Stream: t-compiler/wg-prioritization

Topic: I-prioritize #70924 'rustc' panicked at 'failed to lookup `S


triagebot (May 07 2020 at 09:53, on Zulip):

@WG-prioritization issue #70924 has been requested for prioritization.

o0Ignition0o - Jeremy Lempereur (May 07 2020 at 11:48, on Zulip):

it seems like cargo clean fixed it though
Ok so it seems to break build caches, but cargo clean is a workaround.

o0Ignition0o - Jeremy Lempereur (May 07 2020 at 11:51, on Zulip):

I feel like this kind of regressions could be caught automatically.
Anyway, the issue is not critical as in "nobody can use rust anymore" but it will probably cause a lot of efforts for project maintainers to figure out what happened, and quite a bit of compute minutes to fully clean their CI caches.
I'd go for P-High or P-Medium

LeSeulArtichaut (May 07 2020 at 11:52, on Zulip):

Mhh, it would be great to bisect this

LeSeulArtichaut (May 07 2020 at 11:52, on Zulip):

@bjorn3 Is your bisect bot available?

LeSeulArtichaut (May 07 2020 at 11:53, on Zulip):

If this is a beta regression, I’d say P-high to try to fix this before it hits stable

Santiago Pastorino (May 07 2020 at 13:34, on Zulip):

I'd say that this is at least P-high

Santiago Pastorino (May 07 2020 at 13:34, on Zulip):

seems like a bunch of people are hitting this issue

Santiago Pastorino (May 07 2020 at 13:39, on Zulip):

o0Ignition0o - Jeremy Lempereur said:

I feel like this kind of regressions could be caught automatically.
Anyway, the issue is not critical as in "nobody can use rust anymore" but it will probably cause a lot of efforts for project maintainers to figure out what happened, and quite a bit of compute minutes to fully clean their CI caches.
I'd go for P-High or P-Medium

I have a different feeling, this seems to be hitting a lot of codebases, yeah there's a workaround as probably are workarounds for everything but in my opinion this is P-high or P-critical, I'm leaning a bit more towards P-critical

o0Ignition0o - Jeremy Lempereur (May 07 2020 at 13:57, on Zulip):

ok, that would make sense as well yeah

LeSeulArtichaut (May 07 2020 at 19:08, on Zulip):

Even if there is a workaround, it’s still frustrating to have to delete the cache and do a full compilation, and I guess this is especially true in Rust

Santiago Pastorino (May 08 2020 at 13:23, on Zulip):

but also people may not know they need to delete the cache

Santiago Pastorino (May 08 2020 at 13:23, on Zulip):

I think it's good to wait a bit until we see a repro before prioritizing this one

Santiago Pastorino (May 08 2020 at 13:24, on Zulip):

or we could even give kind of a temporary priority until is clear, whatever the rest prefer is fine to me

LeSeulArtichaut (May 08 2020 at 15:06, on Zulip):

I believe it’s a good idea to assign a high priority to issues that need bisection/MCVE

LeSeulArtichaut (May 08 2020 at 15:07, on Zulip):

I think it might help give those issues the visibility they need

simulacrum (May 08 2020 at 15:07, on Zulip):

I think we should put a high priority after bisection/MCVE, ideally that's not something we'd push compiler team members to do I think

LeSeulArtichaut (May 08 2020 at 15:07, on Zulip):

But I might be wrong, I don’t know if many people look for the high-priority issues first

simulacrum (May 08 2020 at 15:08, on Zulip):

but obviously if e.g. that's not happening then yeah

Santiago Pastorino (May 11 2020 at 20:25, on Zulip):

this one has now a repro

Santiago Pastorino (May 11 2020 at 20:28, on Zulip):

I'm still debating between P-critical and P-high

Santiago Pastorino (May 13 2020 at 16:10, on Zulip):

this is the last one without priority, assigning P-high given that seems to be more consensus on that

triagebot (May 13 2020 at 16:10, on Zulip):

Issue #70924's prioritization request has been removed.

Last update: Jun 05 2020 at 22:05UTC