Revert of [css-grid] Avoid the HashMap for the override containing block size

Revert of [css-grid] Avoid the HashMap for the override containing block size (patchset #1 id:1 of https://codereview.chromium.org/2176633002/ )

Reason for revert:
Speculative Revert for Win10 Failing Tests:
fast/text/emphasis-combined-text.html
svg/W3C-SVG-1.1/filters-composite-02-b.svg

Original issue's description:
> [css-grid] Avoid the HashMap for the override containing block size
>
> The change made in r406557 caused a 12% regression on the grid layout
> performance tests. The root cause is that as part of the changes
> required to implement the orthogonal flows support we decided to use
> physical directions to determine whether a grid item overflows its
> grid area or not. Since we store the logical sizes in a HashMap, we
> implemented some utility methods that, based on the grid container's
> writing mode, retrieve the stored width or height from the HashMap.
>
> Since the previous logic was reusing the already computed logical
> value, the change introduced a considerable performance regression.
>
> This patch address the issue by removing the utility methods and
> determine the physical sizes directly inside the layout logic.
>
> BUG=630448
>
> Committed: https://crrev.com/2cf9f05031414616d3d1eb1e5ffa5013e396d61b
> Cr-Commit-Position: refs/heads/master@{#407762}

TBR=svillar@igalia.com,cbiesinger@chromium.org,rego@igalia.com,jfernandez@igalia.com
# Skipping CQ checks because original CL landed less than 1 days ago.
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=630448
Committed: https://crrev.com/04a768e5991c6e4d0967320e8b563930769439c9
Cr-Commit-Position: refs/heads/master@{#407972}

[robliao, 2016-07-26 23:36:43]

The CQ bit was checked by robliao@chromium.org

[robliao, 2016-07-26 23:36:44]

Created Revert of [css-grid] Avoid the HashMap for the override containing block
size

[commit-bot: I haz the power, 2016-07-26 23:37:23]

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

[commit-bot: I haz the power, 2016-07-26 23:38:58]

Description was changed from

==========
Revert of [css-grid] Avoid the HashMap for the override containing block size
(patchset #1 id:1 of https://codereview.chromium.org/2176633002/ )

Reason for revert:
Speculative Revert for Win10 Failing Tests:
fast/text/emphasis-combined-text.html
svg/W3C-SVG-1.1/filters-composite-02-b.svg

Original issue's description:
> [css-grid] Avoid the HashMap for the override containing block size
> 
> The change made in r406557 caused a 12% regression on the grid layout
> performance tests. The root cause is that as part of the changes
> required to implement the orthogonal flows support we decided to use
> physical directions to determine whether a grid item overflows its
> grid area or not. Since we store the logical sizes in a HashMap, we
> implemented some utility methods that, based on the grid container's
> writing mode, retrieve the stored width or height from the HashMap.
> 
> Since the previous logic was reusing the already computed logical
> value, the change introduced a considerable performance regression.
> 
> This patch address the issue by removing the utility methods and
> determine the physical sizes directly inside the layout logic.
> 
> BUG=630448
> 
> Committed: https://crrev.com/2cf9f05031414616d3d1eb1e5ffa5013e396d61b
> Cr-Commit-Position: refs/heads/master@{#407762}

TBR=svillar@igalia.com,cbiesinger@chromium.org,rego@igalia.com,jfernandez@iga...
# Skipping CQ checks because original CL landed less than 1 days ago.
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=630448
==========

to

==========
Revert of [css-grid] Avoid the HashMap for the override containing block size
(patchset #1 id:1 of https://codereview.chromium.org/2176633002/ )

Reason for revert:
Speculative Revert for Win10 Failing Tests:
fast/text/emphasis-combined-text.html
svg/W3C-SVG-1.1/filters-composite-02-b.svg

Original issue's description:
> [css-grid] Avoid the HashMap for the override containing block size
> 
> The change made in r406557 caused a 12% regression on the grid layout
> performance tests. The root cause is that as part of the changes
> required to implement the orthogonal flows support we decided to use
> physical directions to determine whether a grid item overflows its
> grid area or not. Since we store the logical sizes in a HashMap, we
> implemented some utility methods that, based on the grid container's
> writing mode, retrieve the stored width or height from the HashMap.
> 
> Since the previous logic was reusing the already computed logical
> value, the change introduced a considerable performance regression.
> 
> This patch address the issue by removing the utility methods and
> determine the physical sizes directly inside the layout logic.
> 
> BUG=630448
> 
> Committed: https://crrev.com/2cf9f05031414616d3d1eb1e5ffa5013e396d61b
> Cr-Commit-Position: refs/heads/master@{#407762}

