Add a copy of the "WinClang64 (dbg)" bot to the main waterfall.

Add a copy of the "WinClang64 (dbg)" bot to the main waterfall.

It's a copy of the WinClang FYI bot.

After this:
* restart main win waterfall master
* change win_clang trybot to point to this bot instead of the FYI bot
* remove bot from FYI waterfall

BUG=603935
Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=299968

[Nico, 2016-04-15 17:38:41]

thakis@chromium.org changed reviewers:
+ dpranke@chromium.org

[Dirk Pranke, 2016-04-15 18:26:27]

Does this have the final slave hostname yet? I'm not sure which one you're
actually going to use ...

https://codereview.chromium.org/1892943002/diff/20001/scripts/slave/recipe_mo...
File scripts/slave/recipe_modules/chromium_tests/chromium_win.py (right):

https://codereview.chromium.org/1892943002/diff/20001/scripts/slave/recipe_mo...
scripts/slave/recipe_modules/chromium_tests/chromium_win.py:285: 'WinClang': {
If this is a debug x64 bot, it seems like it would be less confusing to call
this 'Win Clang x64 (dbg)' ?

[Dirk Pranke, 2016-04-15 18:27:13]

Oh, also, it seems like you can change the trybots.py entry as part of this; the
bot doesn't actually need to be live for that to work.

[Nico, 2016-04-15 18:28:04]

Yes, it's the final slave name now.

https://codereview.chromium.org/1892943002/diff/20001/scripts/slave/recipe_mo...
File scripts/slave/recipe_modules/chromium_tests/chromium_win.py (right):

https://codereview.chromium.org/1892943002/diff/20001/scripts/slave/recipe_mo...
scripts/slave/recipe_modules/chromium_tests/chromium_win.py:285: 'WinClang': {
On 2016/04/15 18:26:27, Dirk Pranke wrote:
> If this is a debug x64 bot, it seems like it would be less confusing to call
> this 'Win Clang x64 (dbg)' ?

Hm, the reasoning behind the current name is that I'd like to be able to change
the configuration of the bot if necessary (between debug/release, component/not
component etc), and if the config is in the name that's close to impossible
since the try waterfall refers to this bot by name.

[Nico, 2016-04-15 18:31:00]

I think I'd prefer to do the try bot separately. If the config is busted for
some strange reason, doing the try change now would wedge the cq. If I land this
first and check the bot comes on alright on the main waterfall, I can be more
confident that switching the try master is safe.

[Dirk Pranke, 2016-04-15 18:31:51]

On 2016/04/15 18:28:04, Nico wrote:
> Yes, it's the final slave name now.
> 
>
https://codereview.chromium.org/1892943002/diff/20001/scripts/slave/recipe_mo...
> File scripts/slave/recipe_modules/chromium_tests/chromium_win.py (right):
> 
>
https://codereview.chromium.org/1892943002/diff/20001/scripts/slave/recipe_mo...
> scripts/slave/recipe_modules/chromium_tests/chromium_win.py:285: 'WinClang': {
> On 2016/04/15 18:26:27, Dirk Pranke wrote:
> > If this is a debug x64 bot, it seems like it would be less confusing to call
> > this 'Win Clang x64 (dbg)' ?
> 
> Hm, the reasoning behind the current name is that I'd like to be able to
change
> the configuration of the bot if necessary (between debug/release,
component/not
> component etc), and if the config is in the name that's close to impossible
> since the try waterfall refers to this bot by name.

The trybots pick it up via trybots.py, so you can change both names at once.

However, it would be rather strange to change the configuration of a waterfall
bot
without changing the name; people expect a given builder to do something
predictable,
and we tend to follow these conventions for a reason. Even flipping between GYP
and 
GN has confused people.

Why would you need to change configurations often enough that you wouldn't also
want to change the name at the same time?

