Stream: t-compiler

Topic: weekly meeting 2019-09-05 #54818


nikomatsakis (Sep 05 2019 at 13:34, on Zulip):

Hi @T-compiler/meeting ; the triage meeting will be starting in approx 30 minutes

nikomatsakis (Sep 05 2019 at 13:35, on Zulip):

I will be doing pre-triage in a parallel topic

nikomatsakis (Sep 05 2019 at 13:36, on Zulip):

The check-in calendar has the PGO and pipelining groups up for check-in -- I think that @mw is not here this week, and PGO is mostly him. Similarly, I think that the pipelining effort is largely "done", at least for its initial goals, though I see that @nnethercote was doing some more experiments recently. So maybe we'll skip WG check-in this week.

nikomatsakis (Sep 05 2019 at 13:37, on Zulip):

Next week we have polonius + rfc 2229

nikomatsakis (Sep 05 2019 at 14:01, on Zulip):

Hello @T-compiler/meeting! The weekly meeting will be started now.

nikomatsakis (Sep 05 2019 at 14:02, on Zulip):

While people filter in, let's have any...

Announcements

nikomatsakis (Sep 05 2019 at 14:02, on Zulip):

On the pre-triage side, I only made it through the nominated issues. Mostly uneventful. One new p-high, though I'm not sure if it should be:

Rustc does display a correct error message on type missmatch but does not show line numbers #51635

nikomatsakis (Sep 05 2019 at 14:04, on Zulip):

One announcement is that we scheduled some design meetings. You'll find them on the compiler calendar.

nikomatsakis (Sep 05 2019 at 14:04, on Zulip):

Oh, I should update the README.md in compiler-team repo

nikomatsakis (Sep 05 2019 at 14:05, on Zulip):

Next Friday, Sep 13: [design meeting] rust-analyzer/libraryification (calendar event)

nikomatsakis (Sep 05 2019 at 14:05, on Zulip):

After that, Friday Sep 20: [design meeting] ICE-breaker groups (calendar event)

nikomatsakis (Sep 05 2019 at 14:06, on Zulip):

OK, let's get started.

nikomatsakis (Sep 05 2019 at 14:06, on Zulip):

As I noted above, we'll skip wg check-in this week, maybe the meeting will be a bit shorter

nikomatsakis (Sep 05 2019 at 14:07, on Zulip):

Beta regressions

nikomatsakis (Sep 05 2019 at 14:07, on Zulip):

I think I had hoped to visit some of these after meeting last time, not sure how far I got, there's kind of a lot, but I think many are expected

nikomatsakis (Sep 05 2019 at 14:07, on Zulip):

@Vadim Petrochenkov you filed a lot of these bugs, any you wish to call attention to?

nikomatsakis (Sep 05 2019 at 14:08, on Zulip):

I think there was general consensus that we should close " [1.38] Regression: custom attribute panicked #63895 " -- the tl;dr was that the proc macro is checking something based on the string output and that this changes

nikomatsakis (Sep 05 2019 at 14:08, on Zulip):

in this case, { } changed to {} or something

nikomatsakis (Sep 05 2019 at 14:09, on Zulip):

" [1.38] Regression: proc macro panicked #63894 " looks very similar

Vadim Petrochenkov (Sep 05 2019 at 14:10, on Zulip):

From the opened issues some have some have fixes pending or merged, others are mostly wontfix.

nikomatsakis (Sep 05 2019 at 14:11, on Zulip):

the next two ( [1.38] Regression: cannot determine resolution for the derive macro Debug #63893 ) and ( [1.38] Regression: expected module, found unresolved item crate::mod #63891 ) seem like "wontfix" but probably good to cc lang? basically subtle aspects of name resolution algorithm?

nikomatsakis (Sep 05 2019 at 14:11, on Zulip):

maybe [1.38] Regression: failed to resolve: could not find rustc_serialize in {{root}} #63885 is the same?

Vadim Petrochenkov (Sep 05 2019 at 14:11, on Zulip):

(The issues were opened for all regressions, just to "register" them in some way, regardless of likely wontfix status.)

Vadim Petrochenkov (Sep 05 2019 at 14:11, on Zulip):

https://github.com/rust-lang/rust/issues/63893 is lang.

