Create Source/modules/accessibility/ and move most of core/accessibility/* into it

Create Source/modules/accessibility/ and move most of core/accessibility/* into it

In an attempt to modularize accessibility, we are breaking dependencies and moving files to Source/modules/. Some dependencies were simply dead code (such as RenderMenuList depending on AXMenuList). Others were changed to from #include statements to forward references (such as AXObject in AXObjectCache.h). This last change had some unfortunate side effects, because some unrelated files were depending on non-accessibility-related things like PositionWithAffinity, which is included from AXObject.h and relied on by files that only include AXObjectCache.h. Since this CL breaks that include chain, seemingly unrelated files (such as RenderText.cpp) needed additional #includes in order not to break the compile.

BUG=428285
TBR=darin@chromium.org

Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=185213

[pilgrim_google, 2014-11-10 17:58:53]

pilgrim@chromium.org changed reviewers:
+ dmazzoni@chromium.org

[pilgrim_google, 2014-11-10 17:59:10]

pilgrim@chromium.org changed reviewers:
+ aboxhall@chromium.org

[dmazzoni, 2014-11-11 06:22:00]

dmazzoni@chromium.org changed reviewers:
+ dglazkov@chromium.org

[dmazzoni, 2014-11-11 06:22:01]

+dglazkov

Overall lgtm, seems like a step in the right direction. Thanks!

Please add a bug number to the description, and since this is such a large
change, please summarize any changes you made - especially anything that wasn't
mechanical like updating include paths.

We need to figure out where to move core/accessibility/AXObjectCache.h - maybe
put it in core/dom since it's owned by Document currently?

Also please be careful about merges when landing - there are several small
changes to core/accessibility that either just landed or are in progress right
now.

[mstensho (USE GERRIT), 2014-11-11 08:23:47]

mstensho@opera.com changed reviewers:
+ mstensho@opera.com

[mstensho (USE GERRIT), 2014-11-11 08:23:49]

https://codereview.chromium.org/713933002/diff/1/Source/core/rendering/Render...
File Source/core/rendering/RenderObject.h (right):

https://codereview.chromium.org/713933002/diff/1/Source/core/rendering/Render...
Source/core/rendering/RenderObject.h:33: #include
"core/editing/VisiblePosition.h"
Why is this necessary now? You didn't add any new code to RenderObject.h and
VisiblePosition isn't used here.

https://codereview.chromium.org/713933002/diff/1/Source/modules/accessibility...
File Source/modules/accessibility/gyp-stuff.txt (right):

https://codereview.chromium.org/713933002/diff/1/Source/modules/accessibility...
Source/modules/accessibility/gyp-stuff.txt:1: 'accessibility/AXARIAGrid.cpp',
What's this file?

[pilgrim_google, 2014-11-11 15:10:47]

On 2014/11/11 08:23:49, mstensho wrote:
>
https://codereview.chromium.org/713933002/diff/1/Source/core/rendering/Render...
> File Source/core/rendering/RenderObject.h (right):
> 
>
https://codereview.chromium.org/713933002/diff/1/Source/core/rendering/Render...
> Source/core/rendering/RenderObject.h:33: #include
> "core/editing/VisiblePosition.h"
> Why is this necessary now? You didn't add any new code to RenderObject.h and
> VisiblePosition isn't used here.
> 

AXObjectCache.h no longer includes AXObject.h, and there were a few files
elsewhere in the source tree that included AXObjectCache.h and relied on things
that AXObject.h relied on. I believe this particular dependency was about
PositionWithAffinity.

>
https://codereview.chromium.org/713933002/diff/1/Source/modules/accessibility...
> File Source/modules/accessibility/gyp-stuff.txt (right):
> 
>
https://codereview.chromium.org/713933002/diff/1/Source/modules/accessibility...
> Source/modules/accessibility/gyp-stuff.txt:1: 'accessibility/AXARIAGrid.cpp',
> What's this file?

Oops, removing.