[Nico, 2016-04-15 18:38:30]

On 2016/04/15 18:31:51, Dirk Pranke wrote:
> On 2016/04/15 18:28:04, Nico wrote:
> > Yes, it's the final slave name now.
> > 
> >
>
https://codereview.chromium.org/1892943002/diff/20001/scripts/slave/recipe_mo...
> > File scripts/slave/recipe_modules/chromium_tests/chromium_win.py (right):
> > 
> >
>
https://codereview.chromium.org/1892943002/diff/20001/scripts/slave/recipe_mo...
> > scripts/slave/recipe_modules/chromium_tests/chromium_win.py:285: 'WinClang':
{
> > On 2016/04/15 18:26:27, Dirk Pranke wrote:
> > > If this is a debug x64 bot, it seems like it would be less confusing to
call
> > > this 'Win Clang x64 (dbg)' ?
> > 
> > Hm, the reasoning behind the current name is that I'd like to be able to
> change
> > the configuration of the bot if necessary (between debug/release,
> component/not
> > component etc), and if the config is in the name that's close to impossible
> > since the try waterfall refers to this bot by name.
> 
> The trybots pick it up via trybots.py, so you can change both names at once.

Ah, that's a good point.

> However, it would be rather strange to change the configuration of a waterfall
> bot
> without changing the name; people expect a given builder to do something
> predictable,
> and we tend to follow these conventions for a reason. Even flipping between
GYP
> and 
> GN has confused people.
> 
> Why would you need to change configurations often enough that you wouldn't
also
> want to change the name at the same time?

I was worried about cycle time when I set up the win_clang bot. Now that it's
been around a bit, I'm admittedly less worried. And being able to change the
name in one CL means having a descriptive name isn't a problem, so I'll do that.

However, as far as I know
(https://bugs.chromium.org/p/chromium/issues/detail?id=399990), names longer
than 17 characters will break things, so the name can't be 'WinClang x64 GN
(dbg)' since that's 21 chars. Since the "GN" bit will be redundant soon, how
about 'WinClang64 (dbg)'?

[Dirk Pranke, 2016-04-15 18:40:44]

On 2016/04/15 18:38:30, Nico wrote:
> On 2016/04/15 18:31:51, Dirk Pranke wrote:
> > On 2016/04/15 18:28:04, Nico wrote:
> > > Yes, it's the final slave name now.
> > > 
> > >
> >
>
https://codereview.chromium.org/1892943002/diff/20001/scripts/slave/recipe_mo...
> > > File scripts/slave/recipe_modules/chromium_tests/chromium_win.py (right):
> > > 
> > >
> >
>
https://codereview.chromium.org/1892943002/diff/20001/scripts/slave/recipe_mo...
> > > scripts/slave/recipe_modules/chromium_tests/chromium_win.py:285:
'WinClang':
> {
> > > On 2016/04/15 18:26:27, Dirk Pranke wrote:
> > > > If this is a debug x64 bot, it seems like it would be less confusing to
> call
> > > > this 'Win Clang x64 (dbg)' ?
> > > 
> > > Hm, the reasoning behind the current name is that I'd like to be able to
> > change
> > > the configuration of the bot if necessary (between debug/release,
> > component/not
> > > component etc), and if the config is in the name that's close to
impossible
> > > since the try waterfall refers to this bot by name.
> > 
> > The trybots pick it up via trybots.py, so you can change both names at once.
> 
> Ah, that's a good point.
> 
> > However, it would be rather strange to change the configuration of a
waterfall
> > bot
> > without changing the name; people expect a given builder to do something
> > predictable,
> > and we tend to follow these conventions for a reason. Even flipping between
> GYP
> > and 
> > GN has confused people.
> > 
> > Why would you need to change configurations often enough that you wouldn't
> also
> > want to change the name at the same time?
> 
> I was worried about cycle time when I set up the win_clang bot. Now that it's
> been around a bit, I'm admittedly less worried. And being able to change the
> name in one CL means having a descriptive name isn't a problem, so I'll do
that.
> 
> However, as far as I know
> (https://bugs.chromium.org/p/chromium/issues/detail?id=399990), names longer
> than 17 characters will break things, so the name can't be 'WinClang x64 GN
> (dbg)' since that's 21 chars. Since the "GN" bit will be redundant soon, how
> about 'WinClang64 (dbg)'?

