async/await: improve not-send errors, part 2 #65345

davidtwco · 2019-10-12T17:20:58Z

Part of #64130. Fixes #65667.

This PR improves the errors introduced in #64895 so that they have specialized messages for Send and Sync.

r? @nikomatsakis

src/libcore/marker.rs

src/librustc/traits/error_reporting.rs

src/test/ui/async-await/issue-64130-non-send-future-diags.stderr

davidtwco

Updated the PR (with help from @nikomatsakis) so that it now applies to more cases.

davidtwco · 2019-10-30T17:11:17Z

src/librustc/traits/error_reporting.rs

+                let ty = ty.builtin_deref(false).map(|ty_and_mut| ty_and_mut.ty).unwrap_or(ty);
+                let target_ty = target_ty.builtin_deref(false)
+                    .map(|ty_and_mut| ty_and_mut.ty)
+                    .unwrap_or(target_ty);


This is hack that I wasn't able to find a better solution for.

What problem is it solving, exactly? I guess that we sometimes capture things "by reference"?

I wasn’t able to find a way to check for equality of types that were ty::Ref, nothing seemed to work. ~~There are some messages in the Zulip thread detailing what I tried when I was last looking at this.~~ Nope, misremembered.

davidtwco · 2019-10-30T17:11:41Z

src/test/ui/async-await/async-fn-nonsend.stderr

+LL |     fut().await;
+   |     ^^^^^^^^^^^ await occurs here, with `x` maybe used later
+LL | }
+   | - `x` is later dropped here


This diagnostic is misleading.

Hmm, because of the drop(x) that comes before? I suppose that's true, though it'd ordinarily be quite helpful. Perhaps we can just weaken the language to "x may be dropped here"?

Yeah, I think we could adjust the wording to better explain why the drop call isn’t enough to avoid the type being held over the await point.

I consider this a bit of a bug, also -- it'd be nice if we were more precise around particularly this case)

nikomatsakis · 2019-10-30T18:28:34Z

@davidtwco should this still be considered "draft"?

davidtwco · 2019-10-30T18:34:31Z

@davidtwco should this still be considered "draft"?

Probably not, I’ve unmarked it.

src/test/ui/async-await/issue-64130-4-async-move.stderr

nikomatsakis

uh I had some comments I apparently never posted :(

src/librustc/traits/error_reporting.rs

nikomatsakis · 2019-10-22T19:33:05Z

src/librustc/traits/error_reporting.rs

+        };
+
+        // Look at the last interior type to get a span for the `.await`.
+        let await_span = tables.generator_interior_types.iter().map(|i| i.span).last().unwrap();


huh, so we just take the last span from the table? seems a bit arbitrary. Ideally, it would be tied to the value we are citing..?

but I guess that's pre-existing

nikomatsakis · 2019-10-30T18:29:33Z

src/librustc/traits/error_reporting.rs

+
+        // Get the tables from the infcx if the generator is the function we are
+        // currently type-checking; otherwise, get them by performing a query.
+        // This is needed to avoid cycles.


I wonder if we want a helper function for this

nikomatsakis · 2019-10-30T18:30:13Z

src/librustc/traits/error_reporting.rs

+                let ty = ty.builtin_deref(false).map(|ty_and_mut| ty_and_mut.ty).unwrap_or(ty);
+                let target_ty = target_ty.builtin_deref(false)
+                    .map(|ty_and_mut| ty_and_mut.ty)
+                    .unwrap_or(target_ty);


What problem is it solving, exactly? I guess that we sometimes capture things "by reference"?

JohnCSimon · 2019-11-16T20:09:54Z

Ping from triage:
@davidtwco Can you please address the comments from @nikomatsakis ? Thank you!
cc: @Centril

davidtwco · 2019-11-16T20:49:34Z

Thanks @JohnCSimon, I believe @nikomatsakis wanted to leave more feedback.

Alexendoo · 2019-11-27T20:52:36Z

Ping from triage, any updates? @nikomatsakis

JohnCSimon · 2019-12-07T07:22:03Z

ping from triage:
@davidtwco - can you please address the merge conflict?
once that happens...
@nikomatsakis - please review this PR

Signed-off-by: David Wood <david@davidtw.co>

This commit corrects the diagnostic note for `async move {}` so that `await` is mentioned, rather than `yield`. Signed-off-by: David Wood <david@davidtw.co>

The reason we were invoking `builtin_deref` was to enable comparisons when the type was `&T`. For the reasons outlined in the comment, those comparisons failed because the regions disagreed.

nikomatsakis · 2019-12-11T18:28:45Z

@bors r+

(Tests pass locally)

bors · 2019-12-11T18:28:46Z

📌 Commit 5cd9f22 has been approved by nikomatsakis

bors · 2019-12-11T19:39:12Z

⌛ Testing commit 5cd9f22 with merge 27d6f55...

