Move border-*-width out of BorderValue and store on SurroundData in ComputedStyle instead

Move border-*-width out of BorderValue and store on SurroundData in ComputedStyle instead

This patch moves the storage of border-*-width away from BorderValue.
We instead store them on the generated SurroundData group on
ComputedStyle instead.

Since BorderValue is used by other classes such as OutlineValue and
MultiColData we have changed BorderData to store the
border-*-(style/color) information on BorderColorAndStyle which is the
BorderValue class without the border-*-width information.

The long term intention is to move all information away from
BorderColorAndStyle - it is a temporary class to store information as
we generate border-*-* iteratively.

Please note that we convert BorderColorAndStyle to a BorderValue when we
send information over to the Layout system to make the interfacing easy.
This will be changed to creating a new BorderValue from the individual
values in the long run or something along those lines
(pending discussion.)

Generated diff: https://gist.github.com/nainar/296dda6afe541b6ca6da18a2923d0ef9/revisions

BUG=628043
Review-Url: https://codereview.chromium.org/2861773004
Cr-Commit-Position: refs/heads/master@{#470272}
Committed: https://chromium.googlesource.com/chromium/src/+/5a563421457c31c1bb6cdac72c899db056f1adc7

[nainar, 2017-05-04 01:50:04]

Description was changed from

==========
Move border-*-width out of BorderValue and store on SurroundData in
ComputedStyle instead

This patch moves the storage of border-*-width away from BorderValue.
We instead soter them on the generated SurroundData group on
ComputedStyle instead.

Since BorderValue is used by other classes such as OutlineValue and
MultiColData we have changed BorderData to store the
border-*-(style/color) information on BorderColorAndStyle which is the
BorderValue class without the border-*-width information.

The long term intention is to move all information away from
BorderColorAndStyle - it is a temporary class to store information as
we generate border-*-* iteratively.

Generated diff:
https://gist.github.com/nainar/9d9df98c2123e3d8b953b8d26be273dc/revisions

BUG=628043
==========

to

==========
Move border-*-width out of BorderValue and store on SurroundData in
ComputedStyle instead

This patch moves the storage of border-*-width away from BorderValue.
We instead soter them on the generated SurroundData group on
ComputedStyle instead.

Since BorderValue is used by other classes such as OutlineValue and
MultiColData we have changed BorderData to store the
border-*-(style/color) information on BorderColorAndStyle which is the
BorderValue class without the border-*-width information.

The long term intention is to move all information away from
BorderColorAndStyle - it is a temporary class to store information as
we generate border-*-* iteratively.

Please note that we convert BorderColorAndStyle to a BorderValue when we
send information over to the Layout system to make the interfacing easy.
This will be changed to creating a new BorderValue from the individual
values in the long run or something along those lines
(pending discussion.)

Generated diff:
https://gist.github.com/nainar/9d9df98c2123e3d8b953b8d26be273dc/revisions

BUG=628043
==========

[nainar, 2017-05-04 01:51:05]

Description was changed from

==========
Move border-*-width out of BorderValue and store on SurroundData in
ComputedStyle instead

This patch moves the storage of border-*-width away from BorderValue.
We instead soter them on the generated SurroundData group on
ComputedStyle instead.

Since BorderValue is used by other classes such as OutlineValue and
MultiColData we have changed BorderData to store the
border-*-(style/color) information on BorderColorAndStyle which is the
BorderValue class without the border-*-width information.

The long term intention is to move all information away from
BorderColorAndStyle - it is a temporary class to store information as
we generate border-*-* iteratively.

Please note that we convert BorderColorAndStyle to a BorderValue when we
send information over to the Layout system to make the interfacing easy.
This will be changed to creating a new BorderValue from the individual
values in the long run or something along those lines
(pending discussion.)

Generated diff:
https://gist.github.com/nainar/9d9df98c2123e3d8b953b8d26be273dc/revisions

BUG=628043
==========

to

==========
Move border-*-width out of BorderValue and store on SurroundData in
ComputedStyle instead

This patch moves the storage of border-*-width away from BorderValue.
We instead soter them on the generated SurroundData group on
ComputedStyle instead.

Since BorderValue is used by other classes such as OutlineValue and
MultiColData we have changed BorderData to store the
border-*-(style/color) information on BorderColorAndStyle which is the
BorderValue class without the border-*-width information.

The long term intention is to move all information away from
BorderColorAndStyle - it is a temporary class to store information as
we generate border-*-* iteratively.

