4

[[e-mail]]

Comment on attachment 8397039 [details] [diff] [review]
Fix

>     // No Adopt on nsACString :(
>-    _retval.Assign(str, len);
>+    if (!_retval.Assign(str, len, mozilla::fallible_t())) {
>+      rv = NS_ERROR_OUT_OF_MEMORY;
>+    }
>     moz_free(str);
The comment lies! It probably predates the "latest" string API overhaul. _retVal.Adopt(str, len) should be fine.

Comment 5

[Makoto Kato [:m_kato]]

(In reply to [e-mail] from comment #4)
> Comment on attachment 8397039 [details] [diff] [review]
> Fix
> 
> >     // No Adopt on nsACString :(
> >-    _retval.Assign(str, len);
> >+    if (!_retval.Assign(str, len, mozilla::fallible_t())) {
> >+      rv = NS_ERROR_OUT_OF_MEMORY;
> >+    }
> >     moz_free(str);
> The comment lies! It probably predates the "latest" string API overhaul.
> _retVal.Adopt(str, len) should be fine.

really?  This issue will continue to bug 989835.

Comment 6

[Carsten Book [:Tomcat]]

https://hg.mozilla.org/mozilla-central/rev/c8ecf1ba4139

Comment 7

[Anthony Hughes (:ashughes) [QA]]

Suburp, can you confirm the latest Nightly no longer crashes for you?

Comment 8

[Anthony Hughes (:ashughes) [QA]]

I'm still seeing two crashes with this signature post-fix.

Firefox 31 2014-04-09: https://crash-stats.mozilla.com/report/index/c1960493-8b08-44fb-811c-e6f592140410
Firefox 31 2014-04-12: https://crash-stats.mozilla.com/report/index/95729cf6-3de5-4270-9e42-5476f2140414

Should this bug get reopened?

Comment 9

[guigs [:guigs] PST (please needinfo me!)]

There is a user that was experiencing this in the forums for 30-31 [https://support.mozilla.org/en-US/questions/994049] Firefox is hanging or crashing. There are some screenshots in the thread as well.

Comment 10

[Lars Gusowski [:lagu]]

i had yesterday a crashe linked to this report in Firefox 33. Therfore it seems to be not fixed :(

Comment 11

[Lars Gusowski [:lagu]]

https://crash-stats.mozilla.com/report/index/d5e42902-5280-4f53-8b92-799d92140613

Comment 12

[Anthony Hughes (:ashughes) [QA]]

(In reply to lars.gusowski+bugzilla from comment #11)
> https://crash-stats.mozilla.com/report/index/d5e42902-5280-4f53-8b92-
> 799d92140613

Thanks for sharing your report, Lars. You might be getting a different crash with the same signature, your stack looks different than what was originally reported here. I'll leave it to the developers on this bug to determine if this bug report should be reopened or if you should file a new bug report.

Comment 13

[Liz Henry (:lizzard)]

m_kato, can you look at lars's crash from comment 11?  Do you think this is a different crash with the same signature? We have several new reports with the same signature:

https://crash-stats.mozilla.com/query/?product=Firefox&version=Firefox%3A31.0b&range_value=1&range_unit=weeks&date=06%2F19%2F2014+22%3A00%3A00&query_search=signature&query_type=contains&query=NS_ABORT_OOM%28unsigned+int%29+|+nsAString_internal%3A%3AAssign%28wchar_t+const*%2C+unsigned+int&reason=&release_channels=&build_id=&process_type=any&hang_type=any

Thanks!

Comment 14

[Makoto Kato [:m_kato]]

(In reply to Liz Henry :lizzard from comment #13)
> m_kato, can you look at lars's crash from comment 11?  Do you think this is
> a different crash with the same signature? We have several new reports with
> the same signature:
> 
> https://crash-stats.mozilla.com/query/?product=Firefox&version=Firefox%3A31.
> 0b&range_value=1&range_unit=weeks&date=06%2F19%2F2014+22%3A00%3A00&query_sear
> ch=signature&query_type=contains&query=NS_ABORT_OOM%28unsigned+int%29+|+nsASt
> ring_internal%3A%3AAssign%28wchar_t+const*%2C+unsigned+int&reason=&release_ch
> annels=&build_id=&process_type=any&hang_type=any
> 
> Thanks!

It is different.   I think we should add nsAString_internal::Assign to skiplist of crash reporter.

Comment 15

[Makoto Kato [:m_kato]]

filed as bug 1033883 for skiplist request.

Comment 16

[Suburp]

(In reply to Anthony Hughes, QA Mentor (:ashughes) from comment #7)
> Suburp, can you confirm the latest Nightly no longer crashes for you?

I couldn't reproduce the crash on the 31, but that doesn't mean it's fixed permanently, as these other comments suggest.

Comment 17

[Anthony Hughes (:ashughes) [QA]]

(In reply to Suburp from comment #16)
> (In reply to Anthony Hughes, QA Mentor (:ashughes) from comment #7)
> > Suburp, can you confirm the latest Nightly no longer crashes for you?
> 
> I couldn't reproduce the crash on the 31, but that doesn't mean it's fixed
> permanently, as these other comments suggest.

Thanks for checking. The other comments will be addressed in a follow-up bug. Marking this verified fixed.