WebRTC: Enable bot_update for standalone recipe.

WebRTC: Enable bot_update for standalone recipe.

Depends on https://codereview.chromium.org/627493002/ being
landed first.

BUG=webrtc:3534, 419516
TEST=Passing build with https://codereview.chromium.org/627493002/ applied:
scripts/tools/run_recipe.py webrtc/standalone buildername="Linux64 Debug" mastername=client.webrtc slavename=x revision='"7417"' use_mirror=True
Also tried a trybot:
scripts/tools/run_recipe.py webrtc/standalone buildername="linux" mastername=tryserver.webrtc slavename=x revision='"7417"' patchset=1 issue=12659005 rietveld=https://webrtc-codereview.appspot.com use_mirror=True root=src

Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=292403

[hinoka, 2014-10-02 20:51:01]

hinoka@chromium.org changed reviewers:
+ hinoka@chromium.org

[hinoka, 2014-10-02 20:51:03]

I think this is really 3 patches.  Can you split them into three?

1. Enable tryserver.webrtc.  This lgtm
2. Add --rietveld_server flag to bot_update api recipes.  this also lgtm
3. Add custom_vars to bot_update.  I don't really like that we're depending on
this (this implies that devs and bots have differing configs, which is confusing
for everyone).  But if this goes away with WebRTC's switch to Git, its probably
okay.  lgtm

[hinoka, 2014-10-02 20:52:59]

Actually, 4 patches - Adding bot_update calls into the recipe.  This should be a
patch independent of enabling bot_update.  The reason for that is because
bot_update contains code to clean up the build/ dir if we switch bot_update from
enable->disable.

If you revert bot_update in recipes along with bot_update being active at the
same time, the git solution wouldn't get cleaned out.

[kjellander_chromium, 2014-10-03 06:00:58]

On 2014/10/02 20:52:59, hinoka wrote:
> Actually, 4 patches - Adding bot_update calls into the recipe.  This should be
a
> patch independent of enabling bot_update.  The reason for that is because
> bot_update contains code to clean up the build/ dir if we switch bot_update
from
> enable->disable.
> 
> If you revert bot_update in recipes along with bot_update being active at the
> same time, the git solution wouldn't get cleaned out.

Thanks for the feedback. 
I will rework this CL for sure. I mostly put it up here as WIP to be able to
mention it in https://codereview.chromium.org/627493002/ (notice there's no
reviewer or published comments yet).
I should have stated this clearly in the description.

[kjellander_chromium, 2014-10-03 11:55:08]

I agree. Like I said earlier this was not meant to be reviewed yet.

On 2014/10/02 20:51:03, hinoka wrote:
> I think this is really 3 patches.  Can you split them into three?
> 
> 1. Enable tryserver.webrtc.  This lgtm

I decided I'll do this as a separate step after enabling bot_update in
client.webrtc instead.

> 2. Add --rietveld_server flag to bot_update api recipes.  this also lgtm

This is handled in https://codereview.chromium.org/615373002/

> 3. Add custom_vars to bot_update.  
This is handled in the following two CLs:
* https://codereview.chromium.org/619133002/ deps2git: Add support for
overriding DEPS vars
* https://codereview.chromium.org/623093002/ bot_update: Pass custom_vars dict
to deps2git.

> I don't really like that we're depending on
> this (this implies that devs and bots have differing configs, which is
confusing
> for everyone).  But if this goes away with WebRTC's switch to Git, its
probably
> okay.  lgtm

This doesn't go away just because we switch to Git, but I can certainly make an
effort 
of switching everyone over to use the default name for the code ('src') if you
desire
we remove that code.

[kjellander_chromium, 2014-10-03 11:56:35]

On 2014/10/03 06:00:58, kjellander wrote:
> On 2014/10/02 20:52:59, hinoka wrote:
> > Actually, 4 patches - Adding bot_update calls into the recipe.  This should
be
> a
> > patch independent of enabling bot_update.  The reason for that is because
> > bot_update contains code to clean up the build/ dir if we switch bot_update
> from
> > enable->disable.

I've landed that in https://codereview.chromium.org/628523002/ now.
I also decided to disable bot_update for client.webrtc (it was already enabled
in 
order to support the Android APK bots when they built from a Chromium checkout,
which
they no longer do).
I'll certainly enable it again once all the prerequisite CLs are landed.

> > If you revert bot_update in recipes along with bot_update being active at
the
> > same time, the git solution wouldn't get cleaned out.
> 
> Thanks for the feedback. 
> I will rework this CL for sure. I mostly put it up here as WIP to be able to
> mention it in https://codereview.chromium.org/627493002/ (notice there's no
> reviewer or published comments yet).
> I should have stated this clearly in the description.

[kjellander_chromium, 2014-10-09 10:50:51]

Now this is rebased and ready to land once
https://codereview.chromium.org/615373002/ is landed (no longer included here).

The work of eliminating the need of custom_vars is now also done.

[kjellander_chromium, 2014-10-09 19:50:26]

On 2014/10/09 10:50:51, kjellander wrote:
> Now this is rebased and ready to land once
> https://codereview.chromium.org/615373002/ is landed (no longer included
here).
> 
> The work of eliminating the need of custom_vars is now also done.

(it was done in https://code.google.com/p/webrtc/source/detail?r=7405)

[Ryan Tseng, 2014-10-10 05:20:01]

hinoka@google.com changed reviewers:
+ hinoka@google.com

[Ryan Tseng, 2014-10-10 05:20:02]

still lgtm

[kjellander_chromium, 2014-10-10 09:24:43]

On 2014/10/10 05:20:02, Ryan Tseng wrote:
> still lgtm

Thanks. I made the change a lot smaller and will try rolling out to one trybot
per platform initially, to avoid breaking too much if this goes wrong.
I'll TBR CLs with more of PS#4 if this succeeds.

[kjellander_chromium, 2014-10-10 13:01:52]

The CQ bit was checked by kjellander@chromium.org

[commit-bot: I haz the power, 2014-10-10 13:02:11]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/472963003/820001

[commit-bot: I haz the power, 2014-10-10 13:03:29]

Committed patchset #5 (id:820001) as 292403

[kjellander_chromium, 2014-10-10 13:11:32]

A revert of this CL (patchset #5 id:820001) has been created in
https://codereview.chromium.org/644813006/ by kjellander@chromium.org.

The reason for reverting is: Turns out our Mac trybots are using a too old
version of Git (git/1.7.4.1) the patch step fails like this:

Applying the patch.
Failed to apply patch for codereview.settings:
While running git apply --index -3 -p1 --verbose;
  error: unknown switch `3'.