Please note that we convert BorderColorAndStyle to a BorderValue when we
send information over to the Layout system to make the interfacing easy.
This will be changed to creating a new BorderValue from the individual
values in the long run or something along those lines
(pending discussion.)

Generated diff:
https://gist.github.com/nainar/296dda6afe541b6ca6da18a2923d0ef9/revisions

BUG=628043
==========

[nainar, 2017-05-04 01:52:45]

The CQ bit was checked by nainar@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2017-05-04 01:53:00]

Dry run: CQ is trying da patch.

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

[nainar, 2017-05-04 03:06:55]

nainar@chromium.org changed reviewers:
+ shend@chromium.org

[nainar, 2017-05-04 03:06:56]

shend@, PTAL? Thanks!

https://codereview.chromium.org/2861773004/diff/1/third_party/WebKit/Source/b...
File
third_party/WebKit/Source/build/scripts/templates/ComputedStyleBase.cpp.tmpl
(right):

https://codereview.chromium.org/2861773004/diff/1/third_party/WebKit/Source/b...
third_party/WebKit/Source/build/scripts/templates/ComputedStyleBase.cpp.tmpl:15:
{{field.type_name}} {{field.name}};
This is cleaning up an erroneous bracket added in the generated code. This path
of the code was never executed before.

[commit-bot: I haz the power, 2017-05-04 03:19:40]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2017-05-04 03:19:40]

Dry run: Try jobs failed on following builders:
  linux_chromium_rel_ng on master.tryserver.chromium.linux (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.linux/builders/linux_chromium_...)

[nainar, 2017-05-05 06:59:50]

nainar@chromium.org changed reviewers:
- shend@chromium.org

[nainar, 2017-05-05 07:11:08]

The CQ bit was checked by nainar@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2017-05-05 07:11:24]

Dry run: CQ is trying da patch.

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

[nainar, 2017-05-05 07:16:48]

nainar@chromium.org changed reviewers:
+ shend@chromium.org

[nainar, 2017-05-05 07:16:50]

shend@, 

PTAL? 

Thanks!

https://codereview.chromium.org/2861773004/diff/40001/third_party/WebKit/Sour...
File
third_party/WebKit/Source/build/scripts/templates/ComputedStyleBase.cpp.tmpl
(right):

https://codereview.chromium.org/2861773004/diff/40001/third_party/WebKit/Sour...
third_party/WebKit/Source/build/scripts/templates/ComputedStyleBase.cpp.tmpl:15:
{{field.type_name}} {{field.name}};
Cleaning up an erroneous bracket. This code was never hit before.

https://codereview.chromium.org/2861773004/diff/40001/third_party/WebKit/Sour...
File third_party/WebKit/Source/build/scripts/templates/ComputedStyleBase.h.tmpl
(right):

https://codereview.chromium.org/2861773004/diff/40001/third_party/WebKit/Sour...
third_party/WebKit/Source/build/scripts/templates/ComputedStyleBase.h.tmpl:141:
static unsigned WidthToFixedPoint(float width) {
need a copy of this here to initialize the border-*-width correctly.

https://codereview.chromium.org/2861773004/diff/40001/third_party/WebKit/Sour...
File third_party/WebKit/Source/core/style/BorderColorAndStyle.h (right):

https://codereview.chromium.org/2861773004/diff/40001/third_party/WebKit/Sour...
third_party/WebKit/Source/core/style/BorderColorAndStyle.h:1: /*
This is a temp class. See CL description

https://codereview.chromium.org/2861773004/diff/40001/third_party/WebKit/Sour...
File third_party/WebKit/Source/core/style/BorderValue.h (left):

https://codereview.chromium.org/2861773004/diff/40001/third_party/WebKit/Sour...
third_party/WebKit/Source/core/style/BorderValue.h:35: // In order to conserve
memory, the border width uses fixed point,
Since this is needed by both this and ComputedStyleBase moved these constants to
ComputedStyleConstants let me know if you have a better suggestion.

https://codereview.chromium.org/2861773004/diff/40001/third_party/WebKit/Sour...
File third_party/WebKit/Source/core/style/ComputedStyle.h (right):

https://codereview.chromium.org/2861773004/diff/40001/third_party/WebKit/Sour...
third_party/WebKit/Source/core/style/ComputedStyle.h:2954:
SET_VAR(surround_data_, border_.top_, BorderColorAndStyle());
hmm on second thought this should probably also reset BorderTopWidth.

