1056199 - update talos to the latest revision

Status: RESOLVED FIXED

(Testing :: Talos, defect)

Comment 1

[Joel Maher ( :jmaher ) (UTC -8)]

Attachment: update-talos.json

getting all the ducks in a row

Comment 2

[Joel Maher ( :jmaher ) (UTC -8)]

Attachment: and the latest version with http cache2 pref changed (2.0)

Comment 3

[Ryan VanderMeulen [:RyanVM]]

https://hg.mozilla.org/integration/mozilla-inbound/rev/39426d83e4bb

Comment 4

[Ryan VanderMeulen [:RyanVM]]

https://hg.mozilla.org/mozilla-central/rev/39426d83e4bb

Comment 5

[Ryan VanderMeulen [:RyanVM]]

Aurora and Beta are both currently on talos rev b826a38ef31f. Looking at the pushlog range between that rev and tip, I think we should just go ahead and take the same rev we just landed on m-c. There's some nice fixes we'd like to get backported as far as possible anyway if they don't break anything.
https://hg.mozilla.org/build/talos/pushloghtml?fromchange=b826a38ef31f&tochange=dea55d0aff4d

Aurora: https://tbpl.mozilla.org/?tree=Try&rev=b8922df8fe87
Beta:   https://tbpl.mozilla.org/?tree=Try&rev=98fa5c68a946

Comment 6

[Ryan VanderMeulen [:RyanVM]]

(In reply to Ryan VanderMeulen [:RyanVM UTC-4] from comment #5)
> Aurora: https://tbpl.mozilla.org/?tree=Try&rev=b8922df8fe87
> Beta:   https://tbpl.mozilla.org/?tree=Try&rev=98fa5c68a946

Green across the board \m/. I assume you'll want to look more closely at the results to make sure nothing's out of whack, though, Joel. Let me know if you're happy with the results and I'll go ahead and push :)

Comment 7

[Joel Maher ( :jmaher ) (UTC -8)]

this will be hard to compare the results as we only build PGO on aurora/beta, this try push doesn't build in PGO mode, but for osx which doesn't have pgo, nothing crazy stands out on aurora, but beta looks pretty scary :(

Comment 8

[Ryan VanderMeulen [:RyanVM]]

Discussed on IRC with Joel. Given the time constraints we're up against, we decided to go ahead and push this to both to get the data ASAP. If there's a real perf regression from this update, it's almost certainly pointing to a real previously-masked regression that we'd be shipping to users without knowing otherwise. The sooner we know if that's the case, the sooner we can decide whether cache2 should ship in Fx32 or not.

https://hg.mozilla.org/releases/mozilla-aurora/rev/6926573971cf
https://hg.mozilla.org/releases/mozilla-beta/rev/ec3e586813b5