Delete Automation[Tab/Renderer]Helper and users.

Delete Automation[Tab/Renderer]Helper and fix users.

For Chrome Endure, use the memory benchmarking V8 extension instead.
For the old ChromeDriver, just mark the messages as deprecated on the client side. I'll remove all the old ChromeDriver code later. I want to keep it here since we release ChromeDriver from HEAD, and we might need to do critical fixes for Chrome 28 and below still.
BUG=240580
Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=202205

[dennisjeffrey, 2013-05-14 17:15:03]

Dai, could you check out the changes to perf_endure.py here?  It affects
Chrome Endure's use of the Deep Memory Profiler.


On Tue, May 14, 2013 at 10:09 AM, <kkania@chromium.org> wrote:

> Reviewers: jam, dennis_jeffrey,
>
> Description:
> Delete Automation[Tab/Renderer]Helper and fix users.
>
> For Chrome Endure, use the memory benchmarking V8 extension instead.
> For the old ChromeDriver, just mark the messages as deprecated on the
> client
> side. I'll remove all the old ChromeDriver code later. I want to keep it
> here
> since we release ChromeDriver from HEAD, and we might need to do critical
> fixes
> for Chrome 28 and below still.
> BUG=240580
>
> Please review this at
https://codereview.chromium.**org/14977013/<https://codereview.chromium.org/1...
>
> SVN Base:
svn://svn.chromium.org/chrome/**trunk/src<http://svn.chromium.org/chrome/trunk/src>
>
> Affected files:
>   M chrome/browser/automation/**automation_provider_observers.**h
>   M chrome/browser/automation/**automation_provider_observers.**cc
>   D chrome/browser/automation/**automation_tab_helper.h
>   D chrome/browser/automation/**automation_tab_helper.cc
>   D chrome/browser/automation/**automation_tab_helper_**browsertest.cc
>   D chrome/browser/automation/**automation_tab_helper_**unittest.cc
>   D chrome/browser/automation/**mock_tab_event_observer.h
>   D chrome/browser/automation/**mock_tab_event_observer.cc
>   M chrome/browser/automation/**testing_automation_provider.h
>   M chrome/browser/automation/**testing_automation_provider.cc
>   M chrome/browser/ui/browser_tab_**contents.cc
>   M chrome/browser/ui/startup/**startup_browser_creator.cc
>   M chrome/chrome_browser.gypi
>   M chrome/chrome_renderer.gypi
>   M chrome/chrome_tests.gypi
>   M chrome/chrome_tests_unit.gypi
>   D chrome/renderer/automation/**OWNERS
>   D chrome/renderer/automation/**automation_renderer_helper.h
>   D chrome/renderer/automation/**automation_renderer_helper.cc
>   D chrome/renderer/automation/**automation_renderer_helper_**
> browsertest.cc
>   M chrome/renderer/chrome_**content_renderer_client.cc
>   M chrome/test/automation/**automation_json_requests.h
>   M chrome/test/automation/**automation_json_requests.cc
>   M chrome/test/automation/tab_**proxy.h
>   M chrome/test/automation/tab_**proxy.cc
>   M chrome/test/functional/perf_**endure.py
>   M chrome/test/pyautolib/pyauto.**py
>   M chrome/test/webdriver/**webdriver_automation.cc
>   M content/public/common/content_**switches.cc
>   M tools/code_coverage/coverage_**posix.py
>   M tools/valgrind/gtest_exclude/**browser_tests.gtest-drmemory_**
> win32.txt
>
>
>

[dennisjeffrey, 2013-05-14 17:16:11]

(to Dai's chromium.org account this time)


On Tue, May 14, 2013 at 10:15 AM, Dennis Jeffrey
<dennisjeffrey@google.com>wrote:

> Dai, could you check out the changes to perf_endure.py here?  It affects
> Chrome Endure's use of the Deep Memory Profiler.
>
>
> On Tue, May 14, 2013 at 10:09 AM, <kkania@chromium.org> wrote:
>
>> Reviewers: jam, dennis_jeffrey,
>>
>> Description:
>> Delete Automation[Tab/Renderer]Helper and fix users.
>>
>> For Chrome Endure, use the memory benchmarking V8 extension instead.
>> For the old ChromeDriver, just mark the messages as deprecated on the
>> client
>> side. I'll remove all the old ChromeDriver code later. I want to keep it
>> here
>> since we release ChromeDriver from HEAD, and we might need to do critical
>> fixes
>> for Chrome 28 and below still.
>> BUG=240580
>>
>> Please review this at
https://codereview.chromium.**org/14977013/<https://codereview.chromium.org/1...
>>
>> SVN Base:
svn://svn.chromium.org/chrome/**trunk/src<http://svn.chromium.org/chrome/trunk/src>
>>
>> Affected files:
>>   M chrome/browser/automation/**automation_provider_observers.**h
>>   M chrome/browser/automation/**automation_provider_observers.**cc
>>   D chrome/browser/automation/**automation_tab_helper.h
>>   D chrome/browser/automation/**automation_tab_helper.cc
>>   D chrome/browser/automation/**automation_tab_helper_**browsertest.cc
>>   D chrome/browser/automation/**automation_tab_helper_**unittest.cc
>>   D chrome/browser/automation/**mock_tab_event_observer.h
>>   D chrome/browser/automation/**mock_tab_event_observer.cc
>>   M chrome/browser/automation/**testing_automation_provider.h
>>   M chrome/browser/automation/**testing_automation_provider.cc
>>   M chrome/browser/ui/browser_tab_**contents.cc
>>   M chrome/browser/ui/startup/**startup_browser_creator.cc
>>   M chrome/chrome_browser.gypi
>>   M chrome/chrome_renderer.gypi
>>   M chrome/chrome_tests.gypi
>>   M chrome/chrome_tests_unit.gypi
>>   D chrome/renderer/automation/**OWNERS
>>   D chrome/renderer/automation/**automation_renderer_helper.h
>>   D chrome/renderer/automation/**automation_renderer_helper.cc
>>   D chrome/renderer/automation/**automation_renderer_helper_**
>> browsertest.cc
>>   M chrome/renderer/chrome_**content_renderer_client.cc
>>   M chrome/test/automation/**automation_json_requests.h
>>   M chrome/test/automation/**automation_json_requests.cc
>>   M chrome/test/automation/tab_**proxy.h
>>   M chrome/test/automation/tab_**proxy.cc
>>   M chrome/test/functional/perf_**endure.py
>>   M chrome/test/pyautolib/pyauto.**py
>>   M chrome/test/webdriver/**webdriver_automation.cc
>>   M content/public/common/content_**switches.cc
>>   M tools/code_coverage/coverage_**posix.py
>>   M tools/valgrind/gtest_exclude/**browser_tests.gtest-drmemory_**
>> win32.txt
>>
>>
>>
>

