Implement RFC 2056 trivial constraints in where clauses#48557

matthewjasper · 2018-02-26T21:32:24Z

This is an implementation of the new behaviour for #48214. Tests are mostly updated to show the effects of this. Feature gate hasn't been added yet.

Some things that are worth noting and are maybe not want we want

&mut T: Copy doesn't allow as much as someone might expect because there is often an implicit reborrow.
~~There isn't a check that a where clause is well-formed any more, so where Vec<str>: Debug is now allowed (without a str: Sized bound).~~

r? @nikomatsakis

rust-highfive · 2018-02-26T21:32:35Z

Thanks for the pull request, and welcome! The Rust team is excited to review your changes, and you should hear from @nikomatsakis (or someone else) soon.

If any changes to this PR are deemed necessary, please add them as extra commits. This ensures that the reviewer can see what has changed since they last reviewed the code. Due to the way GitHub handles out-of-date commits, this should also make it reasonably obvious what issues have or haven't been addressed. Large or tricky changes may require several passes of review and changes.

Please see the contribution instructions for more information.

bors · 2018-02-27T03:25:29Z

☔ The latest upstream changes (presumably #48449) made this pull request unmergeable. Please resolve the merge conflicts.

cramertj · 2018-02-27T07:48:00Z

This is so cool! :)

shepmaster · 2018-03-04T15:49:33Z

trivial constaints

Typo. You may want to double check any other places for the same one.

nikomatsakis

D'oh! I had this review with a few initial questions for some time but never submitted it! Sorry about that. I've been wondering why you didn't respond ;)

nikomatsakis · 2018-02-28T18:55:57Z

src/librustc/infer/mod.rs

Can you clarify why you added this param_env.is_global() check?

Hmm, looks like I added this before I put in the changes to ParamEnv::and from #48411. It doesn't appear to be needed now.

Ok, let's remove then =)

nikomatsakis · 2018-02-28T18:56:47Z

src/librustc/traits/mod.rs

Also, can you clarify how you avoid the soundness problems discussed here =)

Is that related to the global check?

I couldn't create any soundness problems after removing this, so I can't really be sure that I have avoided it. See the tests for an idea of what I attempted.

It looks like the current caching is (at the time of making this PR) very conservative, which probably explains this though.

rust/src/librustc/traits/select.rs

Lines 1271 to 1282 in 3edb3cc

fn can_use_global_caches(&self, param_env: ty::ParamEnv<'tcx>) -> bool {

// If there are any where-clauses in scope, then we always use

// a cache local to this particular scope. Otherwise, we

// switch to a global cache. We used to try and draw

// finer-grained distinctions, but that led to a serious of

// annoying and weird bugs like #22019 and #18290. This simple

// rule seems to be pretty clearly safe and also still retains

// a very high hit rate (~95% when compiling rustc).

if !param_env.caller_bounds.is_empty() {

return false;

}

nikomatsakis · 2018-03-13T19:01:03Z

src/librustc/traits/select.rs

Well, I guess this is part of how we avoid the soundness problems too. I am trying to decide if this is_global check on the param environment suffices. I think it probably does given the current definitions of elaboration, but won't once we have a better definition of implied bounds (but that will require the newer solver, in which this PR is sort of easier to implement anyway).

I am imagining cases like this:

trait Foo where u32: Bar { } trait Bar { }

Now, if we were to fully elaborate, T: Foo would imply u32: Bar.

Well, I guess this is part of how we avoid the soundness problems too.

Yes, but this only seemed to cause bounds to not apply in my testing.

I am imagining cases like this:

trait Foo where u32: Bar { } trait Bar { }

Now, if we were to fully elaborate, T: Foo would imply u32: Bar.

True, but this is surely just as much a problem with fully elaborating

trait Foo where Vec<T>: Copy { } trait Bar { }

nikomatsakis · 2018-03-13T19:03:04Z

src/librustc/traits/select.rs

This is interesting. Should we remove Never from the enum? (Did you do so? If so, I missed it.)

nikomatsakis · 2018-03-13T19:07:49Z

src/librustc/ty/mod.rs

I just landed a change much like this.

