Factor out common BorderImageLength code in CSSToStyleMap

Factor out common BorderImageLength code in CSSToStyleMap

CSSToStyleMap::mapNinePieceImageQuad had a lot of repetitive code for
BorderImageLength. Factor it out to a static function.

Afterwards, the default constructor for BorderImageLengthBox gets used
less so make the code more explicit by removing the default
constructor entirely. Instead add a constructor that initializes the
members to a specific value.

Since this revealed one error in a code comment (the initial value for
-webkit-mask-box-image-width is 'auto' not zero) fix that as well.

BUG=322014
Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=162646

[davve, 2013-11-21 15:48:37]

Here comes a clean-up patch after my battle with Length::Relative.

Julien, feel free to suggest other reviewers if you get tired of my border-image
patches.

[Julien - ping for review, 2013-11-21 21:24:02]

The code change is great, just a comment below.

> Julien, feel free to suggest other reviewers if you get tired of my
border-image patches.

Me, getting tired of border-image patches, nevar!

From the description:
> the initial value for -webkit-mask-box-image-width is 'auto' not zero

We should add some testing for that as it seems this isn't tested anywhere.

[davve, 2013-11-22 09:13:32]

On 2013/11/21 21:24:02, Julien Chaffraix - AEDT wrote:
> From the description:
> > the initial value for -webkit-mask-box-image-width is 'auto' not zero
> 
> We should add some testing for that as it seems this isn't tested anywhere.

Done.

[Julien - ping for review, 2013-11-24 22:53:01]

lgtm with comments.

https://codereview.chromium.org/81123002/diff/80001/LayoutTests/fast/backgrou...
File LayoutTests/fast/backgrounds/mask-box-image-width.html (right):

https://codereview.chromium.org/81123002/diff/80001/LayoutTests/fast/backgrou...
LayoutTests/fast/backgrounds/mask-box-image-width.html:15: .test *:nth-child(2)
{
Let's avoid nth-child selector, which makes the test less readable, and use
explicit classes.

https://codereview.chromium.org/81123002/diff/80001/LayoutTests/fast/backgrou...
LayoutTests/fast/backgrounds/mask-box-image-width.html:25: }
If we are going to do some testing, how about percentage? CSSValue(s) that are
not auto?

https://codereview.chromium.org/81123002/diff/80001/LayoutTests/fast/backgrou...
LayoutTests/fast/backgrounds/mask-box-image-width.html:36: description("Test
getComputedStyle().webkitMaskBoxImageWidth");
That's a bit tense for a description IMO (it is accurate though).

How about "This test checks that -webkit-mask-box-image-width is correctly
parsed and returned from getComputedStyle()." or something equivalent?

https://codereview.chromium.org/81123002/diff/80001/Source/core/css/CSSToStyl...
File Source/core/css/CSSToStyleMap.cpp (right):

https://codereview.chromium.org/81123002/diff/80001/Source/core/css/CSSToStyl...
Source/core/css/CSSToStyleMap.cpp:613: static BorderImageLength
toBorderImageLength(CSSPrimitiveValue& value, const RenderStyle* currStyle,
const RenderStyle* rootStyle, float multiplier)
s/currStyle/currentStyle/ (3 characters isn't much of a saving IMO)

[davve, 2013-11-25 22:15:16]

https://codereview.chromium.org/81123002/diff/80001/LayoutTests/fast/backgrou...
File LayoutTests/fast/backgrounds/mask-box-image-width.html (right):

https://codereview.chromium.org/81123002/diff/80001/LayoutTests/fast/backgrou...
LayoutTests/fast/backgrounds/mask-box-image-width.html:15: .test *:nth-child(2)
{
On 2013/11/24 22:53:01, Julien Chaffraix - AEDT wrote:
> Let's avoid nth-child selector, which makes the test less readable, and use
> explicit classes.

In combination with the next issue (about adding more tests), this doesn't scale
well anyway so I'll restructure this and set the inline style on the fly from
script instead.

https://codereview.chromium.org/81123002/diff/80001/LayoutTests/fast/backgrou...
LayoutTests/fast/backgrounds/mask-box-image-width.html:25: }
On 2013/11/24 22:53:01, Julien Chaffraix - AEDT wrote:
> If we are going to do some testing, how about percentage? CSSValue(s) that are
> not auto?

Done.

https://codereview.chromium.org/81123002/diff/80001/LayoutTests/fast/backgrou...
LayoutTests/fast/backgrounds/mask-box-image-width.html:36: description("Test
getComputedStyle().webkitMaskBoxImageWidth");
On 2013/11/24 22:53:01, Julien Chaffraix - AEDT wrote:
> That's a bit tense for a description IMO (it is accurate though).
> 
> How about "This test checks that -webkit-mask-box-image-width is correctly
> parsed and returned from getComputedStyle()." or something equivalent?

Done.

https://codereview.chromium.org/81123002/diff/80001/Source/core/css/CSSToStyl...
File Source/core/css/CSSToStyleMap.cpp (right):

https://codereview.chromium.org/81123002/diff/80001/Source/core/css/CSSToStyl...
Source/core/css/CSSToStyleMap.cpp:613: static BorderImageLength
toBorderImageLength(CSSPrimitiveValue& value, const RenderStyle* currStyle,
const RenderStyle* rootStyle, float multiplier)
On 2013/11/24 22:53:01, Julien Chaffraix - AEDT wrote:
> s/currStyle/currentStyle/ (3 characters isn't much of a saving IMO) 

Done.

[commit-bot: I haz the power, 2013-11-25 22:16:14]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/davve@opera.com/81123002/170001

[commit-bot: I haz the power, 2013-11-26 00:36:43]

Change committed as 162646