1067534 - Move some android and b2g manifest configs out of mozharness and into the tree

Status: RESOLVED FIXED

(Release Engineering :: Applications: MozharnessCore, defect)

Description

[Andrew Halberstadt [:ahal]]

These are still defined in mozharness, they should live in-tree under testing/config/mozharness.

Comment 1

[Andrew Halberstadt [:ahal]]

Attachment: Make in-tree configs supercede extra_args for android

The idea of this patch is to only append the 'extra_args' to the command (which are defined in mozharness) if the option hasn't already been defined by in-tree configs. This allows us to use the config in mozharness for older branches, while letting the in-tree versions ride the train.

In order to do this, I had to assume all 'extra_args' were of the form ['--foo=bar'] instead of ['--foo', 'bar']. This might not be a great assumption to make, but I justify it in the knowledge that we eventually want all 'extra_args' in-tree.

Comment 2

[Andrew Halberstadt [:ahal]]

Attachment: (m-c) Define b2g/android manifests in in-tree config

This is a patch against m-c that makes some of the configs in mozharness obsolete (from Gecko 35 onwards).

Comment 3

[Andrew Halberstadt [:ahal]]

The previous two patches are being tested on Ash:
https://tbpl.mozilla.org/?tree=Ash&rev=899dc596b794

Comment 4

[Chris Manchester (limited bugmail, email directly)]

Comment on attachment 8489636 [details] [diff] [review]
Make in-tree configs supercede extra_args for android

Review of attachment 8489636 [details] [diff] [review]:
-----------------------------------------------------------------

::: scripts/android_emulator_unittest.py
@@ +413,5 @@
> +            argname = arg.split('=')[0]
> +            # only add the extra arg if it wasn't already defined by in-tree configs
> +            if any(a.split('=')[0] == argname for a in cmd):
> +                continue
> +            cmd.append(arg)

This is sort of too bad but I guess there isn't any other way.

Comment 5

[Chris Manchester (limited bugmail, email directly)]

Comment on attachment 8489637 [details] [diff] [review]
(m-c) Define b2g/android manifests in in-tree config

Review of attachment 8489637 [details] [diff] [review]:
-----------------------------------------------------------------

::: testing/config/mozharness/b2g_emulator_config.py
@@ +9,5 @@
>          "--remote-webserver=%(remote_webserver)s", "--ignore-window-size",
>          "--xre-path=%(xre_path)s", "--symbols-path=%(symbols_path)s", "--busybox=%(busybox)s",
>          "--total-chunks=%(total_chunks)s", "--this-chunk=%(this_chunk)s",
>          "--extra-profile-file=jsreftest/tests/user.js",
> +        "jsreftest/tests/jstests.list",

Might update the bug summary to reflect this isn't an xpcshell only change.

Comment 6

[Andrew Halberstadt [:ahal]]

I don't see anything suspicious in that Ash run:
https://hg.mozilla.org/build/mozharness/rev/707df471ab2c

Here's a cypress run to keep an eye on, just in case:
https://tbpl.mozilla.org/?tree=Cypress&rev=27253887d2cc

Comment 7

[Simone Bruno [:simone]]

In prod with reconfig on 2014-09-22 08:20 PT

Comment 8

[Andrew Halberstadt [:ahal]]

Comment on attachment 8489637 [details] [diff] [review]
(m-c) Define b2g/android manifests in in-tree config

Try run looks good:
https://hg.mozilla.org/integration/mozilla-inbound/rev/f41dc09ee34b

Comment 9

[Wes Kocher (:KWierso) (Not reading bugmail; email directly if needed)]

https://hg.mozilla.org/mozilla-central/rev/f41dc09ee34b

Comment 10

[Ryan VanderMeulen [:RyanVM]]

https://hg.mozilla.org/releases/mozilla-aurora/rev/61f812b1c934