Stream: t-compiler/wg-prioritization/alerts

Topic: I-prioritize #76785 350MB-long type name regression in 1.46


triagebot (Sep 16 2020 at 16:07, on Zulip):

@WG-prioritization/alerts issue #76785 has been requested for prioritization.

Procedure

Joshua Nelson (Sep 16 2020 at 16:09, on Zulip):

issues like this have been coming up a lot lately, maybe they should be combined into one?

DPC (Sep 16 2020 at 16:10, on Zulip):

seperate issues are fine imo, sometimes they get solved at different times

DPC (Sep 16 2020 at 16:10, on Zulip):

unless it is an exact duplicate

Mason Stallmo (Sep 16 2020 at 16:26, on Zulip):

Would this end up being something like a P-Medium since it's a stable to stable regression but has a workaround?

Joshua Nelson (Sep 16 2020 at 16:27, on Zulip):

I think I misunderstood the issue, it's a diagnostics regression

Joshua Nelson (Sep 16 2020 at 16:28, on Zulip):

and it has a fix that's almost ready too

Joshua Nelson (Sep 16 2020 at 16:28, on Zulip):

so I'd go with P-low maybe? since it's a diagnostic regression

Stu (Sep 16 2020 at 16:36, on Zulip):

I don't understand.
How is this a diagnostic regression?

Joshua Nelson (Sep 16 2020 at 16:37, on Zulip):

the bug is not that the limit decreased, but that it doesn't show why the limit decreased

Stu (Sep 16 2020 at 16:37, on Zulip):

Ah I see, that makes sense.

Joshua Nelson (Sep 16 2020 at 16:38, on Zulip):

or maybe it should count as tracking both :shrug: but in that case there are other issues it duplicates

LeSeulArtichaut (Sep 16 2020 at 17:29, on Zulip):

Well I believe that the compiler team (e.g. the reviewer of the culprit of the regression) could decide whether it is right to have a lowered limit

Santiago Pastorino (Sep 16 2020 at 18:13, on Zulip):

this one is pending prioritization

Santiago Pastorino (Sep 16 2020 at 18:13, on Zulip):

I'd be fine with P-low or P-medium whatever to be honest

Santiago Pastorino (Sep 16 2020 at 18:13, on Zulip):

saw that @Joshua Nelson thinks P-low, should we go with it then? or others have different opinion

apiraino (Sep 16 2020 at 18:17, on Zulip):

does it impact code that's an touching a pretty edge case, if I am correct?

Joshua Nelson (Sep 16 2020 at 18:19, on Zulip):

I still think this is two separate issues

Joshua Nelson (Sep 16 2020 at 18:20, on Zulip):

so we should decide which bit we're prioritizing: the diagnostic problem, which is not a regression, or the type length limit regression

Joshua Nelson (Sep 16 2020 at 18:20, on Zulip):

apiraino said:

does it impact code that's an touching a pretty edge case, if I am correct?

not really, there are a lot of recent type length limit issues

Joshua Nelson (Sep 16 2020 at 18:20, on Zulip):

https://github.com/rust-lang/rust/issues?q=is%3Aissue+is%3Aopen+type+length+limit

apiraino (Sep 16 2020 at 18:20, on Zulip):

interesting :thinking:

Santiago Pastorino (Sep 16 2020 at 19:05, on Zulip):

@Joshua Nelson can you comment on the issue what you're saying

Santiago Pastorino (Sep 16 2020 at 19:05, on Zulip):

if they are 2 issues in one, I'd give the priority of the highest one :)

Joshua Nelson (Sep 16 2020 at 19:05, on Zulip):

sure

Santiago Pastorino (Sep 16 2020 at 19:06, on Zulip):

let me read the issue better :)

Santiago Pastorino (Sep 16 2020 at 19:07, on Zulip):

I guess they are talking about the limit thing and not exactly about the diagnostic

Joshua Nelson (Sep 16 2020 at 19:07, on Zulip):

done

Santiago Pastorino (Sep 16 2020 at 19:08, on Zulip):

ok, I'd go with P-medium in this case

Santiago Pastorino (Sep 16 2020 at 19:09, on Zulip):

well whatever :P, I'd agree with the first one that says something that is P-low or P-medium and would leave it up to them to tag it properly

apiraino (Sep 17 2020 at 07:54, on Zulip):

issue has been closed as dupe of #75992

Last update: Apr 16 2021 at 23:15UTC