Pass all compile targets to "analyze"

Pass all compile targets to "analyze"

See https://groups.google.com/a/chromium.org/d/msg/infra-dev/3kUd4emLyBA/zPCnKUgk39MJ
for context.

BUG=493998
Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=295873

[Paweł Hajdan Jr., 2015-06-26 14:51:28]

phajdan.jr@chromium.org changed reviewers:
+ sergiyb@chromium.org, sky@chromium.org

[sky, 2015-06-26 15:29:00]

https://codereview.chromium.org/1212413002/diff/1/scripts/slave/recipes/chrom...
File scripts/slave/recipes/chromium_trybot.py (right):

https://codereview.chromium.org/1212413002/diff/1/scripts/slave/recipes/chrom...
scripts/slave/recipes/chromium_trybot.py:467: compile_targets))
Can we add some sort of assertion that we aren't passing in all?

[Paweł Hajdan Jr., 2015-06-26 15:42:48]

https://codereview.chromium.org/1212413002/diff/1/scripts/slave/recipes/chrom...
File scripts/slave/recipes/chromium_trybot.py (right):

https://codereview.chromium.org/1212413002/diff/1/scripts/slave/recipes/chrom...
scripts/slave/recipes/chromium_trybot.py:467: compile_targets))
On 2015/06/26 at 15:29:00, sky wrote:
> Can we add some sort of assertion that we aren't passing in all?

Good point. Done. WDYT?

[sky, 2015-06-29 02:28:26]

https://codereview.chromium.org/1212413002/diff/20001/scripts/slave/recipes/c...
File scripts/slave/recipes/chromium_trybot.py (right):

https://codereview.chromium.org/1212413002/diff/20001/scripts/slave/recipes/c...
scripts/slave/recipes/chromium_trybot.py:466: all_compile_targets(api, tests +
tests_including_triggered) +
This certainly works, but shouldn't we instead warn/assert that compile_targets
shouldn't included all? With the way you have it now all is silently ignored.
Also, I think it's 'All'.

[Sergiy Byelozyorov, 2015-06-29 11:09:37]

l-g-t-m % what sky said

[Paweł Hajdan Jr., 2015-06-29 12:38:22]

https://codereview.chromium.org/1212413002/diff/20001/scripts/slave/recipes/c...
File scripts/slave/recipes/chromium_trybot.py (right):

https://codereview.chromium.org/1212413002/diff/20001/scripts/slave/recipes/c...
scripts/slave/recipes/chromium_trybot.py:466: all_compile_targets(api, tests +
tests_including_triggered) +
On 2015/06/29 at 02:28:25, sky wrote:
> This certainly works, but shouldn't we instead warn/assert that
compile_targets shouldn't included all? With the way you have it now all is
silently ignored.

That's right. Do you think it's blocking this CL?

We might have a chicken and egg problem here: some bots may be using "all"
because of a hack in chromium_tests.analyze (that I was trying to get rid of in
https://codereview.chromium.org/1187863002).

My plan is to get rid of the need for that hack. I'd just like to move in small
steps to avoid breakages, and this CL is just a small step. This is mostly based
on
https://groups.google.com/a/chromium.org/d/msg/infra-dev/3kUd4emLyBA/W3OcZo8-...
("In short, we should change the filter/api.py script to ignore the
'build_targets' param, and use just the 'targets' param 
in the output; once we've done that, we should remove 'build_targets'
altogether.").

I believe that in order to ignore "build_targets" (a.k.a. "compile_targets") I'd
first need to ensure all targets we may want to compile are passed in "targets"
(a.k.a. "exes" or "matching_exes").

I could make a CL which does all of that in one step, it just sounds very risky.
I thought I'd first add things to exes (i.e. this CL). It shouldn't break
anything. Then I'd look into bots compiling "all" (no special alerts are needed,
it's all in the configs), and then making a change to the recipe to ignore
"build_targets" (compile_targets) and only use matching_exes while removing
"all" from the configs.

Does above plan make sense to you? Feel free to make recommendations for a
better one.

> Also, I think it's 'All'.

That's why code below uses t.lower() - it seems to me it's All on Windows but
all on other platforms. Note that all_compile_targets is a function call that is
only moved in this CL without modification.

[sky, 2015-06-29 17:07:44]

LGTM

[Paweł Hajdan Jr., 2015-06-30 07:35:29]

The CQ bit was checked by phajdan.jr@chromium.org

[commit-bot: I haz the power, 2015-06-30 07:35:41]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1212413002/20001

[commit-bot: I haz the power, 2015-06-30 07:40:09]

Committed patchset #2 (id:20001) as
http://src.chromium.org/viewvc/chrome?view=rev&revision=295873