Vadim Petrochenkov (Sep 05 2019 at 14:12, on Zulip):

https://github.com/rust-lang/rust/issues/63891 is an outright bug (Def::Err created without reporting an error).

nikomatsakis (Sep 05 2019 at 14:13, on Zulip):

ok

nikomatsakis (Sep 05 2019 at 14:13, on Zulip):

what about [1.38] Regression: failed to resolve: could not find rustc_serialize in {{root}} #63885 ?

centril (Sep 05 2019 at 14:14, on Zulip):

Probably no action required.

centril (Sep 05 2019 at 14:14, on Zulip):

Seems to just affect tests

Vadim Petrochenkov (Sep 05 2019 at 14:14, on Zulip):

This is a consequense of derive(RustcEncodable, RustcDecodable) being implemented both in the compiler and in a third party crate.

nikomatsakis (Sep 05 2019 at 14:15, on Zulip):

@Vadim Petrochenkov does this comment sound vaguely correct for #63885?

Vadim Petrochenkov (Sep 05 2019 at 14:15, on Zulip):

Only tests in those third-party crate (and its fork) are affected.

nikomatsakis (Sep 05 2019 at 14:15, on Zulip):

Seems to just affect tests

yeah I was more looking for a rough explanation of what happened

nikomatsakis (Sep 05 2019 at 14:15, on Zulip):

i.e., is it a "bug fix" or what

centril (Sep 05 2019 at 14:15, on Zulip):

This is a consequense of derive(RustcEncodable, RustcDecodable) being implemented both in the compiler and in a third party crate.

That makes it sound like the opposite of a bug fix? :P

centril (Sep 05 2019 at 14:16, on Zulip):

We should be working towards removing those two derives from the compiler imo and moving them to proc macros for rustc internals

Vadim Petrochenkov (Sep 05 2019 at 14:17, on Zulip):

This is not a bug-fix, but rather migration of derive(RustcEncodable, RustcDecodable) to 2018 edition.

nikomatsakis (Sep 05 2019 at 14:18, on Zulip):

ok, but it's kind of specific to the rustc-serialize crate itself I guess somehow?

nikomatsakis (Sep 05 2019 at 14:18, on Zulip):

If you are able to do a short write-up of the interaction, that'd be awesome. I can imagine though that this crate is in a rather...unique position.

nikomatsakis (Sep 05 2019 at 14:18, on Zulip):

I'd like to move on through our triage though

nikomatsakis (Sep 05 2019 at 14:19, on Zulip):

stable to nightly regressions

nikomatsakis (Sep 05 2019 at 14:19, on Zulip):

In this TiKV PR we upgraded the compiler. Our tidb_query component tripped the redundant_semicolon lint (and seemingly in turn clippy::no_effect) but rustc/clippy did not tell us the line number.

Vadim Petrochenkov (Sep 05 2019 at 14:19, on Zulip):

ok, but it's kind of specific to the rustc-serialize crate itself I guess somehow?

Yes, that's an issue specific to rustc-serialize itself.

nikomatsakis (Sep 05 2019 at 14:19, on Zulip):

@Esteban K├╝ber writes:

Introduced in #62984. It has caused similar effects in other cases like #63947. CC #63679.

centril (Sep 05 2019 at 14:20, on Zulip):

Seems like a good issue to write some mentoring instructions for

centril (Sep 05 2019 at 14:20, on Zulip):

I'll cc nathanwhit and varkor -- done

nikomatsakis (Sep 05 2019 at 14:21, on Zulip):

yeah; I guess we could just make the lint allow-by-default in the limit

nikomatsakis (Sep 05 2019 at 14:21, on Zulip):

I don't really know what the problem is

nikomatsakis (Sep 05 2019 at 14:21, on Zulip):

otoh I guess there's an easy enough workaround

centril (Sep 05 2019 at 14:21, on Zulip):

maybe forgot some span somewhere...?

nikomatsakis (Sep 05 2019 at 14:21, on Zulip):

presumably

nikomatsakis (Sep 05 2019 at 14:22, on Zulip):

I'm going to mark it as P-high I guess

nikomatsakis (Sep 05 2019 at 14:22, on Zulip):

it's a regression

nikomatsakis (Sep 05 2019 at 14:22, on Zulip):

and seems like it could be a fairly annoying one, if not overly serious

centril (Sep 05 2019 at 14:22, on Zulip):

sgtm

nikomatsakis (Sep 05 2019 at 14:23, on Zulip):

why did you cc @varkor ? were they mentoring or involved with the original PR?