Best to do it in this function or in all callsites?

[commit-bot: I haz the power, 2017-05-05 09:42:00]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2017-05-05 09:42:01]

Dry run: Try jobs failed on following builders:
  linux_chromium_asan_rel_ng on master.tryserver.chromium.linux (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.linux/builders/linux_chromium_...)

[shend, 2017-05-07 23:08:31]

https://codereview.chromium.org/2861773004/diff/40001/third_party/WebKit/Sour...
File
third_party/WebKit/Source/build/scripts/templates/ComputedStyleBase.cpp.tmpl
(right):

https://codereview.chromium.org/2861773004/diff/40001/third_party/WebKit/Sour...
third_party/WebKit/Source/build/scripts/templates/ComputedStyleBase.cpp.tmpl:15:
{{field.type_name}} {{field.name}};
On 2017/05/05 at 07:16:49, nainar wrote:
> Cleaning up an erroneous bracket. This code was never hit before.

Nice catch!

https://codereview.chromium.org/2861773004/diff/40001/third_party/WebKit/Sour...
File third_party/WebKit/Source/build/scripts/templates/ComputedStyleBase.h.tmpl
(right):

https://codereview.chromium.org/2861773004/diff/40001/third_party/WebKit/Sour...
third_party/WebKit/Source/build/scripts/templates/ComputedStyleBase.h.tmpl:141:
static unsigned WidthToFixedPoint(float width) {
On 2017/05/05 at 07:16:50, nainar wrote:
> need a copy of this here to initialize the border-*-width correctly.

Can this be in BorderValue or some utilities file instead?

https://codereview.chromium.org/2861773004/diff/40001/third_party/WebKit/Sour...
File third_party/WebKit/Source/core/css/resolver/StyleResolver.cpp (right):

https://codereview.chromium.org/2861773004/diff/40001/third_party/WebKit/Sour...
third_party/WebKit/Source/core/css/resolver/StyleResolver.cpp:1968:
(cached_ua_style->border_left_width !=
nit: I would move this up to just after line 1959, since it was originally
"there".

https://codereview.chromium.org/2861773004/diff/40001/third_party/WebKit/Sour...
File third_party/WebKit/Source/core/style/BorderValue.h (left):

https://codereview.chromium.org/2861773004/diff/40001/third_party/WebKit/Sour...
third_party/WebKit/Source/core/style/BorderValue.h:35: // In order to conserve
memory, the border width uses fixed point,
On 2017/05/05 at 07:16:50, nainar wrote:
> Since this is needed by both this and ComputedStyleBase moved these constants
to ComputedStyleConstants let me know if you have a better suggestion.

If you moved WidthToFixedPoint to this file, would that fix this?

https://codereview.chromium.org/2861773004/diff/40001/third_party/WebKit/Sour...
File third_party/WebKit/Source/core/style/BorderValue.h (right):

https://codereview.chromium.org/2861773004/diff/40001/third_party/WebKit/Sour...
third_party/WebKit/Source/core/style/BorderValue.h:48: BorderValue(const
BorderColorAndStyle& data, const float width) {
nit: probably don't need the const on the 'width' (imo, pass-by-value parameters
should have const like you've done here, but the rest of the codebase doesn't
seem to follow this rule).

https://codereview.chromium.org/2861773004/diff/40001/third_party/WebKit/Sour...
File third_party/WebKit/Source/core/style/ComputedStyle.h (right):

https://codereview.chromium.org/2861773004/diff/40001/third_party/WebKit/Sour...
third_party/WebKit/Source/core/style/ComputedStyle.h:550: return
static_cast<float>(BorderRightWidthInternal()) /
The fixed point stuff feels like it should be a class. Maybe as future work, we
could make a FixedPoint class that handles this stuff in a generic way?

https://codereview.chromium.org/2861773004/diff/40001/third_party/WebKit/Sour...
third_party/WebKit/Source/core/style/ComputedStyle.h:2878: BorderValue*
border_value =
Have we measured the perf impact of this dynamic allocation?
Also, I'm not familiar with Blink GC magic, but isn't this a memory leak?
Could we just do:

BorderValue BorderLeft() const {
  return BorderValue(surround_data_->border_.Left(), BorderLeftWidth());
}

https://codereview.chromium.org/2861773004/diff/40001/third_party/WebKit/Sour...
third_party/WebKit/Source/core/style/ComputedStyle.h:2954:
SET_VAR(surround_data_, border_.top_, BorderColorAndStyle());
On 2017/05/05 at 07:16:50, nainar wrote:
> hmm on second thought this should probably also reset BorderTopWidth.
> 
> Best to do it in this function or in all callsites?