TBR=svillar@igalia.com,cbiesinger@chromium.org,rego@igalia.com,jfernandez@iga...
# Skipping CQ checks because original CL landed less than 1 days ago.
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=630448
==========

[commit-bot: I haz the power, 2016-07-26 23:38:59]

Committed patchset #1 (id:1)

[commit-bot: I haz the power, 2016-07-26 23:41:39]

Description was changed from

==========
Revert of [css-grid] Avoid the HashMap for the override containing block size
(patchset #1 id:1 of https://codereview.chromium.org/2176633002/ )

Reason for revert:
Speculative Revert for Win10 Failing Tests:
fast/text/emphasis-combined-text.html
svg/W3C-SVG-1.1/filters-composite-02-b.svg

Original issue's description:
> [css-grid] Avoid the HashMap for the override containing block size
> 
> The change made in r406557 caused a 12% regression on the grid layout
> performance tests. The root cause is that as part of the changes
> required to implement the orthogonal flows support we decided to use
> physical directions to determine whether a grid item overflows its
> grid area or not. Since we store the logical sizes in a HashMap, we
> implemented some utility methods that, based on the grid container's
> writing mode, retrieve the stored width or height from the HashMap.
> 
> Since the previous logic was reusing the already computed logical
> value, the change introduced a considerable performance regression.
> 
> This patch address the issue by removing the utility methods and
> determine the physical sizes directly inside the layout logic.
> 
> BUG=630448
> 
> Committed: https://crrev.com/2cf9f05031414616d3d1eb1e5ffa5013e396d61b
> Cr-Commit-Position: refs/heads/master@{#407762}

TBR=svillar@igalia.com,cbiesinger@chromium.org,rego@igalia.com,jfernandez@iga...
# Skipping CQ checks because original CL landed less than 1 days ago.
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=630448
==========

to

==========
Revert of [css-grid] Avoid the HashMap for the override containing block size
(patchset #1 id:1 of https://codereview.chromium.org/2176633002/ )

Reason for revert:
Speculative Revert for Win10 Failing Tests:
fast/text/emphasis-combined-text.html
svg/W3C-SVG-1.1/filters-composite-02-b.svg

Original issue's description:
> [css-grid] Avoid the HashMap for the override containing block size
>
> The change made in r406557 caused a 12% regression on the grid layout
> performance tests. The root cause is that as part of the changes
> required to implement the orthogonal flows support we decided to use
> physical directions to determine whether a grid item overflows its
> grid area or not. Since we store the logical sizes in a HashMap, we
> implemented some utility methods that, based on the grid container's
> writing mode, retrieve the stored width or height from the HashMap.
>
> Since the previous logic was reusing the already computed logical
> value, the change introduced a considerable performance regression.
>
> This patch address the issue by removing the utility methods and
> determine the physical sizes directly inside the layout logic.
>
> BUG=630448
>
> Committed: https://crrev.com/2cf9f05031414616d3d1eb1e5ffa5013e396d61b
> Cr-Commit-Position: refs/heads/master@{#407762}

TBR=svillar@igalia.com,cbiesinger@chromium.org,rego@igalia.com,jfernandez@iga...
# Skipping CQ checks because original CL landed less than 1 days ago.
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=630448

Committed: https://crrev.com/04a768e5991c6e4d0967320e8b563930769439c9
Cr-Commit-Position: refs/heads/master@{#407972}
==========

[commit-bot: I haz the power, 2016-07-26 23:41:40]

Patchset 1 (id:??) landed as
https://crrev.com/04a768e5991c6e4d0967320e8b563930769439c9
Cr-Commit-Position: refs/heads/master@{#407972}

[jfernandez, 2016-07-27 11:42:43]

Are you really sure patch for issue #2176633002 was the cause of the layout
tests failures ?

[robliao, 2016-07-27 17:14:58]

On 2016/07/27 11:42:43, jfernandez wrote:
> Are you really sure patch for issue #2176633002 was the cause of the layout
> tests failures ?

No, hence the speculative revert. This was the only substantive change when
those tests started failing. I'll be relanding since they're still failing.