Stream: t-compiler

Topic: RLS 1.0 triage and maintenance


nikomatsakis (Mar 01 2019 at 21:27, on Zulip):

So, I think that we should be tracking RLS 1.0 bugs like we would any other compiler bug. This is a probably just a specific example of us wanting to expand our triage system so that it can cover more than just the rust-lang/rust crate.

I'm not sure exactly what is needed to make that happen, presumably establishing compatible labels, and using the nomination / P-high system? Are there people doing regular triage (e.g., release team or what have you, cc @centril / @simulacrum) of the RLS bugs?

cc @pnkfelix and @Igor Matuszewski

simulacrum (Mar 01 2019 at 21:27, on Zulip):

Where are RLS bugs?

nikomatsakis (Mar 01 2019 at 21:32, on Zulip):

Not sure but I opened up https://github.com/rust-lang/rust/issues/58858 to help us track this

pnkfelix (Mar 04 2019 at 10:48, on Zulip):

@simulacrum at the very least there is a A-rls label on github

pnkfelix (Mar 04 2019 at 10:48, on Zulip):

but it clearly hasn't seen much use. Which is great because we can decide how we want to use it ourselves

Igor Matuszewski (Mar 04 2019 at 21:16, on Zulip):

if it's any help, I completed a thorough issue triage over at https://github.com/rust-lang/rls/issues

Igor Matuszewski (Mar 04 2019 at 21:18, on Zulip):

beside bug reports, I believe the reported issues boil down to: better/smarter project detection (do something smarter than just do a cargo check equivalent), macro-generated/related identifiers not working and latency/architecture improvement around calling rustc

Igor Matuszewski (Mar 04 2019 at 21:18, on Zulip):

and possibly some LSP features we should ideally catch up with

nikomatsakis (Mar 08 2019 at 17:03, on Zulip):

This is awesome @Igor Matuszewski -- did you adjust labels and things to represent this grouping?

nikomatsakis (Mar 08 2019 at 17:04, on Zulip):

What steps do we need to ensure that triage continues. I think we need at least:

nikomatsakis (Mar 08 2019 at 17:04, on Zulip):

One nice thing about the T-compiler label is that it is an easy way to see "untriaged" bug reports I guess

Igor Matuszewski (Mar 09 2019 at 12:29, on Zulip):

@nikomatsakis I mostly focused on the enhancement/bug split, there’s also „protocol” label for the LSP related stuff

Igor Matuszewski (Mar 09 2019 at 12:29, on Zulip):

We also have a P-high label used but unsure if we can somehow make an issue query across repositories

Igor Matuszewski (Mar 09 2019 at 12:34, on Zulip):

I can go over them again and group them accordingly (project layout, latency etc.)

nikomatsakis (Mar 11 2019 at 19:10, on Zulip):

We also have a P-high label used but unsure if we can somehow make an issue query across repositories

github can do issue searches across repositories if they are all in the same github org

Last update: Nov 16 2019 at 01:05UTC