That's where it's from. 😄

nikomatsakis · 2018-03-15T18:33:21Z

@matthewjasper so -- much as this PR somehow makes me mildly nervous -- I can't think of a reason this won't work =) I suggest you rebase and remove the WIP comment from the header, so I can r+ ...

matthewjasper · 2018-03-15T21:55:51Z

rebased

A few further changes:

wfcheck changes reverted, it appears these aren't needed to allow inconsistent bonds and any further changes are probably be beyond the scope of the RFC.
The trivial_bounds lint from the RFC is now implemented. I don't really like the name, so suggestions are welcome.
No feature gate, I'm not sure really how best to add one with these changes.

@nikomatsakis Done, but this still needs a feature gate, right?

nikomatsakis · 2018-03-16T15:43:06Z

@matthewjasper

Done, but this still needs a feature gate, right?

Um, yes, right. I was.. uh.. just testing you to see if you'd remember.

nikomatsakis · 2018-03-20T17:24:43Z

@matthewjasper let me know when you add feature gate then =)

matthewjasper · 2018-03-21T22:50:21Z

@nikomatsakis Feature gate now added.
The feature gate is added as another check in wfcheck so that the feature gate errors can have a custom note.

I haven't feature-gated built-in trait bound errors (e.g. where [T]: Sized), because they are a strange exception, and Clone only recently became a built-in trait.

The commits are now a bit of a mess, but I can clean them up later.

bors · 2018-03-22T20:01:56Z

☔ The latest upstream changes (presumably #49264) made this pull request unmergeable. Please resolve the merge conflicts.

nikomatsakis · 2018-03-23T17:05:38Z

@matthewjasper thanks -- sorry for being slow, will re-review

nikomatsakis · 2018-05-15T18:20:15Z

src/librustc_typeck/check/wfcheck.rs

Nit: behavoir

Nit: behavior

nikomatsakis

r=me upon rebase

matthewjasper · 2018-05-15T20:54:19Z

rebased.
And another soundness issue raised for good measure.

cramertj · 2018-05-15T20:55:10Z

@bors r=nikomatsakis

bors · 2018-05-15T20:55:10Z

📌 Commit be2900c has been approved by nikomatsakis

bors · 2018-05-16T09:03:45Z

⌛ Testing commit be2900c with merge 448cc57...

@nikomatsakis

Implement RFC 2056 trivial constraints in where clauses This is an implementation of the new behaviour for #48214. Tests are mostly updated to show the effects of this. Feature gate hasn't been added yet. Some things that are worth noting and are maybe not want we want * `&mut T: Copy` doesn't allow as much as someone might expect because there is often an implicit reborrow. * ~There isn't a check that a where clause is well-formed any more, so `where Vec<str>: Debug` is now allowed (without a `str: Sized` bound).~ r? @nikomatsakis

bors · 2018-05-16T11:18:00Z

☀️ Test successful - status-appveyor, status-travis
Approved by: nikomatsakis
Pushing 448cc57 to master...

@nikomatsakis

Prevent main from having a where clause. Closes #50714 Should this have a crater run? cc #48557, #48214 r? @nikomatsakis

rust-highfive assigned nikomatsakis Feb 26, 2018

rust-highfive added the S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. label Feb 26, 2018

bors added S-waiting-on-author Status: This is awaiting some action (such as code changes or more information) from the author. and removed S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. labels Feb 27, 2018

matthewjasper changed the title ~~[WIP] Implement RFC 2056 trivial constaints in where clauses~~ [WIP] Implement RFC 2056 trivial constraints in where clauses Mar 4, 2018

nikomatsakis reviewed Mar 9, 2018

View reviewed changes

nikomatsakis added S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. and removed S-waiting-on-author Status: This is awaiting some action (such as code changes or more information) from the author. labels Mar 12, 2018

nikomatsakis reviewed Mar 13, 2018

View reviewed changes

nikomatsakis added S-waiting-on-author Status: This is awaiting some action (such as code changes or more information) from the author. and removed S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. labels Mar 15, 2018

matthewjasper force-pushed the allow-trvial-bounds branch 2 times, most recently from c03bacd to b080ee2 Compare March 15, 2018 21:55

matthewjasper changed the title ~~[WIP] Implement RFC 2056 trivial constraints in where clauses~~ Implement RFC 2056 trivial constraints in where clauses Mar 15, 2018

matthewjasper force-pushed the allow-trvial-bounds branch 2 times, most recently from 760cec8 to fcfbfe0 Compare March 21, 2018 22:35