@nikomatsakis

…provements, r=nikomatsakis async/await: improve not-send errors, part 2 Part of #64130. Fixes #65667. This PR improves the errors introduced in #64895 so that they have specialized messages for `Send` and `Sync`. r? @nikomatsakis

bors · 2019-12-11T22:52:59Z

☀️ Test successful - checks-azure
Approved by: nikomatsakis
Pushing 27d6f55 to master...

rust-highfive assigned nikomatsakis Oct 12, 2019

rust-highfive added the S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. label Oct 12, 2019

Centril reviewed Oct 12, 2019

View reviewed changes

src/libcore/marker.rs Outdated Show resolved Hide resolved

Centril reviewed Oct 12, 2019

View reviewed changes

src/librustc/traits/error_reporting.rs Outdated Show resolved Hide resolved

src/librustc/traits/error_reporting.rs Show resolved Hide resolved

Aaron1011 reviewed Oct 12, 2019

View reviewed changes

src/test/ui/async-await/issue-64130-non-send-future-diags.stderr Outdated Show resolved Hide resolved

davidtwco force-pushed the issue-64130-async-send-sync-error-improvements branch from cba981a to 15ebeab Compare October 13, 2019 16:21

This was referenced Oct 22, 2019

Async/Await: Confusing error message when using non-Send type in async fn #65667

Closed

async/await regression regarding borrows across yields #65436

Closed

This comment has been minimized.

Sign in to view

davidtwco force-pushed the issue-64130-async-send-sync-error-improvements branch from 964759e to 6d16511 Compare October 30, 2019 17:10

davidtwco commented Oct 30, 2019

View reviewed changes

davidtwco marked this pull request as ready for review October 30, 2019 18:34

davidtwco commented Oct 31, 2019

View reviewed changes

src/test/ui/async-await/issue-64130-4-async-move.stderr Outdated Show resolved Hide resolved

This comment has been minimized.

Sign in to view

davidtwco force-pushed the issue-64130-async-send-sync-error-improvements branch from 473d24e to 6c18dd6 Compare November 8, 2019 21:10

nikomatsakis reviewed Nov 8, 2019

View reviewed changes

JohnCSimon 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 Nov 16, 2019

davidtwco 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 Nov 16, 2019

nikomatsakis mentioned this pull request Nov 19, 2019

Show the location of a type usage for trait bound failures #66444

Closed

JohnCSimon 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 Dec 7, 2019

async/await: more improvements to non-send errors

438455d

Signed-off-by: David Wood <david@davidtw.co>

async/await: correct diag note for async move

f0b5114

This commit corrects the diagnostic note for `async move {}` so that `await` is mentioned, rather than `yield`. Signed-off-by: David Wood <david@davidtw.co>

davidtwco force-pushed the issue-64130-async-send-sync-error-improvements branch from 6c18dd6 to f0b5114 Compare December 8, 2019 16:48

This was referenced Dec 10, 2019

Print nicer async/await trait errors for generators in any place in the error 'stack' #67116

Closed

Reduce type errors in async fns #67025

Closed

erase regions instead of using builtin_deref

5cd9f22

The reason we were invoking `builtin_deref` was to enable comparisons when the type was `&T`. For the reasons outlined in the comment, those comparisons failed because the regions disagreed.

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-author Status: This is awaiting some action (such as code changes or more information) from the author. labels Dec 11, 2019

bors added the merged-by-bors This PR was explicitly merged by bors. label Dec 11, 2019

bors merged commit 5cd9f22 into rust-lang:master Dec 11, 2019

hellow554 mentioned this pull request Dec 12, 2019

ICE: no name for expr || #67252

Closed

tmandry mentioned this pull request Dec 17, 2019

non-send futures error at point of use, but could cite point of definition in some cases #64130

Closed

yoshuawuyts mentioned this pull request Dec 17, 2019

improve diagnostics for non-send Futures that might be solved by introducing temporary variables #67376

Closed

davidtwco deleted the issue-64130-async-send-sync-error-improvements branch December 19, 2019 14:57

eddyb mentioned this pull request Apr 6, 2020

note_obligation_cause reporting for non-async/await not optimal #70818

Closed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

async/await: improve not-send errors, part 2 #65345

async/await: improve not-send errors, part 2 #65345

davidtwco commented Oct 12, 2019 •

edited

Loading

This comment has been minimized.

davidtwco left a comment

davidtwco Oct 30, 2019

nikomatsakis Oct 30, 2019

davidtwco Nov 8, 2019 •

edited

Loading

davidtwco Oct 30, 2019

nikomatsakis Nov 8, 2019

davidtwco Nov 8, 2019

nikomatsakis Nov 19, 2019

nikomatsakis commented Oct 30, 2019

davidtwco commented Oct 30, 2019

This comment has been minimized.

nikomatsakis left a comment

nikomatsakis Oct 22, 2019

nikomatsakis Oct 22, 2019

nikomatsakis Oct 30, 2019

nikomatsakis Oct 30, 2019

JohnCSimon commented Nov 16, 2019

davidtwco commented Nov 16, 2019

Alexendoo commented Nov 27, 2019

JohnCSimon commented Dec 7, 2019

nikomatsakis commented Dec 11, 2019

bors commented Dec 11, 2019

bors commented Dec 11, 2019

bors commented Dec 11, 2019

async/await: improve not-send errors, part 2 #65345

async/await: improve not-send errors, part 2 #65345

Conversation

davidtwco commented Oct 12, 2019 • edited Loading

This comment has been minimized.

davidtwco left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

davidtwco Nov 8, 2019 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

nikomatsakis commented Oct 30, 2019

davidtwco commented Oct 30, 2019

This comment has been minimized.

nikomatsakis left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

JohnCSimon commented Nov 16, 2019

davidtwco commented Nov 16, 2019

Alexendoo commented Nov 27, 2019

JohnCSimon commented Dec 7, 2019

nikomatsakis commented Dec 11, 2019

bors commented Dec 11, 2019

bors commented Dec 11, 2019

bors commented Dec 11, 2019

davidtwco commented Oct 12, 2019 •

edited

Loading

davidtwco Nov 8, 2019 •

edited

Loading