WebUI test framework: fix Mock4JS verification, allow chrome.send passthrough.

WebUI test framework: fix Mock4JS verification, allow chrome.send passthrough.

1) Add test for Bug 99970 Comment 6 & fix
2) Allow chrome.send to be mocked outside of preLoad call.
3) add chrome.originalSend to allow mocked messages to pass through to C++.

Also to address need for .cc file to be correct for clang, and prevent ugliness in js with a lot of GEN lines, allow the header and .cc to not have a suffix, and the generated file to end in -gen.cc.  While doing this, addressed crbug.com/102794 to put in <(INTERMEDIATE_DIR) preventing collisions across rules.

R=flackr@chromium.org,wyck@chromium.org
BUG=103740, 99970, 102794
TEST=browser_tests --gtest_filter=Mock4JS*.* AND ChromeSend*.*

Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=109678

Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=109711

[wyck, 2011-11-11 00:51:34]

I see linux_chromeos_arm ICE'd on your try.  I don't understand the rest of what
happened.
  CXX(target) out/Release/obj.target/crypto/crypto/rsa_private_key.o
crypto/p224.cc: In function 'void crypto::p224::Negate(const
crypto::p224::Point&, crypto::p224::Point*)':
crypto/p224.cc:648:1: internal compiler error: in do_SUBST, at combine.c:681
Please submit a full bug report,
with preprocessed source if appropriate.
See <https://support.codesourcery.com/GNUToolchain/> for instructions.
make: *** [out/Release/obj.target/crypto/crypto/p224.o] Error 1

[Sheridan Rawlins, 2011-11-11 02:21:00]

I'm pretty sure that's unrelated, but I am getting clang errors.  While
addressing (creating separate .cc file - Thanks vandebo@ for the discussion) I
also addressed crbug.com/102794, which was suggested by rsleevi@ in another CL.

I'm running trybots again with the (hopefully) clang-proof combination.

-SCR

On 2011/11/11 00:51:34, wyck wrote:
> I see linux_chromeos_arm ICE'd on your try.  I don't understand the rest of
what
> happened.
>   CXX(target) out/Release/obj.target/crypto/crypto/rsa_private_key.o
> crypto/p224.cc: In function 'void crypto::p224::Negate(const
> crypto::p224::Point&, crypto::p224::Point*)':
> crypto/p224.cc:648:1: internal compiler error: in do_SUBST, at combine.c:681
> Please submit a full bug report,
> with preprocessed source if appropriate.
> See <https://support.codesourcery.com/GNUToolchain/> for instructions.
> make: *** [out/Release/obj.target/crypto/crypto/p224.o] Error 1

[flackr, 2011-11-11 15:34:54]

LGTM with nit

http://codereview.chromium.org/8515016/diff/1005/chrome/test/data/webui/chrom...
File chrome/test/data/webui/chrome_send_browsertest.h (right):

http://codereview.chromium.org/8515016/diff/1005/chrome/test/data/webui/chrom...
chrome/test/data/webui/chrome_send_browsertest.h:5: #define
CHROME_TEST_DATA_WEBUI_CHROME_SEND_BROWSERTEST_H_
+ #pragma once

[Sheridan Rawlins, 2011-11-11 17:04:51]

http://codereview.chromium.org/8515016/diff/1005/chrome/test/data/webui/chrom...
File chrome/test/data/webui/chrome_send_browsertest.h (right):

http://codereview.chromium.org/8515016/diff/1005/chrome/test/data/webui/chrom...
chrome/test/data/webui/chrome_send_browsertest.h:5: #define
CHROME_TEST_DATA_WEBUI_CHROME_SEND_BROWSERTEST_H_
On 2011/11/11 15:34:54, flackr wrote:
> + #pragma once

Done.

[Ryan Sleevi, 2011-11-11 18:48:39]

http://codereview.chromium.org/8515016/diff/9002/chrome/chrome_tests.gypi
File chrome/chrome_tests.gypi (right):

http://codereview.chromium.org/8515016/diff/9002/chrome/chrome_tests.gypi#new...
chrome/chrome_tests.gypi:2013:
'<(INTERMEDIATE_DIR)/chrome/<(RULE_INPUT_DIRNAME)/<(RULE_INPUT_ROOT)-gen.cc',
Just a minor nit - when using <(INTERMEDIATE_DIR), each intermediate is unique
per target, so you don't need to prefix with /chrome/. It's fine w/ or w/o
though.

[Sheridan Rawlins, 2011-11-11 19:07:46]

http://codereview.chromium.org/8515016/diff/9002/chrome/chrome_tests.gypi
File chrome/chrome_tests.gypi (right):

http://codereview.chromium.org/8515016/diff/9002/chrome/chrome_tests.gypi#new...
chrome/chrome_tests.gypi:2013:
'<(INTERMEDIATE_DIR)/chrome/<(RULE_INPUT_DIRNAME)/<(RULE_INPUT_ROOT)-gen.cc',
I like the similarity to the location under src/ and I'll keep the /chrome/,
though it's not required.  Thanks for the leeway to choose.

On 2011/11/11 18:48:40, Ryan Sleevi wrote:
> Just a minor nit - when using <(INTERMEDIATE_DIR), each intermediate is unique
> per target, so you don't need to prefix with /chrome/. It's fine w/ or w/o
> though.

[wyck, 2011-11-11 19:11:40]

On 2011/11/10 22:46:25, Sheridan Rawlins wrote:

LGTM

[commit-bot: I haz the power, 2011-11-11 20:47:28]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/scr@chromium.org/8515016/10004

[commit-bot: I haz the power, 2011-11-11 22:02:25]

Change committed as 109711