Revert of Change HistoryInfoMap to use unordered_map.

Revert of Change HistoryInfoMap to use unordered_map. (patchset #2 id:20001 of https://codereview.chromium.org/1482443002/ )

Reason for revert:
This introduced a clear regression in the omnibox performance. See bug 574181.

Original issue's description:
> Change HistoryInfoMap to use unordered_map.
>
> This change speeds up the omnibox by 5-10% in my (limited) tests, using a large profile and checking the value of the Omnibox.CharTypedToRepaintLatency histogram.
>
> This was also validated when profiling, where the find method was the top function where CPU cycles were consumed. After switching to a hash map, it was a non factor.
>
> Note: As far as I can tell, the order is not important, so this change doesn't alter expected behavior.
>
> BUG=574181
>
> Committed: https://crrev.com/6ddf83a44cb65d2540e8b78c4ff56d240a1d0597
> Cr-Commit-Position: refs/heads/master@{#362532}

TBR=pkasting@chromium.org,mpearson@chromium.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=574181
Committed: https://crrev.com/554207590d8fb848bd21997a23093efc35ce1dd9
Cr-Commit-Position: refs/heads/master@{#367625}

[Georges Khalil, 2016-01-05 20:09:26]

Created Revert of Change HistoryInfoMap to use unordered_map.

[commit-bot: I haz the power, 2016-01-05 20:10:54]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1566443002/1
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1566443002/1

[commit-bot: I haz the power, 2016-01-05 20:12:46]

Description was changed from

==========
Revert of Change HistoryInfoMap to use unordered_map. (patchset #2 id:20001 of
https://codereview.chromium.org/1482443002/ )

Reason for revert:
This introduced a clear regression in the omnibox performance. See bug 574181.

Original issue's description:
> Change HistoryInfoMap to use unordered_map.
> 
> This change speeds up the omnibox by 5-10% in my (limited) tests, using a
large profile and checking the value of the Omnibox.CharTypedToRepaintLatency
histogram.
> 
> This was also validated when profiling, where the find method was the top
function where CPU cycles were consumed. After switching to a hash map, it was a
non factor.
> 
> Note: As far as I can tell, the order is not important, so this change doesn't
alter expected behavior.
> 
> BUG=574181
> 
> Committed: https://crrev.com/6ddf83a44cb65d2540e8b78c4ff56d240a1d0597
> Cr-Commit-Position: refs/heads/master@{#362532}

TBR=pkasting@chromium.org,mpearson@chromium.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=574181
==========

to

==========
Revert of Change HistoryInfoMap to use unordered_map. (patchset #2 id:20001 of
https://codereview.chromium.org/1482443002/ )

Reason for revert:
This introduced a clear regression in the omnibox performance. See bug 574181.

Original issue's description:
> Change HistoryInfoMap to use unordered_map.
> 
> This change speeds up the omnibox by 5-10% in my (limited) tests, using a
large profile and checking the value of the Omnibox.CharTypedToRepaintLatency
histogram.
> 
> This was also validated when profiling, where the find method was the top
function where CPU cycles were consumed. After switching to a hash map, it was a
non factor.
> 
> Note: As far as I can tell, the order is not important, so this change doesn't
alter expected behavior.
> 
> BUG=574181
> 
> Committed: https://crrev.com/6ddf83a44cb65d2540e8b78c4ff56d240a1d0597
> Cr-Commit-Position: refs/heads/master@{#362532}

TBR=pkasting@chromium.org,mpearson@chromium.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=574181
==========

[commit-bot: I haz the power, 2016-01-05 20:12:47]

Committed patchset #1 (id:1)

[commit-bot: I haz the power, 2016-01-05 20:13:39]

Description was changed from

==========
Revert of Change HistoryInfoMap to use unordered_map. (patchset #2 id:20001 of
https://codereview.chromium.org/1482443002/ )

Reason for revert:
This introduced a clear regression in the omnibox performance. See bug 574181.

Original issue's description:
> Change HistoryInfoMap to use unordered_map.
> 
> This change speeds up the omnibox by 5-10% in my (limited) tests, using a
large profile and checking the value of the Omnibox.CharTypedToRepaintLatency
histogram.
> 
> This was also validated when profiling, where the find method was the top
function where CPU cycles were consumed. After switching to a hash map, it was a
non factor.
> 
> Note: As far as I can tell, the order is not important, so this change doesn't
alter expected behavior.
> 
> BUG=574181
> 
> Committed: https://crrev.com/6ddf83a44cb65d2540e8b78c4ff56d240a1d0597
> Cr-Commit-Position: refs/heads/master@{#362532}

TBR=pkasting@chromium.org,mpearson@chromium.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=574181
==========

to

==========
Revert of Change HistoryInfoMap to use unordered_map. (patchset #2 id:20001 of
https://codereview.chromium.org/1482443002/ )

Reason for revert:
This introduced a clear regression in the omnibox performance. See bug 574181.

Original issue's description:
> Change HistoryInfoMap to use unordered_map.
>
> This change speeds up the omnibox by 5-10% in my (limited) tests, using a
large profile and checking the value of the Omnibox.CharTypedToRepaintLatency
histogram.
>
> This was also validated when profiling, where the find method was the top
function where CPU cycles were consumed. After switching to a hash map, it was a
non factor.
>
> Note: As far as I can tell, the order is not important, so this change doesn't
alter expected behavior.
>
> BUG=574181
>
> Committed: https://crrev.com/6ddf83a44cb65d2540e8b78c4ff56d240a1d0597
> Cr-Commit-Position: refs/heads/master@{#362532}

TBR=pkasting@chromium.org,mpearson@chromium.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=574181

Committed: https://crrev.com/554207590d8fb848bd21997a23093efc35ce1dd9
Cr-Commit-Position: refs/heads/master@{#367625}
==========

[commit-bot: I haz the power, 2016-01-05 20:13:40]

Patchset 1 (id:??) landed as
https://crrev.com/554207590d8fb848bd21997a23093efc35ce1dd9
Cr-Commit-Position: refs/heads/master@{#367625}

[Peter Kasting, 2016-01-05 20:19:45]

LGTM