I wonder if the paths are still that long w/ GN; I have a vague feeling that
they're shorter,
but I could be totally wrong.

'WinClang64 (dbg)' sgtm.

[Nico, 2016-04-15 18:44:26]

Description was changed from

==========
Add a copy of the WinClang FYI bot to the main waterfall.

After this:
* restart main win waterfall master
* change win_clang trybot to point to this bot instead of the FYI bot
* remove bot from FYI waterfall

BUG=603935
==========

to

==========
Add a copy of the "WinClang64 (dbg)" bot to the main waterfall.

It's a copy of the WinClang FYI bot.

After this:
* restart main win waterfall master
* change win_clang trybot to point to this bot instead of the FYI bot
* remove bot from FYI waterfall

BUG=603935
==========

[Nico, 2016-04-15 18:44:34]

done

[Dirk Pranke, 2016-04-15 20:01:12]

lgtm.

[Nico, 2016-04-15 20:01:58]

The CQ bit was checked by thakis@chromium.org

[commit-bot: I haz the power, 2016-04-15 20:02:05]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1892943002/40001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1892943002/40001

[commit-bot: I haz the power, 2016-04-15 20:09:35]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2016-04-15 20:09:36]

Try jobs failed on following builders:
  Build Presubmit on tryserver.infra (JOB_FAILED,
https://build.chromium.org/p/tryserver.infra/builders/Build%20Presubmit/build...)

[Nico, 2016-04-15 21:12:27]

Main waterfall ng-trybot coverage: 70.00
Regression, the following builders lack in-sync tryserver coverage:
	master.chromium.win:WinClang64 (dbg)

I guess that means I must do the trybot change in this CL after all.

[Nico, 2016-04-15 21:20:52]

Ok, did that.

[Nico, 2016-04-15 21:20:55]

The CQ bit was checked by thakis@chromium.org

[Nico, 2016-04-15 21:20:56]

The patchset sent to the CQ was uploaded after l-g-t-m from dpranke@chromium.org
Link to the patchset: https://codereview.chromium.org/1892943002/#ps60001
(title: "trybot")

[commit-bot: I haz the power, 2016-04-15 21:21:01]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1892943002/60001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1892943002/60001

[commit-bot: I haz the power, 2016-04-15 21:24:59]

Description was changed from

==========
Add a copy of the "WinClang64 (dbg)" bot to the main waterfall.

It's a copy of the WinClang FYI bot.

After this:
* restart main win waterfall master
* change win_clang trybot to point to this bot instead of the FYI bot
* remove bot from FYI waterfall

BUG=603935
==========

to

==========
Add a copy of the "WinClang64 (dbg)" bot to the main waterfall.

It's a copy of the WinClang FYI bot.

After this:
* restart main win waterfall master
* change win_clang trybot to point to this bot instead of the FYI bot
* remove bot from FYI waterfall

BUG=603935

Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=299968
==========

[commit-bot: I haz the power, 2016-04-15 21:25:01]

Committed patchset #4 (id:60001) as
http://src.chromium.org/viewvc/chrome?view=rev&revision=299968

[Dirk Pranke, 2016-04-15 21:33:06]

On 2016/04/15 21:12:27, Nico wrote:
> Main waterfall ng-trybot coverage: 70.00
> Regression, the following builders lack in-sync tryserver coverage:
> 	master.chromium.win:WinClang64 (dbg)
> 
> I guess that means I must do the trybot change in this CL after all.

You can actually whitelist the bot in the test, but I would've flipped the
trybot over ...

[Nico, 2016-04-15 21:37:08]

Yeah, I saw the whitelist (erikchen used it for the 10.11 os x bot), but at that
point I felt I was fighting the system too much :-)