[sync] Fix deadlock in SessionModelAssociator for Android.

[sync] Fix deadlock in SessionModelAssociator for Android.

If there exists session sync nodes that use the cache GUID as the ID,
Chrome for Android goes into a deadlock while associating the tabs.

Since SessionModelAssociator::DeleteForeignSession uses its own
syncer::WriteTransaction, we must ensure that we only call that method
without currently holding a syncer::WriteTransaction.

BUG=175782
Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=182416

[nyquist, 2013-02-13 19:56:19]

PTAL (yfriedman mostly FYI)

[Nicolas Zea, 2013-02-13 19:57:19]

lgtm

[Yaron, 2013-02-13 21:01:46]

lgtm

https://codereview.chromium.org/12253010/diff/1/chrome/browser/sync/glue/sess...
File chrome/browser/sync/glue/session_model_associator.cc (right):

https://codereview.chromium.org/12253010/diff/1/chrome/browser/sync/glue/sess...
chrome/browser/sync/glue/session_model_associator.cc:682: if
(current_machine_tag_.compare(transaction_tag) != 0)
While I'm not excited about add a comment about implementation details, it does
seem worth calling out why this is done afterwards. Either that, or refactor the
code to re-use a transaction and/or a test.

Ok to do refactor/cleanup in a follow-up given the intent to merge to branch.

[commit-bot: I haz the power, 2013-02-13 21:34:43]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/nyquist@chromium.org/12253010/2003

[commit-bot: I haz the power, 2013-02-14 00:59:19]

Retried try job too often on win_rel for step(s) sync_integration_tests
http://build.chromium.org/p/tryserver.chromium/buildstatus?builder=win_rel&nu...

[commit-bot: I haz the power, 2013-02-14 01:03:13]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/nyquist@chromium.org/12253010/2003

[commit-bot: I haz the power, 2013-02-14 08:15:12]

Change committed as 182416