1357437 - Crash in mozilla::dom::workers::serviceWorkerScriptCache::`anonymous namespace''::CompareNetwork::Initialize

Status: RESOLVED FIXED

(Core :: DOM: Service Workers, defect)

Description

[Marcia Knous [:marcia]]

This bug was filed from the Socorro interface and is 
report bp-06c8aaaa-f784-4bc4-b888-417700170418.
=============================================================

Seen while looking at crash stats - crashes started using 20170418030220: http://bit.ly/2nZDP1Z


Possible regression range based on Build ID: https://hg.mozilla.org/mozilla-central/pushloghtml?fromchange=c697e756f738ce37abc56f31bfbc48f55625d617&tochange=bb38d935d699e0529f9e0bb35578d381026415c4

Bug 1290951? ni on :HoPang

Comment 1

[Guilherme Lima]

Hi, I'm the user from the crash report mentioned in the description. I always crash at https://www.uol.com.br/, either while loading or reloading the page. Started happening in the latest Nightly (18-04-2017). Nobody else on Mozillazine can reproduce, unfortunately.

Comment 2

[Marcia Knous [:marcia]]

(In reply to Guilherme Lima from comment #1)
> Hi, I'm the user from the crash report mentioned in the description. I
> always crash at https://www.uol.com.br/, either while loading or reloading
> the page. Started happening in the latest Nightly (18-04-2017). Nobody else
> on Mozillazine can reproduce, unfortunately.

Hello Guilherme - Can you paste the information from about:support? It appears the crash is only happening on Mac, and I cannot reproduce it on my machine.

Comment 3

[Guilherme Lima]

Attachment: aboutSupport.txt

Comment 4

[Rodrigo]

I can reproduce this every time on Windows 10 with the latest Nightly (18-04-2017) at https://www.faceit.com/.
This my crash sig: https://crash-stats.mozilla.com/report/index/08d0c724-0767-41c1-964f-385f10170419

Comment 5

[Marcia Knous [:marcia]]

Looks as if Comment 2 was incorrect - I don't see any mac crashes. It looks as if from the two comments some people are easily able to repro.

Comment 6

[Ben Hsu [:HoPang]]

Attachment: QFIX

Comment 7

[Ben Hsu [:HoPang]]

Thank you all!

Those information does help a lot. Do you mind applying the QFIX and testing those site again? With those patches, I can no longer reproduce them with refreshing the pages (or loading when there already a service worker). I'll polish the patch if it works.

Comment 8

[Calixte Denizet (:calixte)]

The signature "mozilla::dom::workers::serviceWorkerScriptCache::(anonymous namespace)::CompareManager::FetchScript" appears only on OSX and "CompareManager::FetchScript" appears only on Linux.
The signature "mozilla::dom::workers::serviceWorkerScriptCache::`anonymous namespace''::CompareNetwork::Initialize" is #4 on topcrash (for content process).

Comment 9

[Ben Hsu [:HoPang]]

Attachment: Short circuit script fetching when the registering is already failed

Comment 10

[Ben Hsu [:HoPang]]

Hi all, and sorry for my carelessness.

This crash is caused by lacking an appropriate error handling when fetching a script to compare. Though I think the patch provided can fix this crash, however, I think we should back out the code at this moment, since I need some more time to check it all over again.

Comment 11

[Carsten Book [:Tomcat]]

bug 1290951 was backed out now

Comment 12

[trg16]

This crash has started again, 100% reproducible at https://www.uol.com.br/
https://crash-stats.mozilla.com/report/index/b057efbc-d4b4-4f6e-a2b3-fbe390170511

Comment 13

[Stefan Fleiter (:sfleiter)]

For me uol.com.br does not cause a crash, but https://plus.google.com/ does immediately with this signature, see https://crash-stats.mozilla.com/report/index/528dd160-be31-49f4-b165-e2b7e0170511

Comment 14

[Stefan Fleiter (:sfleiter)]

Does not crash anymore.
For me only was the nightly from my crash report above.

Comment 15

[Hsin-Yi Tsai (she/her) [:hsinyi]]

As bug 1290951 was backed out from 55, anything else we need to do in this bug or can we close this?

Comment 16

[Ben Hsu [:HoPang]]

Sorry for the late, since bug 1290951 is going to reland again, I personally think that we should keep this bug alive for a little bit longer, in case there is anything going buggy. However, I don't have a strong preference here.

Comment 17

[Hsin-Yi Tsai (she/her) [:hsinyi]]

(In reply to Ben Hsu [:HoPang] from comment #16)
> Sorry for the late, since bug 1290951 is going to reland again, I personally
> think that we should keep this bug alive for a little bit longer, in case
> there is anything going buggy. However, I don't have a strong preference
> here.

Sure for keeping this open for monitoring the relanding on 56. And that sounds to me that there's no action we currently consider to take for beta 55. I changed the firefox55-status to "fixed" as after bug 1290951 was backed out from 55, crash then stopped.  :HoPang, let me know if I misunderstood anything.

Comment 18

[Hsin-Yi Tsai (she/her) [:hsinyi]]

The crash stopped after bug 1290951 was backed out then reland. Going to close this.