Make components_browsertests swarmable and add to main waterfall.

Make components_browsertests swarmable and add to main waterfall.

- Makes components_browsertests swarmable.
- Adds components_browsertests to the main waterfall for linux, mac and
  win bots.
- Updates the chromium tryserver to run components_browsertests as
  swarmable.

BUG=311270
Committed: https://crrev.com/7a3df3b7171fa87f556d22803fdc02a67e975865
Cr-Commit-Position: refs/heads/master@{#312886}

[nyquist, 2015-01-21 22:08:51]

nyquist@chromium.org changed reviewers:
+ jam@chromium.org

[nyquist, 2015-01-21 22:08:51]

jam: PTAL

[jam, 2015-01-22 16:42:56]

you'll also want to modify chromium_memory_trybot.json to get the asan trybot

https://codereview.chromium.org/863023003/diff/20001/components/components_br...
File components/components_browsertests.isolate (right):

https://codereview.chromium.org/863023003/diff/20001/components/components_br...
components/components_browsertests.isolate:7: '../gin/v8.isolate',
you sure you need this?

https://codereview.chromium.org/863023003/diff/20001/components/components_br...
components/components_browsertests.isolate:15: '../net/data/',
just to be clear, these two directories are needed?

https://codereview.chromium.org/863023003/diff/20001/components/components_br...
components/components_browsertests.isolate:23:
'<(PRODUCT_DIR)/content_shell.pak',
do you know what is using this pak file? it's strange that
component_browsertests uses it

https://codereview.chromium.org/863023003/diff/20001/components/components_br...
components/components_browsertests.isolate:66: '<(PRODUCT_DIR)/Content
Shell.app/',
are you sure you need this?

[jam, 2015-01-22 16:42:56]

you'll also want to modify chromium_memory_trybot.json to get the asan trybot

[nyquist, 2015-01-22 22:47:33]

jam: PTAL

Also added to chromium_memory_trybot.json.

https://codereview.chromium.org/863023003/diff/20001/components/components_br...
File components/components_browsertests.isolate (right):

https://codereview.chromium.org/863023003/diff/20001/components/components_br...
components/components_browsertests.isolate:7: '../gin/v8.isolate',
On 2015/01/22 16:42:56, jam wrote:
> you sure you need this?

Yes, otherwise the tests would fail with no visible output. It took me some time
to realize this was needed.

The output would look like this:
...
[27/27] DistillerPageWebContentsTest.TestTitleAndContentAreNeverEmpty (126 ms)
Too many failing tests (27), skipping retries.
27 tests failed:
...

https://codereview.chromium.org/863023003/diff/20001/components/components_br...
components/components_browsertests.isolate:15: '../net/data/',
On 2015/01/22 16:42:56, jam wrote:
> just to be clear, these two directories are needed?

Apparently not. Removed a few more.

https://codereview.chromium.org/863023003/diff/20001/components/components_br...
components/components_browsertests.isolate:23:
'<(PRODUCT_DIR)/content_shell.pak',
On 2015/01/22 16:42:55, jam wrote:
> do you know what is using this pak file? it's strange that
> component_browsertests uses it

The test DistillerPageWebContentsTest.TestTitleAndContentAreNeverEmpty fails
without it.

It reads string resources defined in //components/dom_distiller_strings.grdp,
which is part of //components/components_strings.grd and creates
components/strings/components_strings_en-US.pak (I think). This is repacked
together with components_resources.pak into
<(PRODUCT_DIR)/components_resources.pak (as an action in the
'components_browsertests' target).
I've checked the contents of <(PRODUCT_DIR)/components_resources.pak using
data_pack.py and it does contain the strings I need.

Also, it needs an HTML file that is part of components_resources.pak.

However, it does not seem to work without also adding content_shell.pak. I am
definitely open to suggestions though.

The components_browsertests target is modeled around content_browsertests, and
uses the content shell.

https://codereview.chromium.org/863023003/diff/20001/components/components_br...
components/components_browsertests.isolate:66: '<(PRODUCT_DIR)/Content
Shell.app/',
On 2015/01/22 16:42:56, jam wrote:
> are you sure you need this?

I believe so, though I don't have a Mac. However, the
DistillerPageWebContentsTest extends ContentBrowserTest which uses the path:
"Content Shell.app/Contents/MacOS/Content Shell".

Since we modeled components_browsertests around content_browsertests, and they
need it, I assumed we would have similar dependencies.

[jam, 2015-01-23 17:35:08]

lgtm

https://codereview.chromium.org/863023003/diff/20001/components/components_br...
File components/components_browsertests.isolate (right):

https://codereview.chromium.org/863023003/diff/20001/components/components_br...
components/components_browsertests.isolate:23:
'<(PRODUCT_DIR)/content_shell.pak',
On 2015/01/22 22:47:33, nyquist wrote:
> On 2015/01/22 16:42:55, jam wrote:
> > do you know what is using this pak file? it's strange that
> > component_browsertests uses it
> 
> The test DistillerPageWebContentsTest.TestTitleAndContentAreNeverEmpty fails
> without it.
> 
> It reads string resources defined in //components/dom_distiller_strings.grdp,
> which is part of //components/components_strings.grd and creates
> components/strings/components_strings_en-US.pak (I think). This is repacked
> together with components_resources.pak into
> <(PRODUCT_DIR)/components_resources.pak (as an action in the
> 'components_browsertests' target).
> I've checked the contents of <(PRODUCT_DIR)/components_resources.pak using
> data_pack.py and it does contain the strings I need.
> 
> Also, it needs an HTML file that is part of components_resources.pak.
> 
> However, it does not seem to work without also adding content_shell.pak. I am
> definitely open to suggestions though.
> 
> The components_browsertests target is modeled around content_browsertests, and
> uses the content shell.

I see, ok.

[nyquist, 2015-01-23 17:59:34]

The CQ bit was checked by nyquist@chromium.org

[commit-bot: I haz the power, 2015-01-23 18:00:47]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/863023003/40001

[commit-bot: I haz the power, 2015-01-23 18:57:33]

Committed patchset #3 (id:40001)

[commit-bot: I haz the power, 2015-01-23 18:59:09]

Patchset 3 (id:??) landed as
https://crrev.com/7a3df3b7171fa87f556d22803fdc02a67e975865
Cr-Commit-Position: refs/heads/master@{#312886}

[jam, 2015-01-24 00:22:57]

A revert of this CL (patchset #3 id:40001) has been created in
https://codereview.chromium.org/856323003/ by jam@chromium.org.

The reason for reverting is: Cause mac build flakiness, see
http://build.chromium.org/p/chromium.chrome/builders/Google%20Chrome%20Mac/bu...

I think the '<(PRODUCT_DIR)/Content Shell.app/', line in the isolate needs to
go. Looking closer, we should be launching the components_browsertest binary for
the renderer as well. Right now it launches Content Shell for the renderer,
which seems odd..