[mstensho (USE GERRIT), 2014-11-11 15:27:38]

On 2014/11/11 15:10:47, pilgrim_google wrote:
> AXObjectCache.h no longer includes AXObject.h, and there were a few files
> elsewhere in the source tree that included AXObjectCache.h and relied on
things
> that AXObject.h relied on. I believe this particular dependency was about
> PositionWithAffinity.

Isn't that something to worry about in those compilation units needing it,
instead of including it from RenderObject.h?

[pilgrim_google, 2014-11-11 16:02:30]

On 2014/11/11 15:27:38, mstensho wrote:
> On 2014/11/11 15:10:47, pilgrim_google wrote:
> > AXObjectCache.h no longer includes AXObject.h, and there were a few files
> > elsewhere in the source tree that included AXObjectCache.h and relied on
> things
> > that AXObject.h relied on. I believe this particular dependency was about
> > PositionWithAffinity.
> 
> Isn't that something to worry about in those compilation units needing it,
> instead of including it from RenderObject.h?

I have no opinion on that. RenderObject.h has a createPositionWithAffinity
method, so all the affected files included RenderObject.h because that's where
there were getting their PositionWithAffinity objects. It seemed logical and
solved the problem so I did it. But I don't care enough to fight over it.

[mstensho (USE GERRIT), 2014-11-11 17:55:49]

> I have no opinion on that. RenderObject.h has a createPositionWithAffinity
> method, so all the affected files included RenderObject.h because that's where
> there were getting their PositionWithAffinity objects. It seemed logical and
> solved the problem so I did it. But I don't care enough to fight over it.

PositionWithAffinity is forward-declared for that very reason, I suppose. I find
avoiding unnecessary includes important enough, especially in a file that's
included in 600-700 super-heavy compilation units.

[pilgrim_google, 2014-11-11 18:56:42]

On 2014/11/11 17:55:49, mstensho wrote:
> > I have no opinion on that. RenderObject.h has a createPositionWithAffinity
> > method, so all the affected files included RenderObject.h because that's
where
> > there were getting their PositionWithAffinity objects. It seemed logical and
> > solved the problem so I did it. But I don't care enough to fight over it.
> 
> PositionWithAffinity is forward-declared for that very reason, I suppose. I
find
> avoiding unnecessary includes important enough, especially in a file that's
> included in 600-700 super-heavy compilation units.

Ah, that makes sense. Thanks for explaining. Fixed in latest patch.

[mstensho (USE GERRIT), 2014-11-12 08:58:50]

On 2014/11/11 18:56:42, pilgrim_google wrote:
> On 2014/11/11 17:55:49, mstensho wrote:
> > > I have no opinion on that. RenderObject.h has a createPositionWithAffinity
> > > method, so all the affected files included RenderObject.h because that's
> where
> > > there were getting their PositionWithAffinity objects. It seemed logical
and
> > > solved the problem so I did it. But I don't care enough to fight over it.
> > 
> > PositionWithAffinity is forward-declared for that very reason, I suppose. I
> find
> > avoiding unnecessary includes important enough, especially in a file that's
> > included in 600-700 super-heavy compilation units.
> 
> Ah, that makes sense. Thanks for explaining. Fixed in latest patch.

Great, thanks. Then I have no further comments. :)

[pilgrim_google, 2014-11-12 14:45:28]

The CQ bit was checked by pilgrim@chromium.org

[commit-bot: I haz the power, 2014-11-12 14:46:22]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/713933002/60001

[commit-bot: I haz the power, 2014-11-12 14:53:16]

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

[commit-bot: I haz the power, 2014-11-12 14:53:17]

Try jobs failed on following builders:
  blink_presubmit on tryserver.blink
(http://build.chromium.org/p/tryserver.blink/builders/blink_presubmit/builds/1...)

[pilgrim_google, 2014-11-12 14:55:09]

The CQ bit was checked by pilgrim@chromium.org

[commit-bot: I haz the power, 2014-11-12 14:55:34]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/713933002/60001

[commit-bot: I haz the power, 2014-11-12 15:48:26]

Committed patchset #4 (id:60001) as 185213