Chromium Code Reviews
chromiumcodereview-hr@appspot.gserviceaccount.com (chromiumcodereview-hr) | Please choose your nickname with Settings | Help | Chromium Project | Gerrit Changes | Sign out
(347)

Issue 6273011: [SYNC] Speculative fix for autofill change processor destroying on the wrong ... (Closed)

Created:
9 years, 11 months ago by Nicolas Zea
Modified:
9 years, 7 months ago
CC:
chromium-reviews, Raghu Simha, ncarter (slow), tim (not reviewing), idana
Visibility:
Public.

Description

[SYNC] Speculative fix for autofill change processor destroying on the wrong thread. The data type manager should now tell any datatype controller that is not already stopping or not running to stop. Namely, this covers the MODEL_STARTING and ASSOCIATING cases, which weren't being covered before. Tests that can reproduce these race conditions still needed. BUG=61804 TEST=unit,sync_integration,self Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=72458

Patch Set 1 #

Unified diffs Side-by-side diffs Delta from patch set Stats (+2 lines, -1 line) Patch
M chrome/browser/sync/glue/data_type_manager_impl.cc View 1 chunk +2 lines, -1 line 0 comments Download

Messages

Total messages: 3 (0 generated)
Nicolas Zea
Mind taking a look? I think this should handle the bug. I'll try and figure ...
9 years, 11 months ago (2011-01-25 01:54:48 UTC) #1
tim (not reviewing)
Thanks for looking at this. Might the assumption have been that the other two states ...
9 years, 11 months ago (2011-01-25 02:00:28 UTC) #2
Nicolas Zea
9 years, 11 months ago (2011-01-25 02:04:59 UTC) #3
Not sure, bug don't think so. At least the autofill data type controller already
already had code in it for handling the cases where it was associating or
model_starting when stop() was called. Comitting.

On 2011/01/25 02:00:28, timsteele wrote:
> Thanks for looking at this.  Might the assumption have been that the other two
> states would abort somehow?
> 
> LGTM

Powered by Google App Engine
This is Rietveld 408576698