Split webkit_storage target into webkit_storage_{browser,common,renderer}

Split webkit_storage target into webkit_storage_{browser,common,renderer}

This does:
- Remove all webkit/browser/foo/*.gypi, webkit/common/foo/*.gypi for storage
- Add three top-level gyp's in webkit/: storage_{browser,common,renderer}.gyp

storage_common and storage_renderer might be able to be subsumed in
bigger targets for common|renderer, but believe we can do that later
when necessary.

BUG=239109
TBR=mnaganov,thestig,avi
NOTRY=true

Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=205463

[kinuko, 2013-06-06 07:39:59]

Ok at least this now compiles. Can you take a look?

[kinuko, 2013-06-06 11:42:11]

On 2013/06/06 07:39:59, kinuko wrote:
> Ok at least this now compiles. Can you take a look?

Hmm, how come win cannot compile while win_rel does?  Digging...

[jamesr, 2013-06-07 23:54:34]

On 2013/06/06 11:42:11, kinuko wrote:
> On 2013/06/06 07:39:59, kinuko wrote:
> > Ok at least this now compiles. Can you take a look?
> 
> Hmm, how come win cannot compile while win_rel does?  Digging...

That normally means a component issue - win_rel uses the static build, win uses
the component build.

[michaeln, 2013-06-08 02:40:06]

I don't grok gyp and gpyi hardly at all. I've got a start at merging webkit_base
into webkit_common.
https://codereview.chromium.org/16189015/

[kinuko, 2013-06-10 22:02:13]

Ok, win build has become good now.  (I was using the same WEBKIT_STORAGE_EXPORT
in common and browser|renderer which made dllexport and dllimport screwed up).
Can you take a look?  Thanks,

[jamesr, 2013-06-10 22:07:21]

lgtm

using the same EXPORT macro for the browser/ and renderer/ code will produce
inconsistent declspec errors if anything #includes one from the other, but
nobody should be doing that anyway.

https://codereview.chromium.org/15716007/diff/105001/webkit/common/fileapi/fi...
File webkit/common/fileapi/file_system_util.cc (right):

https://codereview.chromium.org/15716007/diff/105001/webkit/common/fileapi/fi...
webkit/common/fileapi/file_system_util.cc:19: const char* kPersistentDir =
"/persistent";
this seems like a slight change for the worse - if you want to change the [] to
* then you should add another level of 'const', but imo the [] version is a
shorter way of expressing the level of const-ness

[kinuko, 2013-06-10 23:27:41]

Thanks!

On 2013/06/10 22:07:21, jamesr wrote:
> lgtm
> 
> using the same EXPORT macro for the browser/ and renderer/ code will produce
> inconsistent declspec errors if anything #includes one from the other, but
> nobody should be doing that anyway.

Yes, I have a TODO there and I'm making follow-up changes.

>
https://codereview.chromium.org/15716007/diff/105001/webkit/common/fileapi/fi...
> File webkit/common/fileapi/file_system_util.cc (right):
> 
>
https://codereview.chromium.org/15716007/diff/105001/webkit/common/fileapi/fi...
> webkit/common/fileapi/file_system_util.cc:19: const char* kPersistentDir =
> "/persistent";
> this seems like a slight change for the worse - if you want to change the []
to
> * then you should add another level of 'const', but imo the [] version is a
> shorter way of expressing the level of const-ness

Oops, this wasn't intended.  I'll revert them.

[kinuko, 2013-06-11 05:23:00]

+mnaganov for android_webview/ changes
+thestig for chrome/ changes
+avi for content/ changes

PTL (will likely land with TBR).
Basically this change should only reduce unnecessary dependencies.

[commit-bot: I haz the power, 2013-06-11 07:35:53]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/kinuko@chromium.org/15716007/124001

[commit-bot: I haz the power, 2013-06-11 07:37:28]

Change committed as 205463

[Nico, 2013-10-23 18:56:46]

The new storage_ gyp files in this patch need to include 

+  'variables': {
+    'chromium_code': 1,
+  },

at the top, else they are built with warnings disabled.