New pyauto automation hook to get browser process information.

New pyauto automation hook to get browser process information.

Also modified the pyauto tests in process_count.py to use the new
automation hook.  Verified that this works on my local ChromeOS device,
my local Linux machine, and the Win/Mac trybots.  Re-enabling the tests
on Win/Mac in the CONTINUOUS suite to see if they now pass continuously.

BUG=94650, 95142, 93822
TEST=None

Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=102406

[Nirnimesh, 2011-09-21 22:42:35]

http://codereview.chromium.org/7976016/diff/1/chrome/browser/automation/autom...
File chrome/browser/automation/automation_provider_observers.cc (right):

http://codereview.chromium.org/7976016/diff/1/chrome/browser/automation/autom...
chrome/browser/automation/automation_provider_observers.cc:2762: const
std::vector<ProcessData>& browser_processes = processes();
This includes all processes right (extensions, plugins, renderers)?.
browser_processes is a misnomer here. Use some other variable name without
'browser' in it. all_processes?

http://codereview.chromium.org/7976016/diff/1/chrome/browser/automation/autom...
chrome/browser/automation/automation_provider_observers.cc:2806: // The
following condition avoids a NOTREACHED in debug builds when the
No need to repeat this comment

http://codereview.chromium.org/7976016/diff/1/chrome/test/functional/process_...
File chrome/test/functional/process_count.py (right):

http://codereview.chromium.org/7976016/diff/1/chrome/test/functional/process_...
chrome/test/functional/process_count.py:46: self.pp =
pprint.PrettyPrinter(indent=2)
I think it would be nice to make the pretty printer available to all tests, so
define this in pyauto.py such that self.pprint can be used everywhere.

http://codereview.chromium.org/7976016/diff/1/chrome/test/pyautolib/pyauto.py
File chrome/test/pyautolib/pyauto.py (right):

http://codereview.chromium.org/7976016/diff/1/chrome/test/pyautolib/pyauto.py...
chrome/test/pyautolib/pyauto.py:1394: dictionaries containing information about
each browser process name.
do you expect multiple browser processes?

[dennis_jeffrey, 2011-09-22 18:25:34]

http://codereview.chromium.org/7976016/diff/1/chrome/browser/automation/autom...
File chrome/browser/automation/automation_provider_observers.cc (right):

http://codereview.chromium.org/7976016/diff/1/chrome/browser/automation/autom...
chrome/browser/automation/automation_provider_observers.cc:2762: const
std::vector<ProcessData>& browser_processes = processes();
On 2011/09/21 22:42:35, Nirnimesh wrote:
> This includes all processes right (extensions, plugins, renderers)?.
> browser_processes is a misnomer here. Use some other variable name without
> 'browser' in it. all_processes?

Yes, it includes all processes associated with a browser.  I can see the
confusion in my variable name between "all processes associated with a browser"
and the particular type of process known as the "browser process".  I changed
the variable name as requested here.

http://codereview.chromium.org/7976016/diff/1/chrome/browser/automation/autom...
chrome/browser/automation/automation_provider_observers.cc:2806: // The
following condition avoids a NOTREACHED in debug builds when the
On 2011/09/21 22:42:35, Nirnimesh wrote:
> No need to repeat this comment

Done.

http://codereview.chromium.org/7976016/diff/1/chrome/test/functional/process_...
File chrome/test/functional/process_count.py (right):

http://codereview.chromium.org/7976016/diff/1/chrome/test/functional/process_...
chrome/test/functional/process_count.py:46: self.pp =
pprint.PrettyPrinter(indent=2)
On 2011/09/21 22:42:35, Nirnimesh wrote:
> I think it would be nice to make the pretty printer available to all tests, so
> define this in pyauto.py such that self.pprint can be used everywhere.

Good idea.  There are a lot of places throughout the pyauto code where
PrettyPrinter objects are created.  Do you think it's worth it to do a cleanup
CL to have them all use the new PrettyPrinter object available to all?

http://codereview.chromium.org/7976016/diff/1/chrome/test/pyautolib/pyauto.py
File chrome/test/pyautolib/pyauto.py (right):

http://codereview.chromium.org/7976016/diff/1/chrome/test/pyautolib/pyauto.py...
chrome/test/pyautolib/pyauto.py:1394: dictionaries containing information about
each browser process name.
On 2011/09/21 22:42:35, Nirnimesh wrote:
> do you expect multiple browser processes?

Yes.  It doesn't seem to happen on Linux or ChromeOS, but I noticed on the
Win/Mac trybots that we get browser entries returned for other browsers like IE,
Firefox, and Safari.

