Remove the pointer from StringImpl

Remove the pointer from StringImpl

This CL removes the data pointer from StringImpl, reducing
sizeof(StringImpl) from 24 bytes to 12 bytes.  We now always
store the character data after the StringImpl object itself,
which means we don't need a pointer to keep track of it.

This CL saves 120 kB of memory on typical web pages (which
have ~10k strings).

R=eseidel
BUG=262320
Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=155309

Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=155404

Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=155464

[eseidel, 2013-08-01 00:38:16]

lgtm

[commit-bot: I haz the power, 2013-08-01 00:38:37]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/abarth@chromium.org/21437003/4001

[commit-bot: I haz the power, 2013-08-01 06:19:12]

Change committed as 155309

[abarth-chromium, 2013-08-01 16:47:31]

ASSERTION FAILED: hasHash()
Backtrace:
	WTF::StringImpl::assertHashIsCorrect [0x0023DC3A+58]
(c:\b\build\slave\webkit-win-latest-dbg\build\src\third_party\webkit\source\wtf\text\stringimpl.h:474)
	WebCore::HTMLNames::init [0x1128E459+169]
(c:\b\build\slave\webkit-win-latest-dbg\build\src\out\debug\gen\blink\htmlnames.cpp:6526)
	WebCore::init [0x11AE07EC+44]
(c:\b\build\slave\webkit-win-latest-dbg\build\src\third_party\webkit\source\core\init.cpp:58)
	WebKit::initializeWithoutV8 [0x10243A05+197]
(c:\b\build\slave\webkit-win-latest-dbg\build\src\third_party\webkit\source\web\webkit.cpp:154)
	WebKit::initialize [0x10243495+37]
(c:\b\build\slave\webkit-win-latest-dbg\build\src\third_party\webkit\source\web\webkit.cpp:102)
	TestWebKitPlatformSupport::TestWebKitPlatformSupport [0x00408B35+437]
(c:\b\build\slave\webkit-win-latest-dbg\build\src\webkit\support\test_webkit_platform_support.cc:62)
	`anonymous namespace'::TestEnvironment::TestEnvironment [0x004035A4+244]
(c:\b\build\slave\webkit-win-latest-dbg\build\src\webkit\support\webkit_support.cc:93)
	webkit_support::SetUpTestEnvironmentForUnitTests [0x004032B8+232]
(c:\b\build\slave\webkit-win-latest-dbg\build\src\webkit\support\webkit_support.cc:134)
	content::SetUpTestEnvironmentForWebKitUnitTests [0x00401138+8]
(c:\b\build\slave\webkit-win-latest-dbg\build\src\content\test\webkit_unit_test_support.cc:13)
	main [0x004010AB+43]
(c:\b\build\slave\webkit-win-latest-dbg\build\src\third_party\webkit\source\web\tests\runalltests.cpp:57)
	__tmainCRTStartup [0x0040164F+447]
(f:\dd\vctools\crt_bld\self_x86\crt\src\crtexe.c:555)
	mainCRTStartup [0x0040147F+15]
(f:\dd\vctools\crt_bld\self_x86\crt\src\crtexe.c:371)
	BaseThreadInitThunk [0x7498339A+18]
	RtlInitializeExceptionChain [0x77269EF2+99]
	RtlInitializeExceptionChain [0x77269EC5+54]
No data was available to update the JSON results

[abarth-chromium, 2013-08-01 22:16:30]

I asked Justin to test this CL on a win64 debug build and he didn't hit the
assert.  It's very strange...

[commit-bot: I haz the power, 2013-08-02 03:33:38]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/abarth@chromium.org/21437003/25001

[commit-bot: I haz the power, 2013-08-02 05:25:38]

Change committed as 155404

[abarth-chromium, 2013-08-03 01:19:31]

I updated StaticString.pm to use placement new instead of reinterpret_cast. 
Hopefully Windows will like that better.

[commit-bot: I haz the power, 2013-08-03 01:20:10]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/abarth@chromium.org/21437003/38002

[commit-bot: I haz the power, 2013-08-03 03:57:06]

Change committed as 155464