CQ matches waterfall: is Blink in scope?

[Paweł Hajdan, Jr.]

I'd like to build consensus about scope of "CQ matches main waterfall" CY pillar.

We have bugs like https://code.google.com/p/chromium/issues/detail?id=415063 (that one is about Oilpan coverage on blink CQ).

This seems relatively easy to do (not sure about capacity), but I'm not sure if we've treated Blink as in scope for this CY pillar. The bug currently has Infra-CodeYellow label but is not blocking the tracking bug https://code.google.com/p/chromium/issues/detail?id=380709 .

Since there are plans to merge Blink and Chromium repos (which will mean merging the waterfalls, CQs and trybots if I understand correctly), I suggest that we consider blink CQ <-> chromium.webkit match in scope for CY.

I'm totally open to your comments on this, what I said above is not a strong opinion.

Paweł

[Dirk Pranke]

I don't believe we consider Blink part of the CY at all, so, no.

That said, any changes that improve Blink developer productivity are still welcome :).

--
You received this message because you are subscribed to the Google Groups "Chromium Hackability Code Yellow" group.
To unsubscribe from this group and stop receiving emails from it, send an email to hackability-c...@chromium.org.
To post to this group, send email to hackabi...@chromium.org.
To view this discussion on the web visit https://groups.google.com/a/chromium.org/d/msgid/hackability-cy/CAATLsPaSS%3Dog%2Bq2Vcuba_chwzrku-Lo%3DSH19dYfVvOsjsGARqw%40mail.gmail.com.

[Julie Parent]

The Code Yellow is specific to the Chromium tree, so, metrics for all pillars are explicitly defined for chromium, not blink (Blink was in a much better state the Chromium when we started).

That said, if the Blink merge happens before the exit of the code yellow, then blink will directly impact chromium metrics and thus become in scope, and, as Dirk said, all changes to Blink are certainly most welcome.

To view this discussion on the web visit https://groups.google.com/a/chromium.org/d/msgid/hackability-cy/CAEoffTADdHm18DKrOfJOrNKK3GD4he1s3AwW16y8vGJJPZrYrg%40mail.gmail.com.

[Eric Seidel]

My high-level understanding is the biggest remaining blocker for the
Blink merge is being able to run all Blink changes through the
Chromium CQ. So related to all this (but not an explicit part of the
CY).

> https://groups.google.com/a/chromium.org/d/msgid/hackability-cy/CAPSmAARK9LLOysoGu_N5xmjOX7Ex5tACbti4NgRj%2Bnpi36w-NA%40mail.gmail.com.

[Julie Parent]

Correct.  Which is blocked on swarming capacity, which we are following up on now.  Within a few days, we'll have estimates of when we expect to have the capacity to handle the additional load from blink, and I will update this (and related) threads on the matter.

[Ojan Vafai]

I don't think this is in scope for the CY, but if anything were, it would only be the bots that would survive after the blink merge, which would be the "WebKit" prefixed bots that don't have "(deps)" in their names.

FWIW, addressing this would be a big benefit to blink development.

Here are the bots. The ones with a star already have CQ coverage. For the ones without a star, they do *not* need to be in the default CQ list. So we would just need a single non-default CQ trybot.

WebKit XP

*WebKit Win7

WebKit Mac10.6

WebKit Mac10.7

WebKit Mac10.8

WebKit Mac10.8 (retina)

*WebKit Mac10.9

*WebKit Linux

WebKit Linux 32

WebKit Android (Nexus4)

There's also the following two. I'm not sure what the status of them is and whether they would need trybots.

WebKit Linux ASAN

WebKit Linux Leak

To view this discussion on the web visit https://groups.google.com/a/chromium.org/d/msgid/hackability-cy/CAPSmAAS%3DX2J_S9k6%2BSzcqDE3sfv5CYKi5Sf5pj5-mUnOBNV4CQ%40mail.gmail.com.

[Paweł Hajdan, Jr.]

Just checking: I'd like to remove the Infra-CodeYellow label from the bug (https://code.google.com/p/chromium/issues/detail?id=415063). Does that sound OK to everyone?

Paweł

[Ojan Vafai]

lgtm

[Dirk Pranke]

Not only does that LGTM, but I've even removed the label for you :). There's no world in which Oilpan is a CY issue ...

-- Dirk