[Dai Mikurube (NOT FULLTIME), 2013-05-14 17:28:03]

For perf_endure.py...

It looks basically good for now, but the argument ('renderer' and 'browser') is
going to be supported with the patch https://codereview.chromium.org/15082004/
by bulach@.

(Calling with no argument would work as 'renderer', but clearly specifying
'renderer' might be better when it's supported.)

[kkania, 2013-05-14 18:06:40]

On 2013/05/14 17:28:03, Dai Mikurube wrote:
> For perf_endure.py...
> 
> It looks basically good for now, but the argument ('renderer' and 'browser')
is
> going to be supported with the patch https://codereview.chromium.org/15082004/
> by bulach@.
> 
> (Calling with no argument would work as 'renderer', but clearly specifying
> 'renderer' might be better when it's supported.)

Thanks for the look Dai. Since PyAuto is deprecated and none of the tests
currently do browser dumps, I'm not going to bother documenting this. However,
just to make sure we're on the same page, the CL you reference checks the reason
is 'browser', which is different than the old code, where there was a
process_type and reason. This change should allow the user to specify 'browser'
as the reason and get dumps.

Dennis, can you give OWNERS approval for chrome/test/func?

[dennis_jeffrey, 2013-05-14 18:11:03]

chrome/test/* LGTM

[kkania, 2013-05-14 18:37:06]

+tsepez for OWNERS review of automation_messages.h; this should be easy since
I'm just deleting stuff

[Tom Sepez, 2013-05-14 18:39:06]

Messages LGTM.

[Dai Mikurube (NOT FULLTIME), 2013-05-14 20:21:39]

On 2013/05/14 18:06:40, kkania wrote:
> On 2013/05/14 17:28:03, Dai Mikurube wrote:
> > For perf_endure.py...
> > 
> > It looks basically good for now, but the argument ('renderer' and 'browser')
> is
> > going to be supported with the patch
https://codereview.chromium.org/15082004/
> > by bulach@.
> > 
> > (Calling with no argument would work as 'renderer', but clearly specifying
> > 'renderer' might be better when it's supported.)
> 
> Thanks for the look Dai. Since PyAuto is deprecated and none of the tests
> currently do browser dumps, I'm not going to bother documenting this. However,
> just to make sure we're on the same page, the CL you reference checks the
reason
> is 'browser', which is different than the old code, where there was a
> process_type and reason. This change should allow the user to specify
'browser'
> as the reason and get dumps.
> 
Ah, sorry, it was my mistake.  In my first intention, the V8 API
HeapProfilerDump() would finally take the same arguments (process_type and
reason) as the old PyAuto API.

The V8 API *MAY* be fixed someday soon, but I think we can start from this
current patch.  When the V8 API is fixed, I'll send a change in perf_endure.py
for that.  (No one else is using the V8 API HeapProfilerDump yet.)

> Dennis, can you give OWNERS approval for chrome/test/func?

[Dai Mikurube (NOT FULLTIME), 2013-05-14 20:22:32]

As of the comment #9, lgtm.

[kkania, 2013-05-15 15:25:54]

On 2013/05/14 20:22:32, Dai Mikurube wrote:
> As of the comment #9, lgtm.

John, can you review everything except
chrome/test/[functional|pyautolib|webdriver]?

[Dai Mikurube (NOT FULLTIME), 2013-05-15 15:28:24]

fyi: Now it's heapProfilerDump(process_type, reason) in
https://codereview.chromium.org/15082004/.  perf_endure.py can use the same
arguments now.

[kkania, 2013-05-16 02:33:12]

whoops, John is ooo;

sky: chrome/
joi: content/

[sky, 2013-05-16 15:03:56]

LGTM

[Jói, 2013-05-16 17:03:58]

LGTM for //content/public.

[commit-bot: I haz the power, 2013-05-24 16:14:32]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/kkania@chromium.org/14977013/57001

[commit-bot: I haz the power, 2013-05-24 22:34:05]

Change committed as 202205