Revert of Fix WebRequest's EventListener::operator<

Revert of Fix WebRequest's EventListener::operator< (patchset #3 id:40001 of https://codereview.chromium.org/2121873002/ )

Reason for revert:
This may have caused a perf regression.  Reverting to verify.  If it did not, I'll reland next week.

BUG=641958, 643025
Original issue's description:
> Fix WebRequest's EventListener::operator<
>
> std::set's count() method is expected to return 0 or 1. But due
> to the fact that EventListener::operator< conditionally ignored
> |embedder_process_id|, the count method could return more than 1
> (e.g. 2) in release builds on Windows (VC++), and fail an assertion.
>
> This assertion failure shows that the assumption about the uniqueness
> of EventListeners (while ignoring the |embedder_process_id| value) does
> not hold, i.e. there may be more than one non-webview EventListeners
> with the same |extension_id| and |sub_event_name|, but different
> |embedder_process_id|.
>
> To fix this, this patch removes the conditional exclusion of
> |embedder_process_id| from operator< and looks up |embedder_process_id|
> before searching through the set.
>
> BUG=589735
> TEST=./unit_tests --gtest_filter=ExtensionWebRequestTest.AddAndRemoveListeners
>
> Committed: https://crrev.com/c89943b168ef2ccc51f3ec0720e552f36ec31351
> Cr-Commit-Position: refs/heads/master@{#404456}

TBR=rockot@chromium.org,rob@robwu.nl
# Not skipping CQ checks because original CL landed more than 1 days ago.
BUG=589735
Committed: https://crrev.com/e579d519620e6583ddc304f22c4c3c71dabbbe20
Cr-Commit-Position: refs/heads/master@{#416094}

[mmenke, 2016-09-01 21:30:44]

The CQ bit was checked by mmenke@chromium.org

[mmenke, 2016-09-01 21:30:44]

Created Revert of Fix WebRequest's EventListener::operator<

[commit-bot: I haz the power, 2016-09-01 21:32:25]

CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[mmenke, 2016-09-01 21:32:57]

Description was changed from

==========
Revert of Fix WebRequest's EventListener::operator< (patchset #3 id:40001 of
https://codereview.chromium.org/2121873002/ )

Reason for revert:
This may have caused https://crbug.com/641958.  Reverting to verify.  If it did
not, I'll reland next week.

Original issue's description:
> Fix WebRequest's EventListener::operator<
> 
> std::set's count() method is expected to return 0 or 1. But due
> to the fact that EventListener::operator< conditionally ignored
> |embedder_process_id|, the count method could return more than 1
> (e.g. 2) in release builds on Windows (VC++), and fail an assertion.
> 
> This assertion failure shows that the assumption about the uniqueness
> of EventListeners (while ignoring the |embedder_process_id| value) does
> not hold, i.e. there may be more than one non-webview EventListeners
> with the same |extension_id| and |sub_event_name|, but different
> |embedder_process_id|.
> 
> To fix this, this patch removes the conditional exclusion of
> |embedder_process_id| from operator< and looks up |embedder_process_id|
> before searching through the set.
> 
> BUG=589735
> TEST=./unit_tests --gtest_filter=ExtensionWebRequestTest.AddAndRemoveListeners
> 
> Committed: https://crrev.com/c89943b168ef2ccc51f3ec0720e552f36ec31351
> Cr-Commit-Position: refs/heads/master@{#404456}

TBR=rockot@chromium.org,rob@robwu.nl
# Not skipping CQ checks because original CL landed more than 1 days ago.
BUG=589735
==========

to

==========
Revert of Fix WebRequest's EventListener::operator< (patchset #3 id:40001 of
https://codereview.chromium.org/2121873002/ )

Reason for revert:
This may have caused a perf regression.  Reverting to verify.  If it did not,
I'll reland next week.

BUG=641958

Original issue's description:
> Fix WebRequest's EventListener::operator<
> 
> std::set's count() method is expected to return 0 or 1. But due
> to the fact that EventListener::operator< conditionally ignored
> |embedder_process_id|, the count method could return more than 1
> (e.g. 2) in release builds on Windows (VC++), and fail an assertion.
> 
> This assertion failure shows that the assumption about the uniqueness
> of EventListeners (while ignoring the |embedder_process_id| value) does
> not hold, i.e. there may be more than one non-webview EventListeners
> with the same |extension_id| and |sub_event_name|, but different
> |embedder_process_id|.
> 
> To fix this, this patch removes the conditional exclusion of
> |embedder_process_id| from operator< and looks up |embedder_process_id|
> before searching through the set.
> 
> BUG=589735
> TEST=./unit_tests --gtest_filter=ExtensionWebRequestTest.AddAndRemoveListeners
> 
> Committed: https://crrev.com/c89943b168ef2ccc51f3ec0720e552f36ec31351
> Cr-Commit-Position: refs/heads/master@{#404456}

TBR=rockot@chromium.org,rob@robwu.nl
# Not skipping CQ checks because original CL landed more than 1 days ago.
BUG=589735
==========

[mmenke, 2016-09-01 21:51:19]

Description was changed from

==========
Revert of Fix WebRequest's EventListener::operator< (patchset #3 id:40001 of
https://codereview.chromium.org/2121873002/ )

Reason for revert:
This may have caused a perf regression.  Reverting to verify.  If it did not,
I'll reland next week.