matthewjasper force-pushed the allow-trvial-bounds branch from 8ce2022 to 1c3854a Compare March 26, 2018 19:14

matthewjasper added 4 commits May 15, 2018 11:43

Feature gate trivial bounds

27183a9

Add trivial bounds lint

dabb820

Add tests for trivial bounds

9b0dfe1

Update existing tests for trivial bounds changes

dc8ce4c

nikomatsakis reviewed May 15, 2018

View reviewed changes

nikomatsakis approved these changes May 15, 2018

View reviewed changes

matthewjasper force-pushed the allow-trvial-bounds branch 2 times, most recently from 40eb071 to ad37f44 Compare May 15, 2018 20:48

Make is_global true for latebound regions

be2900c

matthewjasper force-pushed the allow-trvial-bounds branch from ad37f44 to be2900c Compare May 15, 2018 20:48

This was referenced May 15, 2018

Trait objects can call nonexistent concrete methods #50781

Closed

Prevent main from having a where clause. #50782

Merged

bors added S-waiting-on-bors Status: Waiting on bors to run and complete tests. Bors will change the label on completion. and removed S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. labels May 15, 2018

bors merged commit be2900c into rust-lang:master May 16, 2018

This was referenced May 16, 2018

Add lint for multiple associated types #50682

Merged

Implement label break value (RFC 2046) #50045

Merged

This was referenced May 16, 2018

tidy: Add a check for empty UI test files #50793

Merged

Can't compile brotli with rustc 1.27.0-nightly (2f2a11dfc 2018-05-16) #50825

Closed

bors added a commit that referenced this pull request May 19, 2018

Auto merge of #50782 - matthewjasper:wheres-main, r=matthewjasper

a308575

Prevent main from having a where clause. Closes #50714 Should this have a crater run? cc #48557, #48214 r? @nikomatsakis

nikomatsakis mentioned this pull request Jul 1, 2018

Move self trait predicate to items #51895

Merged

matthewjasper deleted the allow-trvial-bounds branch September 3, 2018 21:32

arielb1 mentioned this pull request Sep 25, 2018

Trait with HRTB blanket impl is implemented when it shouldn't be #54302

Closed

aturon mentioned this pull request Jun 9, 2019

Tracking issue for RFC #2056: Allow trivial constraints to appear in where clauses #48214

Open

5 tasks

	fn can_use_global_caches(&self, param_env: ty::ParamEnv<'tcx>) -> bool {
	// If there are any where-clauses in scope, then we always use
	// a cache local to this particular scope. Otherwise, we
	// switch to a global cache. We used to try and draw
	// finer-grained distinctions, but that led to a serious of
	// annoying and weird bugs like #22019 and #18290. This simple
	// rule seems to be pretty clearly safe and also still retains
	// a very high hit rate (~95% when compiling rustc).
	if !param_env.caller_bounds.is_empty() {
	return false;
	}

Uh oh!

Conversation

matthewjasper commented Feb 26, 2018 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

rust-highfive commented Feb 26, 2018

Uh oh!

bors commented Feb 27, 2018

Uh oh!

cramertj commented Feb 27, 2018

Uh oh!

shepmaster commented Mar 4, 2018

Uh oh!

nikomatsakis left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

nikomatsakis commented Mar 15, 2018

Uh oh!

matthewjasper commented Mar 15, 2018

Uh oh!

nikomatsakis commented Mar 16, 2018

Uh oh!

nikomatsakis commented Mar 20, 2018

Uh oh!

matthewjasper commented Mar 21, 2018

Uh oh!

bors commented Mar 22, 2018

Uh oh!

nikomatsakis commented Mar 23, 2018

Uh oh!

Choose a reason for hiding this comment

Uh oh!

shepmaster May 15, 2018 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

nikomatsakis left a comment

Choose a reason for hiding this comment

Uh oh!

matthewjasper commented May 15, 2018

Uh oh!

cramertj commented May 15, 2018

Uh oh!

bors commented May 15, 2018

Uh oh!

bors commented May 16, 2018

Uh oh!

bors commented May 16, 2018

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

7 participants

matthewjasper commented Feb 26, 2018 •

edited

Loading

shepmaster May 15, 2018 •

edited

Loading