[Nirnimesh, 2011-09-22 19:35:25]

LGTM

http://codereview.chromium.org/7976016/diff/1/chrome/test/functional/process_...
File chrome/test/functional/process_count.py (right):

http://codereview.chromium.org/7976016/diff/1/chrome/test/functional/process_...
chrome/test/functional/process_count.py:46: self.pp =
pprint.PrettyPrinter(indent=2)

> Good idea.  There are a lot of places throughout the pyauto code where
> PrettyPrinter objects are created.  Do you think it's worth it to do a cleanup
> CL to have them all use the new PrettyPrinter object available to all?

Yes. In another CL yes

http://codereview.chromium.org/7976016/diff/12001/chrome/test/pyautolib/pyaut...
File chrome/test/pyautolib/pyauto.py (right):

http://codereview.chromium.org/7976016/diff/12001/chrome/test/pyautolib/pyaut...
chrome/test/pyautolib/pyauto.py:141: self.pretty_printer =
pprint.PrettyPrinter(indent=2)
I meant you should define this in pyauto.py
I realize I'm going out of scope of this CL, so feel free to do this in a
separate Cl

http://codereview.chromium.org/7976016/diff/12001/chrome/test/pyautolib/pyaut...
chrome/test/pyautolib/pyauto.py:143: self.pformat = self.pretty_printer.pformat
This is unused

[Nirnimesh, 2011-09-22 19:36:48]

Corrections

http://codereview.chromium.org/7976016/diff/12001/chrome/test/pyautolib/pyaut...
File chrome/test/pyautolib/pyauto.py (right):

http://codereview.chromium.org/7976016/diff/12001/chrome/test/pyautolib/pyaut...
chrome/test/pyautolib/pyauto.py:141: self.pretty_printer =
pprint.PrettyPrinter(indent=2)
ok. I'm obviously high. Ignore my previous comment

http://codereview.chromium.org/7976016/diff/12001/chrome/test/pyautolib/pyaut...
chrome/test/pyautolib/pyauto.py:142: self.pprint = self.pretty_printer.pprint
Why not just self.pprint = pprint.pprint?

[dennis_jeffrey, 2011-09-22 19:54:02]

http://codereview.chromium.org/7976016/diff/1/chrome/test/functional/process_...
File chrome/test/functional/process_count.py (right):

http://codereview.chromium.org/7976016/diff/1/chrome/test/functional/process_...
chrome/test/functional/process_count.py:46: self.pp =
pprint.PrettyPrinter(indent=2)
On 2011/09/22 19:35:26, Nirnimesh wrote:
> 
> > Good idea.  There are a lot of places throughout the pyauto code where
> > PrettyPrinter objects are created.  Do you think it's worth it to do a
cleanup
> > CL to have them all use the new PrettyPrinter object available to all?
> 
> Yes. In another CL yes

Ok, I'll take care of it in another CL.

http://codereview.chromium.org/7976016/diff/12001/chrome/test/pyautolib/pyaut...
File chrome/test/pyautolib/pyauto.py (right):

http://codereview.chromium.org/7976016/diff/12001/chrome/test/pyautolib/pyaut...
chrome/test/pyautolib/pyauto.py:141: self.pretty_printer =
pprint.PrettyPrinter(indent=2)
On 2011/09/22 19:35:26, Nirnimesh wrote:
> I meant you should define this in pyauto.py
> I realize I'm going out of scope of this CL, so feel free to do this in a
> separate Cl

Done.

http://codereview.chromium.org/7976016/diff/12001/chrome/test/pyautolib/pyaut...
chrome/test/pyautolib/pyauto.py:141: self.pretty_printer =
pprint.PrettyPrinter(indent=2)
On 2011/09/22 19:36:48, Nirnimesh wrote:
> ok. I'm obviously high. Ignore my previous comment

If you've realized you're high, you're not high enough.  Or else you are too
high.

http://codereview.chromium.org/7976016/diff/12001/chrome/test/pyautolib/pyaut...
chrome/test/pyautolib/pyauto.py:142: self.pprint = self.pretty_printer.pprint
On 2011/09/22 19:36:48, Nirnimesh wrote:
> Why not just self.pprint = pprint.pprint?

Cool!  I didn't realize I could do this directly without first creating a
PrettyPrinter object.

http://codereview.chromium.org/7976016/diff/12001/chrome/test/pyautolib/pyaut...
chrome/test/pyautolib/pyauto.py:143: self.pformat = self.pretty_printer.pformat
On 2011/09/22 19:35:26, Nirnimesh wrote:
> This is unused

I use it in process_count.py in the current CL.

[Nirnimesh, 2011-09-22 21:11:23]

LGTM