BUG=641958

Original issue's description:
> Fix WebRequest's EventListener::operator<
> 
> std::set's count() method is expected to return 0 or 1. But due
> to the fact that EventListener::operator< conditionally ignored
> |embedder_process_id|, the count method could return more than 1
> (e.g. 2) in release builds on Windows (VC++), and fail an assertion.
> 
> This assertion failure shows that the assumption about the uniqueness
> of EventListeners (while ignoring the |embedder_process_id| value) does
> not hold, i.e. there may be more than one non-webview EventListeners
> with the same |extension_id| and |sub_event_name|, but different
> |embedder_process_id|.
> 
> To fix this, this patch removes the conditional exclusion of
> |embedder_process_id| from operator< and looks up |embedder_process_id|
> before searching through the set.
> 
> BUG=589735
> TEST=./unit_tests --gtest_filter=ExtensionWebRequestTest.AddAndRemoveListeners
> 
> Committed: https://crrev.com/c89943b168ef2ccc51f3ec0720e552f36ec31351
> Cr-Commit-Position: refs/heads/master@{#404456}

TBR=rockot@chromium.org,rob@robwu.nl
# Not skipping CQ checks because original CL landed more than 1 days ago.
BUG=589735
==========

to

==========
Revert of Fix WebRequest's EventListener::operator< (patchset #3 id:40001 of
https://codereview.chromium.org/2121873002/ )

Reason for revert:
This may have caused a perf regression.  Reverting to verify.  If it did not,
I'll reland next week.

BUG=641958,643025

Original issue's description:
> Fix WebRequest's EventListener::operator<
> 
> std::set's count() method is expected to return 0 or 1. But due
> to the fact that EventListener::operator< conditionally ignored
> |embedder_process_id|, the count method could return more than 1
> (e.g. 2) in release builds on Windows (VC++), and fail an assertion.
> 
> This assertion failure shows that the assumption about the uniqueness
> of EventListeners (while ignoring the |embedder_process_id| value) does
> not hold, i.e. there may be more than one non-webview EventListeners
> with the same |extension_id| and |sub_event_name|, but different
> |embedder_process_id|.
> 
> To fix this, this patch removes the conditional exclusion of
> |embedder_process_id| from operator< and looks up |embedder_process_id|
> before searching through the set.
> 
> BUG=589735
> TEST=./unit_tests --gtest_filter=ExtensionWebRequestTest.AddAndRemoveListeners
> 
> Committed: https://crrev.com/c89943b168ef2ccc51f3ec0720e552f36ec31351
> Cr-Commit-Position: refs/heads/master@{#404456}

TBR=rockot@chromium.org,rob@robwu.nl
# Not skipping CQ checks because original CL landed more than 1 days ago.
BUG=589735
==========

[commit-bot: I haz the power, 2016-09-01 22:41:32]

Description was changed from

==========
Revert of Fix WebRequest's EventListener::operator< (patchset #3 id:40001 of
https://codereview.chromium.org/2121873002/ )

Reason for revert:
This may have caused a perf regression.  Reverting to verify.  If it did not,
I'll reland next week.

BUG=641958,643025

Original issue's description:
> Fix WebRequest's EventListener::operator<
> 
> std::set's count() method is expected to return 0 or 1. But due
> to the fact that EventListener::operator< conditionally ignored
> |embedder_process_id|, the count method could return more than 1
> (e.g. 2) in release builds on Windows (VC++), and fail an assertion.
> 
> This assertion failure shows that the assumption about the uniqueness
> of EventListeners (while ignoring the |embedder_process_id| value) does
> not hold, i.e. there may be more than one non-webview EventListeners
> with the same |extension_id| and |sub_event_name|, but different
> |embedder_process_id|.
> 
> To fix this, this patch removes the conditional exclusion of
> |embedder_process_id| from operator< and looks up |embedder_process_id|
> before searching through the set.
> 
> BUG=589735
> TEST=./unit_tests --gtest_filter=ExtensionWebRequestTest.AddAndRemoveListeners
> 
> Committed: https://crrev.com/c89943b168ef2ccc51f3ec0720e552f36ec31351
> Cr-Commit-Position: refs/heads/master@{#404456}

TBR=rockot@chromium.org,rob@robwu.nl
# Not skipping CQ checks because original CL landed more than 1 days ago.
BUG=589735
==========

to

==========
Revert of Fix WebRequest's EventListener::operator< (patchset #3 id:40001 of
https://codereview.chromium.org/2121873002/ )

Reason for revert:
This may have caused a perf regression.  Reverting to verify.  If it did not,
I'll reland next week.

