Remove duplicated headers from core/

Remove duplicated headers from core/ directory where the implementation file includes the same dependencies as the corresponding primary header.

BUG=317231
Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=162772

[lgombos, 2013-11-10 18:07:51]

I realize this is a large amount of change for one CL, but the changes are all
trivial. The CL is only changing .cpp files (and no .h file) so the impact of
the changes are localized to that .cpp file. The changes are done by a script, I
scanned it trough manually and try bots are happy.

Having said that I am happy to break it up into smaller CLs if that is
preferred.

[Stephen Chennney, 2013-11-11 16:16:01]

On 2013/11/10 18:07:51, lgombos wrote:
> I realize this is a large amount of change for one CL, but the changes are all
> trivial. The CL is only changing .cpp files (and no .h file) so the impact of
> the changes are localized to that .cpp file. The changes are done by a script,
I
> scanned it trough manually and try bots are happy.
> 
> Having said that I am happy to break it up into smaller CLs if that is
> preferred.

How many of these could instead replace the include in the header with a
pre-declaration of a class or struct? That is, how many of the headers just use
pointers or references to classes in the include files? We would prefer that
option, as it reduces compile times in general and helps all files that include
the header, not just the implementation.

[Inactive, 2013-11-11 17:53:30]

On 2013/11/11 16:16:01, Stephen Chenney wrote:
> On 2013/11/10 18:07:51, lgombos wrote:
> > I realize this is a large amount of change for one CL, but the changes are
all
> > trivial. The CL is only changing .cpp files (and no .h file) so the impact
of
> > the changes are localized to that .cpp file. The changes are done by a
script,
> I
> > scanned it trough manually and try bots are happy.
> > 
> > Having said that I am happy to break it up into smaller CLs if that is
> > preferred.
> 
> How many of these could instead replace the include in the header with a
> pre-declaration of a class or struct? That is, how many of the headers just
use
> pointers or references to classes in the include files? We would prefer that
> option, as it reduces compile times in general and helps all files that
include
> the header, not just the implementation.

Yes, it is likely some includes can be removed from the header and replaced by
forward declarations. This would indeed be the ideal situation and would benefit
build time.
However, this needs to be done manually and carefully.

Also, this sort of effort does not only apply to includes that are duplicated in
the cpp and the h files. It applies to *any* include in the headers.

While this CL does not have a huge impact, I believe getting rid of duplicate
headers is a good thing (It is also a low-hanging fruit since it can be done
automatically).
My personal feeling about this is "why not?". I understand we want to get rid of
includes of the includes in the headers as much as possible but this is sort of
an orthogonal (and more complicated) problem.

This work was already done and only removes duplicate code so I would tend to
let it land but this is just my 2 cents. I definitely don't feel strongly either
way.

[eseidel, 2013-11-11 18:26:14]

rslgtm.

[commit-bot: I haz the power, 2013-11-12 22:18:05]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/l.gombos@samsung.com/64643009/1

[commit-bot: I haz the power, 2013-11-12 22:20:25]

Failed to apply patch for Source/core/rendering/svg/RenderSVGModelObject.cpp:
While running patch -p1 --forward --force --no-backup-if-mismatch;
  patching file Source/core/rendering/svg/RenderSVGModelObject.cpp
  Hunk #1 FAILED at 35.
  1 out of 1 hunk FAILED -- saving rejects to file
Source/core/rendering/svg/RenderSVGModelObject.cpp.rej

Patch:       Source/core/rendering/svg/RenderSVGModelObject.cpp
Index: Source/core/rendering/svg/RenderSVGModelObject.cpp
diff --git a/Source/core/rendering/svg/RenderSVGModelObject.cpp
b/Source/core/rendering/svg/RenderSVGModelObject.cpp
index
56312a7af54146da356bc2c3051541e8fdc0e858..adaf130b1f5ad52a63621c9565a30e19f8a1cac9
100644
--- a/Source/core/rendering/svg/RenderSVGModelObject.cpp
+++ b/Source/core/rendering/svg/RenderSVGModelObject.cpp
@@ -35,7 +35,6 @@
 #include "SVGNames.h"
 #include "core/rendering/svg/RenderSVGRoot.h"
 #include "core/rendering/svg/SVGResourcesCache.h"
-#include "core/svg/SVGElement.h"
 
 namespace WebCore {

[commit-bot: I haz the power, 2013-11-27 16:39:58]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/l.gombos@samsung.com/64643009/130001

[atreat, 2013-11-27 17:13:31]

On 2013/11/11 17:53:30, Chris Dumez wrote:
> Yes, it is likely some includes can be removed from the header and replaced by
> forward declarations. This would indeed be the ideal situation and would
benefit
> build time.
> However, this needs to be done manually and carefully.

BTW, this is precisely what clang plugins enable you to do.  The compiler will
know what declarations are necessary for the header file.  I'm just saying this
could be written and automated with a clang plugin.  Would be a good project
IMO.

[commit-bot: I haz the power, 2013-11-27 17:37:13]

Change committed as 162772