Hmm, good question. I would do it here, since BorderLeft() returns a BorderValue
(color + style + width), I would expect ResetBorderLeft() to reset all three
things.

[nainar, 2017-05-08 00:55:59]

The CQ bit was checked by nainar@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2017-05-08 00:56:08]

Dry run: CQ is trying da patch.

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

[nainar, 2017-05-08 00:59:29]

The CQ bit was checked by nainar@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2017-05-08 00:59:42]

Dry run: CQ is trying da patch.

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

[nainar, 2017-05-08 00:59:57]

shend@, 

Made the changes you asked for and answered some questions inline. PTAL? Thanks!

https://codereview.chromium.org/2861773004/diff/40001/third_party/WebKit/Sour...
File third_party/WebKit/Source/build/scripts/templates/ComputedStyleBase.h.tmpl
(right):

https://codereview.chromium.org/2861773004/diff/40001/third_party/WebKit/Sour...
third_party/WebKit/Source/build/scripts/templates/ComputedStyleBase.h.tmpl:141:
static unsigned WidthToFixedPoint(float width) {
Not unless I include BorderValue by hand. Utilities file - hmmm dont know of any
ComputedStyle utilities class...

https://codereview.chromium.org/2861773004/diff/40001/third_party/WebKit/Sour...
File third_party/WebKit/Source/core/css/resolver/StyleResolver.cpp (right):

https://codereview.chromium.org/2861773004/diff/40001/third_party/WebKit/Sour...
third_party/WebKit/Source/core/css/resolver/StyleResolver.cpp:1968:
(cached_ua_style->border_left_width !=
Done.

https://codereview.chromium.org/2861773004/diff/40001/third_party/WebKit/Sour...
File third_party/WebKit/Source/core/style/BorderValue.h (left):

https://codereview.chromium.org/2861773004/diff/40001/third_party/WebKit/Sour...
third_party/WebKit/Source/core/style/BorderValue.h:35: // In order to conserve
memory, the border width uses fixed point,
Nope. Unless we include the BorderValue file in ComputedStyleBase.

https://codereview.chromium.org/2861773004/diff/40001/third_party/WebKit/Sour...
File third_party/WebKit/Source/core/style/BorderValue.h (right):

https://codereview.chromium.org/2861773004/diff/40001/third_party/WebKit/Sour...
third_party/WebKit/Source/core/style/BorderValue.h:48: BorderValue(const
BorderColorAndStyle& data, const float width) {
Done.

https://codereview.chromium.org/2861773004/diff/40001/third_party/WebKit/Sour...
File third_party/WebKit/Source/core/style/ComputedStyle.h (right):

https://codereview.chromium.org/2861773004/diff/40001/third_party/WebKit/Sour...
third_party/WebKit/Source/core/style/ComputedStyle.h:550: return
static_cast<float>(BorderRightWidthInternal()) /
Yes. Added a todo here on me to take a look at this next.

https://codereview.chromium.org/2861773004/diff/40001/third_party/WebKit/Sour...
third_party/WebKit/Source/core/style/ComputedStyle.h:2878: BorderValue*
border_value =
Done.

https://codereview.chromium.org/2861773004/diff/40001/third_party/WebKit/Sour...
third_party/WebKit/Source/core/style/ComputedStyle.h:2954:
SET_VAR(surround_data_, border_.top_, BorderColorAndStyle());
Yup. Agreed. Done.

[commit-bot: I haz the power, 2017-05-08 02:31:40]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2017-05-08 02:31:40]

Dry run: Try jobs failed on following builders:
  linux_chromium_chromeos_ozone_rel_ng on master.tryserver.chromium.linux
(JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.linux/builders/linux_chromium_...)

[shend, 2017-05-08 04:39:02]

lgtm, although some of the return by values might have perf impact. I think it
would be minimal because of return value optimisation and stuff, but just a
heads up anyway.

https://codereview.chromium.org/2861773004/diff/40001/third_party/WebKit/Sour...
File third_party/WebKit/Source/build/scripts/templates/ComputedStyleBase.h.tmpl
(right):

https://codereview.chromium.org/2861773004/diff/40001/third_party/WebKit/Sour...
third_party/WebKit/Source/build/scripts/templates/ComputedStyleBase.h.tmpl:141:
static unsigned WidthToFixedPoint(float width) {
On 2017/05/08 at 00:59:57, nainar wrote:
> Not unless I include BorderValue by hand. Utilities file - hmmm dont know of
any ComputedStyle utilities class...

