Stream: t-compiler

Topic: landing invariance fix #71896


nikomatsakis (May 06 2020 at 14:33, on Zulip):

So @Santiago Pastorino opened https://github.com/rust-lang/rust/pull/71896 which fixes #71550. I'm wondering if we should just land this?

nikomatsakis (May 06 2020 at 14:33, on Zulip):

Or if we want to do a crater run

nikomatsakis (May 06 2020 at 14:33, on Zulip):

I'm a bit worried it'll break random crates in the ecosystem

nikomatsakis (May 06 2020 at 14:34, on Zulip):

and I don't know how to do a warning period, really

nikomatsakis (May 06 2020 at 14:34, on Zulip):

on the other hand, the crater queue is looking long

nikomatsakis (May 06 2020 at 14:34, on Zulip):

not sure who to cc on this topic ;)

Pietro Albini (May 06 2020 at 14:34, on Zulip):

we might get more crater capacity soon

Pietro Albini (May 06 2020 at 14:35, on Zulip):

(yes, I'm getting notified when someone types crater on zulip :D)

nikomatsakis (May 06 2020 at 14:35, on Zulip):

are we doing regular beta/nightly runs on some cadence?

Pietro Albini (May 06 2020 at 14:35, on Zulip):

we do a beta run at the start of each cycle

Pietro Albini (May 06 2020 at 14:35, on Zulip):

the 1.44.0 already finished

nikomatsakis (May 06 2020 at 14:37, on Zulip):

OK.

Pietro Albini (May 06 2020 at 14:40, on Zulip):

@nikomatsakis btw, currently a check run takes ~2 days, so in 10 days your run should start

Pietro Albini (May 06 2020 at 14:40, on Zulip):

we might get more capacity sooner than that too

nikomatsakis (May 06 2020 at 14:40, on Zulip):

I'm somewhat inclined to "just land it" and see what happens

nikomatsakis (May 06 2020 at 14:40, on Zulip):

But I wonder if that's irresponsible of me

nikomatsakis (May 06 2020 at 14:40, on Zulip):

We don't have a totally clear policy on this I guess

Pietro Albini (May 06 2020 at 14:41, on Zulip):

with my release hat I'd prefer to avoid landing it on nightly if we don't know the extent of the breakage

simulacrum (May 06 2020 at 14:43, on Zulip):

I'd be down for prioritizing the crater queue a bit

simulacrum (May 06 2020 at 14:43, on Zulip):

some of it is probably less important

nikomatsakis (May 06 2020 at 14:45, on Zulip):

I was wondering about that

nikomatsakis (May 06 2020 at 14:45, on Zulip):

maybe p=1

nikomatsakis (May 06 2020 at 14:45, on Zulip):

for this PR

simulacrum (May 06 2020 at 14:45, on Zulip):

yeah that seems reasonable to me

simulacrum (May 06 2020 at 14:45, on Zulip):

We could also do some hacky group testing

nikomatsakis (May 06 2020 at 14:45, on Zulip):

I guess to start I will @bors try

simulacrum (May 06 2020 at 14:45, on Zulip):

I think I just did

nikomatsakis (May 06 2020 at 14:45, on Zulip):

Oh, so you did

simulacrum (May 06 2020 at 14:47, on Zulip):

yeah let's p=1 it, then we should have results by mid-weekend

Santiago Pastorino (May 06 2020 at 15:03, on Zulip):

hey, just saw this, all you have said sounds reasonable :+1:

nikomatsakis (Jun 02 2020 at 16:00, on Zulip):

So we didn't reach a decision here

nikomatsakis (Jun 02 2020 at 16:01, on Zulip):

I propose we land it at the start of the next release cycle

nikomatsakis (Jun 02 2020 at 16:01, on Zulip):

I'm not sure if there is anything else we should do before-hand -- reaching out to crater authors for example? It is probably a good idea to at least ping them

nikomatsakis (Jun 02 2020 at 16:02, on Zulip):

cc @pnkfelix :point_up:

Last update: Jun 04 2020 at 18:30UTC