BUG=641958,643025

Original issue's description:
> Fix WebRequest's EventListener::operator<
> 
> std::set's count() method is expected to return 0 or 1. But due
> to the fact that EventListener::operator< conditionally ignored
> |embedder_process_id|, the count method could return more than 1
> (e.g. 2) in release builds on Windows (VC++), and fail an assertion.
> 
> This assertion failure shows that the assumption about the uniqueness
> of EventListeners (while ignoring the |embedder_process_id| value) does
> not hold, i.e. there may be more than one non-webview EventListeners
> with the same |extension_id| and |sub_event_name|, but different
> |embedder_process_id|.
> 
> To fix this, this patch removes the conditional exclusion of
> |embedder_process_id| from operator< and looks up |embedder_process_id|
> before searching through the set.
> 
> BUG=589735
> TEST=./unit_tests --gtest_filter=ExtensionWebRequestTest.AddAndRemoveListeners
> 
> Committed: https://crrev.com/c89943b168ef2ccc51f3ec0720e552f36ec31351
> Cr-Commit-Position: refs/heads/master@{#404456}

TBR=rockot@chromium.org,rob@robwu.nl
# Not skipping CQ checks because original CL landed more than 1 days ago.
BUG=589735
==========

[commit-bot: I haz the power, 2016-09-01 22:41:33]

Committed patchset #1 (id:1)

[commit-bot: I haz the power, 2016-09-01 22:43:38]

Description was changed from

==========
Revert of Fix WebRequest's EventListener::operator< (patchset #3 id:40001 of
https://codereview.chromium.org/2121873002/ )

Reason for revert:
This may have caused a perf regression.  Reverting to verify.  If it did not,
I'll reland next week.

BUG=641958,643025

Original issue's description:
> Fix WebRequest's EventListener::operator<
> 
> std::set's count() method is expected to return 0 or 1. But due
> to the fact that EventListener::operator< conditionally ignored
> |embedder_process_id|, the count method could return more than 1
> (e.g. 2) in release builds on Windows (VC++), and fail an assertion.
> 
> This assertion failure shows that the assumption about the uniqueness
> of EventListeners (while ignoring the |embedder_process_id| value) does
> not hold, i.e. there may be more than one non-webview EventListeners
> with the same |extension_id| and |sub_event_name|, but different
> |embedder_process_id|.
> 
> To fix this, this patch removes the conditional exclusion of
> |embedder_process_id| from operator< and looks up |embedder_process_id|
> before searching through the set.
> 
> BUG=589735
> TEST=./unit_tests --gtest_filter=ExtensionWebRequestTest.AddAndRemoveListeners
> 
> Committed: https://crrev.com/c89943b168ef2ccc51f3ec0720e552f36ec31351
> Cr-Commit-Position: refs/heads/master@{#404456}

TBR=rockot@chromium.org,rob@robwu.nl
# Not skipping CQ checks because original CL landed more than 1 days ago.
BUG=589735
==========

to

==========
Revert of Fix WebRequest's EventListener::operator< (patchset #3 id:40001 of
https://codereview.chromium.org/2121873002/ )

Reason for revert:
This may have caused a perf regression.  Reverting to verify.  If it did not,
I'll reland next week.

BUG=641958,643025

Original issue's description:
> Fix WebRequest's EventListener::operator<
>
> std::set's count() method is expected to return 0 or 1. But due
> to the fact that EventListener::operator< conditionally ignored
> |embedder_process_id|, the count method could return more than 1
> (e.g. 2) in release builds on Windows (VC++), and fail an assertion.
>
> This assertion failure shows that the assumption about the uniqueness
> of EventListeners (while ignoring the |embedder_process_id| value) does
> not hold, i.e. there may be more than one non-webview EventListeners
> with the same |extension_id| and |sub_event_name|, but different
> |embedder_process_id|.
>
> To fix this, this patch removes the conditional exclusion of
> |embedder_process_id| from operator< and looks up |embedder_process_id|
> before searching through the set.
>
> BUG=589735
> TEST=./unit_tests --gtest_filter=ExtensionWebRequestTest.AddAndRemoveListeners
>
> Committed: https://crrev.com/c89943b168ef2ccc51f3ec0720e552f36ec31351
> Cr-Commit-Position: refs/heads/master@{#404456}

TBR=rockot@chromium.org,rob@robwu.nl
# Not skipping CQ checks because original CL landed more than 1 days ago.
BUG=589735

Committed: https://crrev.com/e579d519620e6583ddc304f22c4c3c71dabbbe20
Cr-Commit-Position: refs/heads/master@{#416094}
==========

[commit-bot: I haz the power, 2016-09-01 22:43:40]

Patchset 1 (id:??) landed as
https://crrev.com/e579d519620e6583ddc304f22c4c3c71dabbbe20
Cr-Commit-Position: refs/heads/master@{#416094}