Do not build 'all' on all the try builders by default.

Do not build 'all' on all the try builders by default.

Enforce the parameters options and tests to make sure it's not forgotten.
Significantly reduce the number of unneeded targets that are built.
This will greatly improve the TS performance.

The main reason for this CL is that foo_tests_run are not built by default.
These target archive the build output, so they must not be built unless
necessary. There is already logic to only build the foo_test_run for each
foo_test_swarm testfilter, but that is not useful as long as 'all' is built.

TBR=cmp@chromium.org
R=petermayo@chromium.org,glider@chromium.org
BUG=


Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=167088

[M-A Ruel, 2012-11-08 16:39:08]

Chase for the general idea, others for specific issues. If I don't get a reply
before the weekend, I'll revert the relevant part to 'all' so the previous
behavior is kept. I only care about the main try builders
(linux_rel,mac_rel,win_rel), the rest is utilization improvement gravy.

https://codereview.chromium.org/11365154/diff/2001/masters/master.tryserver.c...
File masters/master.tryserver.chromium/master.cfg (right):

https://codereview.chromium.org/11365154/diff/2001/masters/master.tryserver.c...
masters/master.tryserver.chromium/master.cfg:450: options=['nacl_integration'],
Brad|Mark, no need to build anything else from what I understand?

In fact, I don't know what this builder exists; it could be replaced on a
PRESUBMIT.py with:
return ['linux_rel:nacl_integration']
and if another linux_rel is specified, this step will simply be added to the
default tests so it's not a problem.

https://codereview.chromium.org/11365154/diff/2001/masters/master.tryserver.c...
masters/master.tryserver.chromium/master.cfg:1348:
options=['chromium_builder_tests'],
Peter, I'm not sure there. I can revert to 'all' instead to be safe.

https://codereview.chromium.org/11365154/diff/2001/masters/master.tryserver.c...
masters/master.tryserver.chromium/master.cfg:1443: 'chromium_builder_tests',
Alexander, master.chromium.memory/master_mac_cfg.py uses a full list of the
targets to build, so I'm not sure if 'all' should be kept here. I'll revert to
'all' if unsure.

https://codereview.chromium.org/11365154/diff/2001/masters/master.tryserver.c...
masters/master.tryserver.chromium/master.cfg:1671:
options=['chromium_builder_qa'],
Brad|Mark, I took that from master.client.nacl.sdk/master.cfg, is it fine?

https://codereview.chromium.org/11365154/diff/2001/masters/master.tryserver.c...
masters/master.tryserver.chromium/master.cfg:1708: options=['all'],
No idea if there's a meta-target for the perf targets so I stayed conservative
here.

[Alexander Potapenko, 2012-11-08 16:47:24]

https://codereview.chromium.org/11365154/diff/2001/masters/master.tryserver.c...
File masters/master.tryserver.chromium/master.cfg (right):

https://codereview.chromium.org/11365154/diff/2001/masters/master.tryserver.c...
masters/master.tryserver.chromium/master.cfg:1443: 'chromium_builder_tests',
On 2012/11/08 16:39:09, Marc-Antoine Ruel wrote:
> Alexander, master.chromium.memory/master_mac_cfg.py uses a full list of the
> targets to build, so I'm not sure if 'all' should be kept here. I'll revert to
> 'all' if unsure.

Let it be 'chromium_builder_tests'.

[M-A Ruel, 2012-11-08 16:51:09]

https://codereview.chromium.org/11365154/diff/2001/masters/master.tryserver.c...
File masters/master.tryserver.chromium/master.cfg (right):

https://codereview.chromium.org/11365154/diff/2001/masters/master.tryserver.c...
masters/master.tryserver.chromium/master.cfg:1443: 'chromium_builder_tests',
On 2012/11/08 16:47:24, Alexander Potapenko wrote:
> On 2012/11/08 16:39:09, Marc-Antoine Ruel wrote:
> > Alexander, master.chromium.memory/master_mac_cfg.py uses a full list of the
> > targets to build, so I'm not sure if 'all' should be kept here. I'll revert
to
> > 'all' if unsure.
> 
> Let it be 'chromium_builder_tests'.

Ok thanks. Note that linux_asan is using the manual style to declare the builder
and it does list its targets. I'd prefer to keep fixing this outside of the
scope of this CL so it's just FYI.

[Peter Mayo, 2012-11-09 02:19:47]

https://codereview.chromium.org/11365154/diff/2001/masters/master.tryserver.c...
File masters/master.tryserver.chromium/master.cfg (right):

https://codereview.chromium.org/11365154/diff/2001/masters/master.tryserver.c...
masters/master.tryserver.chromium/master.cfg:1348:
options=['chromium_builder_tests'],
On 2012/11/08 16:39:09, Marc-Antoine Ruel wrote:
> Peter, I'm not sure there. I can revert to 'all' instead to be safe.

Actually making this be aura_builder would cross something off my TODO list.

"win_ash should run the same tests that win_aura does, but in ash instead of on
the windows desktop. it doesn't do that now (because they don't even run without
failing on desktop)"

[M-A Ruel, 2012-11-09 23:23:19]

Ping for the other people, this will get in this weekend.

https://codereview.chromium.org/11365154/diff/2001/masters/master.tryserver.c...
File masters/master.tryserver.chromium/master.cfg (right):

https://codereview.chromium.org/11365154/diff/2001/masters/master.tryserver.c...
masters/master.tryserver.chromium/master.cfg:1348:
options=['chromium_builder_tests'],
On 2012/11/09 02:19:47, Peter Mayo wrote:
> On 2012/11/08 16:39:09, Marc-Antoine Ruel wrote:
> > Peter, I'm not sure there. I can revert to 'all' instead to be safe.
> 
> Actually making this be aura_builder would cross something off my TODO list.

done

[commit-bot: I haz the power, 2012-11-10 12:19:25]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/maruel@chromium.org/11365154/9001

[commit-bot: I haz the power, 2012-11-10 12:20:18]

Change committed as 167088