nikomatsakis (Sep 05 2019 at 14:23, on Zulip):

I'd love to assign to someone :)

centril (Sep 05 2019 at 14:23, on Zulip):

were they mentoring or involved with the original PR?

indeed; they reviewed the PR

centril (Sep 05 2019 at 14:23, on Zulip):

and I cced the PR author

nikomatsakis (Sep 05 2019 at 14:24, on Zulip):

okey dokey

nikomatsakis (Sep 05 2019 at 14:24, on Zulip):

I wrote a comment

nikomatsakis (Sep 05 2019 at 14:24, on Zulip):

moving on

nikomatsakis (Sep 05 2019 at 14:24, on Zulip):

beta nominated PRs

nikomatsakis (Sep 05 2019 at 14:25, on Zulip):

we've got 4

nikomatsakis (Sep 05 2019 at 14:25, on Zulip):

Update xLTO compatibility table in rustc book. #64092

nikomatsakis (Sep 05 2019 at 14:25, on Zulip):

Support "soft" feature-gating using a lint #64066

nikomatsakis (Sep 05 2019 at 14:25, on Zulip):

fix nounwind attribute logic #63909

nikomatsakis (Sep 05 2019 at 14:25, on Zulip):

pprust: Do not print spaces before some tokens #63897

nikomatsakis (Sep 05 2019 at 14:26, on Zulip):

I think that #63909 we kind of decided not to do in lang team meeting? presuming we make progress on that RFC? (cc @Taylor Cramer, @Kyle Strand =))

centril (Sep 05 2019 at 14:27, on Zulip):

I am very confused as to what we decided on the nounwind thing

centril (Sep 05 2019 at 14:27, on Zulip):

I feel as tho we didn't reach a firm conclusion

centril (Sep 05 2019 at 14:27, on Zulip):

and there seems to be some back & forth on the #[unwind(...)] stuff after as well

centril (Sep 05 2019 at 14:28, on Zulip):

pprust: Do not print spaces before some tokens #63897

@nikomatsakis should the language team discuss this?

nikomatsakis (Sep 05 2019 at 14:28, on Zulip):

OK, those all seem like "no brainers" for the most part, so I'm marking them all as beta-accepted

centril (Sep 05 2019 at 14:28, on Zulip):

there seems to be a larger question of pretty printing policy here

centril (Sep 05 2019 at 14:29, on Zulip):

aside from this particular instance

nikomatsakis (Sep 05 2019 at 14:29, on Zulip):

I think we should both accept the PR and we can discuss the general issue

nikomatsakis (Sep 05 2019 at 14:29, on Zulip):

that is, I don't think we need to block the PR

nikomatsakis (Sep 05 2019 at 14:29, on Zulip):

but it's a good topic to bring up anyway and make people aware of

nikomatsakis (Sep 05 2019 at 14:29, on Zulip):

I think I actually created a doc for this week but faile toa dd it to the page, we can add it to the agenda...

centril (Sep 05 2019 at 14:29, on Zulip):

agreed; let's nominate this specific PR as a proxy for the general issue?

nikomatsakis (Sep 05 2019 at 14:30, on Zulip):

sure.

nikomatsakis (Sep 05 2019 at 14:30, on Zulip):

(also, I agree with you that we should revisit the unwind thing)

nikomatsakis (Sep 05 2019 at 14:30, on Zulip):

ok, that's the betas

nikomatsakis (Sep 05 2019 at 14:32, on Zulip):

since we don't have working group check-ins, the only thing left would be to review P-high bugs, but I'm not inclined to do that -- I'd like to maybe schedule a more dedicated time to review that, but next week, once @pnkfelix is back (and @mw is around)

nikomatsakis (Sep 05 2019 at 14:32, on Zulip):

make sense?

nikomatsakis (Sep 05 2019 at 14:32, on Zulip):

if so, any final topics to discuss? or shall we end meeting early?

nikomatsakis (Sep 05 2019 at 14:32, on Zulip):

(also, did I overlook something...)

nikomatsakis (Sep 05 2019 at 14:34, on Zulip):

OK :wave: thanks all for attending! :heart_eyes:

Last update: Nov 22 2019 at 04:30UTC