Ok, keep it as it is then. I think we can resolve this later when we move fixed
point stuff to its own class or something.

[nainar, 2017-05-08 05:08:26]

On 2017/05/08 at 04:39:02, shend wrote:
> lgtm, although some of the return by values might have perf impact. I think it
would be minimal because of return value optimisation and stuff, but just a
heads up anyway.

Yup, thanks for the headsup.

>
https://codereview.chromium.org/2861773004/diff/40001/third_party/WebKit/Sour...
> File
third_party/WebKit/Source/build/scripts/templates/ComputedStyleBase.h.tmpl
(right):
> 
>
https://codereview.chromium.org/2861773004/diff/40001/third_party/WebKit/Sour...
>
third_party/WebKit/Source/build/scripts/templates/ComputedStyleBase.h.tmpl:141:
static unsigned WidthToFixedPoint(float width) {
> On 2017/05/08 at 00:59:57, nainar wrote:
> > Not unless I include BorderValue by hand. Utilities file - hmmm dont know of
any ComputedStyle utilities class...
> 
> Ok, keep it as it is then. I think we can resolve this later when we move
fixed point stuff to its own class or something.

Sounds good.

[nainar, 2017-05-08 05:08:42]

nainar@chromium.org changed reviewers:
+ meade@chromium.org

[nainar, 2017-05-08 05:08:42]

meade@, 

PTAL? Thanks!

[meade_UTC10, 2017-05-09 03:25:06]

lgtm with nits

I assume this is the CL you were talking about in standup today where you didn't
find the pass-by-value to have a noticable performance impact?

Also you misspelled store as "soter" in the CL description...

https://codereview.chromium.org/2861773004/diff/80001/third_party/WebKit/Sour...
File third_party/WebKit/Source/core/style/BorderValue.h (left):

https://codereview.chromium.org/2861773004/diff/80001/third_party/WebKit/Sour...
third_party/WebKit/Source/core/style/BorderValue.h:46: DISALLOW_NEW();
Do you actually need to remove this? I couldn't find a usage of new
BorderValue() anywhere... Also this class isn't garbage collected or reference
counted, so we'll need to be careful with it if you do that.

Is it a future change for when you will need to later? In that case I suggest
doing it with that change for clarity.

[nainar, 2017-05-09 04:14:11]

Description was changed from

==========
Move border-*-width out of BorderValue and store on SurroundData in
ComputedStyle instead

This patch moves the storage of border-*-width away from BorderValue.
We instead soter them on the generated SurroundData group on
ComputedStyle instead.

Since BorderValue is used by other classes such as OutlineValue and
MultiColData we have changed BorderData to store the
border-*-(style/color) information on BorderColorAndStyle which is the
BorderValue class without the border-*-width information.

The long term intention is to move all information away from
BorderColorAndStyle - it is a temporary class to store information as
we generate border-*-* iteratively.

Please note that we convert BorderColorAndStyle to a BorderValue when we
send information over to the Layout system to make the interfacing easy.
This will be changed to creating a new BorderValue from the individual
values in the long run or something along those lines
(pending discussion.)

Generated diff:
https://gist.github.com/nainar/296dda6afe541b6ca6da18a2923d0ef9/revisions

BUG=628043
==========

to

==========
Move border-*-width out of BorderValue and store on SurroundData in
ComputedStyle instead

This patch moves the storage of border-*-width away from BorderValue.
We instead store them on the generated SurroundData group on
ComputedStyle instead.

Since BorderValue is used by other classes such as OutlineValue and
MultiColData we have changed BorderData to store the
border-*-(style/color) information on BorderColorAndStyle which is the
BorderValue class without the border-*-width information.

The long term intention is to move all information away from
BorderColorAndStyle - it is a temporary class to store information as
we generate border-*-* iteratively.

Please note that we convert BorderColorAndStyle to a BorderValue when we
send information over to the Layout system to make the interfacing easy.
This will be changed to creating a new BorderValue from the individual
values in the long run or something along those lines
(pending discussion.)

