Stream: t-compiler/wg-prioritization/alerts

Topic: #81531 internal compiler error: failed to process buffered …


triagebot (Jan 29 2021 at 19:47, on Zulip):

@WG-prioritization/alerts issue #81531 has been requested for prioritization.

Procedure

Camelid (Jan 29 2021 at 19:52, on Zulip):

Maybe worth waiting for more info, but seems like P-high to me.

Camelid (Feb 01 2021 at 03:59, on Zulip):

We have more info now. I think P-high:

macro_rules! foo {
    () => { true; }
}

async fn func() {
    let _ = foo!();
}

fn main() {}
rylev (Feb 01 2021 at 12:34, on Zulip):

Is this any use of a macro inside of an async function?

Stu (Feb 02 2021 at 10:16, on Zulip):

Aaron1011 says this is fixed by #81541.

apiraino (Feb 03 2021 at 14:16, on Zulip):

uhm ... the issue has been reopened, I wonder if it's because #81541 wasn't enough or because the commenter tried too soon on a nightly (but maybe I'm wrong here).

How long before a commit on the master branch ends up into the first nightly?

apiraino (Feb 03 2021 at 14:18, on Zulip):

I mean, this is what I understand

apiraino (Feb 03 2021 at 14:18, on Zulip):

screenshot-20210203-151833.png

oliver (Feb 03 2021 at 14:21, on Zulip):

I think it's outlined here: https://forge.rust-lang.org/release/process.html
just reading it now myself

oliver (Feb 03 2021 at 14:21, on Zulip):

I thought nightly was updated in early evening PST daily

oliver (Feb 03 2021 at 14:23, on Zulip):

there are timestamps here: https://static.rust-lang.org/dist/2021-02-02

apiraino (Feb 03 2021 at 14:27, on Zulip):

uhm ... ok so in theory the comment of 7 hours ago (I think 07:30 PST) should refer to a nightly that has that fix

Anyway, by judging on the number of issues piling up on this, I'd assign a P-high and tomorrow let's try to understand if #81541 was enough or if there's something else behind

oliver (Feb 03 2021 at 14:29, on Zulip):

it depends on when the merged PR is rolled up

oliver (Feb 03 2021 at 14:30, on Zulip):

https://forge.rust-lang.org/release/rollups.html
'every pull request must be tested after merge before it can be pushed to master'

oliver (Feb 03 2021 at 14:31, on Zulip):

actually the way it's worded the CI testing is for non-rolledup PRs

oliver (Feb 03 2021 at 14:32, on Zulip):

https://forge.rust-lang.org/infra/docs/rustc-ci.html#merging-prs-serially-with-bors
'long CI (~3.5 hours for a full run), means we can’t merge too many PRs in a single day, and a single failure greatly impacts our throughput for the day. The maximum number of PRs we can merge in a day is 7.'

apiraino (Feb 03 2021 at 14:36, on Zulip):

oliver said:

actually the way it's worded the CI testing is for non-rolledup PRs

uhm ... I'm a still a bit confused :-) As I don't see any mention of "rollups" in that issue, can I assume that the PR went into master after its CI finished?

apiraino (Feb 03 2021 at 14:39, on Zulip):

oh, closed again a few minutes ago

oliver (Feb 03 2021 at 14:42, on Zulip):

it looks like you committed directly to master, as I understand it the CI job is then arranged
in the queue before somehow being bundled into the next nightly release which happens
every evening PST

oliver (Feb 03 2021 at 14:44, on Zulip):

it might be interesting to have an untested image of master released though that
can be obtained by building from source locally

apiraino (Feb 03 2021 at 14:51, on Zulip):

On godbolt the nightly is currently rustc 1.51.0-nightly (0e63af5da 2021-01-31). I've tried to pull the latest nightly using rustup but I get and even older one (04caa632d 2021-01-30).

I wonder how the commenter could have a newest nightly I can't get

apiraino (Feb 03 2021 at 14:52, on Zulip):

(I hope I'm not adding noise to the discussion, just trying to figure this out)

oliver (Feb 03 2021 at 14:52, on Zulip):

from here: https://static.rust-lang.org/dist/2021-01-20

apiraino (Feb 03 2021 at 14:53, on Zulip):

oh

oliver (Feb 03 2021 at 14:53, on Zulip):

or building locally from the commit 0e63af5da

oliver (Feb 03 2021 at 14:54, on Zulip):

but I'm not clear on what's the interaction with CI if it was built from source

oliver (Feb 03 2021 at 14:55, on Zulip):

I would assume the dist-releases have been CI tested

apiraino (Feb 03 2021 at 14:56, on Zulip):

those looks to me like simple source tarballs, not even landed on nightly (but I really don't know)

Stu (Feb 03 2021 at 14:56, on Zulip):

I've just tried it with rustc 1.51.0-nightly (368275062 2021-02-02) and the last example that was send, after closing the issue the first time, still reproduces

apiraino (Feb 03 2021 at 14:57, on Zulip):

hmm ... I see

Stu (Feb 03 2021 at 14:58, on Zulip):

I will try the new commit that closed it again

triagebot (Feb 03 2021 at 16:52, on Zulip):

Issue #81531's prioritization request has been removed.

Camelid (Feb 03 2021 at 16:53, on Zulip):

(Issue closed)

Camelid (Feb 03 2021 at 16:53, on Zulip):

This appears to be fixed (fully) now.

apiraino (Feb 03 2021 at 17:47, on Zulip):

Camelid said:

This appears to be fixed (fully) now.

I'm not 101% sure, but someone will certainly speak up if not :-)

apiraino (Feb 03 2021 at 17:51, on Zulip):

(to clarify, not 101% sure because issue reporters had not yet had a chance to test)

Last update: Apr 16 2021 at 23:15UTC