Generated diff:
https://gist.github.com/nainar/296dda6afe541b6ca6da18a2923d0ef9/revisions

BUG=628043
==========

[nainar, 2017-05-09 05:45:39]

The CQ bit was checked by nainar@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2017-05-09 05:46:08]

Dry run: CQ is trying da patch.

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

[nainar, 2017-05-09 05:48:11]

meade@, 

Changed the misspelling in the CL description. Thank you for that. And yup this
is the CL I mentioned.

https://codereview.chromium.org/2861773004/diff/80001/third_party/WebKit/Sour...
File third_party/WebKit/Source/core/style/BorderValue.h (left):

https://codereview.chromium.org/2861773004/diff/80001/third_party/WebKit/Sour...
third_party/WebKit/Source/core/style/BorderValue.h:46: DISALLOW_NEW();
On 2017/05/09 at 03:25:06, meade_UTC10 wrote:
> Do you actually need to remove this? I couldn't find a usage of new
BorderValue() anywhere... Also this class isn't garbage collected or reference
counted, so we'll need to be careful with it if you do that.
> 
> Is it a future change for when you will need to later? In that case I suggest
doing it with that change for clarity.

This was an artefact from an old change. Removed this.

[commit-bot: I haz the power, 2017-05-09 09:48:24]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2017-05-09 09:48:25]

Dry run: This issue passed the CQ dry run.

[nainar, 2017-05-09 09:58:35]

The CQ bit was checked by nainar@chromium.org

[nainar, 2017-05-09 09:58:36]

The patchset sent to the CQ was uploaded after l-g-t-m from meade@chromium.org,
shend@chromium.org
Link to the patchset: https://codereview.chromium.org/2861773004/#ps100001
(title: "meade@'s suggestion")

[commit-bot: I haz the power, 2017-05-09 09:59:03]

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, 2017-05-09 10:04:51]

CQ is committing da patch.

Bot data: {"patchset_id": 100001, "attempt_start_ts": 1494323915764600,
"parent_rev": "d6dd415275e2ddf1b1e816de4a9f2b23d4236339", "commit_rev":
"5a563421457c31c1bb6cdac72c899db056f1adc7"}

[commit-bot: I haz the power, 2017-05-09 10:05:06]

Description was changed from

==========
Move border-*-width out of BorderValue and store on SurroundData in
ComputedStyle instead

This patch moves the storage of border-*-width away from BorderValue.
We instead store them on the generated SurroundData group on
ComputedStyle instead.

Since BorderValue is used by other classes such as OutlineValue and
MultiColData we have changed BorderData to store the
border-*-(style/color) information on BorderColorAndStyle which is the
BorderValue class without the border-*-width information.

The long term intention is to move all information away from
BorderColorAndStyle - it is a temporary class to store information as
we generate border-*-* iteratively.

Please note that we convert BorderColorAndStyle to a BorderValue when we
send information over to the Layout system to make the interfacing easy.
This will be changed to creating a new BorderValue from the individual
values in the long run or something along those lines
(pending discussion.)

Generated diff:
https://gist.github.com/nainar/296dda6afe541b6ca6da18a2923d0ef9/revisions

BUG=628043
==========

to

==========
Move border-*-width out of BorderValue and store on SurroundData in
ComputedStyle instead

This patch moves the storage of border-*-width away from BorderValue.
We instead store them on the generated SurroundData group on
ComputedStyle instead.

Since BorderValue is used by other classes such as OutlineValue and
MultiColData we have changed BorderData to store the
border-*-(style/color) information on BorderColorAndStyle which is the
BorderValue class without the border-*-width information.

The long term intention is to move all information away from
BorderColorAndStyle - it is a temporary class to store information as
we generate border-*-* iteratively.

Please note that we convert BorderColorAndStyle to a BorderValue when we
send information over to the Layout system to make the interfacing easy.
This will be changed to creating a new BorderValue from the individual
values in the long run or something along those lines
(pending discussion.)

Generated diff:
https://gist.github.com/nainar/296dda6afe541b6ca6da18a2923d0ef9/revisions

BUG=628043

Review-Url: https://codereview.chromium.org/2861773004
Cr-Commit-Position: refs/heads/master@{#470272}
Committed:
https://chromium.googlesource.com/chromium/src/+/5a563421457c31c1bb6cdac72c89...
==========

[commit-bot: I haz the power, 2017-05-09 10:05:07]

Committed patchset #6 (id:100001) as
https://chromium.googlesource.com/chromium/src/+/5a563421457c31c1bb6cdac72c89...