Simplify higher-ranked LUB/GLB #45853

nikomatsakis · 2017-11-07T22:43:10Z

This is a better version of #44211. It still makes higher-ranked LUB/GLB into a hard equality test, however, it does try to identify that something changed and issue a notice to the user. I wroteup #45852 as a tracking issue for this change.

Currently, this moves straight to a hard-error, on the basis that the crater run in #44211 saw no impact. It might be good to retest -- or perhaps to try for a warning period. Trying to do the latter in a precise way would be somewhat painful, but an imprecise way might suffice -- that is, we could issue warning whenever a LUB/GLB operation succeeds that will later fail, even if it doesn't ultimately impact the type check. I could experiment with this.

I am mildly wary about landing this independently of other code that moves to a universe-based system. In particular, I was nervous that this change would make coherence accepts new pairs of impls that will later be errors. I have the code for the universe-based approach available, I hope to open an PR and run some tests on its impact very shortly. @arielb1 points out that I was being silly.

r? @arielb1

arielb1 · 2017-11-08T10:23:26Z

In particular, I was nervous that this change would make coherence accepts new pairs of impls that will later be errors

Coherence is based only on type equality, while this PR only affects subtyping - I don't believe the LUB code can be run from coherence.

arielb1 · 2017-11-08T15:08:34Z

if that wasn't clear, r=me unless there's some other complication about coherence.

arielb1 · 2017-11-08T17:30:33Z

@bors r+

bors · 2017-11-08T17:30:34Z

📌 Commit e782f33 has been approved by arielb1

bors · 2017-11-12T04:15:10Z

⌛ Testing commit e782f33137258d27566d2217f6e068c2ff24d3c0 with merge 827619703a023bb5e57d6450979a91040aca1f61...

bors · 2017-11-12T05:08:06Z

💔 Test failed - status-appveyor

kennytm · 2017-11-12T05:25:11Z

x86_64-windows-msvc UI test failed:

 error[E0308]: match arms have incompatible types
   --> $DIR/old-lub-glb-object.rs:20:13
    |
 20 |       let z = match 22 {
    |  _____________^
 21 | |         0 => x,
 22 | |         _ => y,
 23 | |     };
-   | |_____^ expected bound lifetime parameter 'a, found concrete lifetime
+   | |_____^ expected bound lifetime parameter 'b, found concrete lifetime
    |
    = note: expected type `&for<'a, 'b> Foo<&'a u8, &'b u8>`
               found type `&for<'a> Foo<&'a u8, &'a u8>`
    = note: this was previously accepted by the compiler but has been phased out
    = note: for more information, see https://github.com/rust-lang/rust/issues/45852
 note: match arm with an incompatible type
   --> $DIR/old-lub-glb-object.rs:22:14
    |
 22 |         _ => y,
    |              ^
 
 error: aborting due to previous error

nikomatsakis · 2017-11-12T10:11:02Z

Argh. I wonder if this is the same problem we saw in #44167. I'm pulling @cengizio's commits here temporarily just to see what AppVeyor says.

arielb1 · 2017-11-12T13:06:01Z

I think we should wait for his PR to land and then rebase this one on top of it

nikomatsakis · 2017-11-13T10:29:03Z

@bors r=arielb1

bors · 2017-11-13T10:29:04Z

📌 Commit 0bbbfaf has been approved by arielb1

bors · 2017-11-13T16:50:19Z

⌛ Testing commit 0bbbfaf with merge 55324a9a5b7b2a7eaa7bae4c8ffb4c1967f16de1...

bors · 2017-11-13T17:42:09Z

💔 Test failed - status-travis

kennytm · 2017-11-13T17:51:03Z

armhf-gnu, exactly the same error as before.

 error[E0308]: match arms have incompatible types
   --> $DIR/old-lub-glb-object.rs:20:13
    |
 20 |       let z = match 22 {
    |  _____________^
 21 | |         0 => x,
 22 | |         _ => y,
 23 | |     };
-   | |_____^ expected bound lifetime parameter 'a, found concrete lifetime
+   | |_____^ expected bound lifetime parameter 'b, found concrete lifetime
    |
    = note: expected type `&for<'a, 'b> Foo<&'a u8, &'b u8>`
               found type `&for<'a> Foo<&'a u8, &'a u8>`
    = note: this was previously accepted by the compiler but has been phased out
    = note: for more information, see https://github.com/rust-lang/rust/issues/45852
 note: match arm with an incompatible type
   --> $DIR/old-lub-glb-object.rs:22:14
    |
 22 |         _ => y,
    |              ^
 
 error: aborting due to previous error

Edit: x86_64-pc-windows-msvc also produces the same error like before.

nikomatsakis · 2017-11-13T20:15:05Z

@kennytm thanks, I'll take a look more closely then. Thought we had squelched that.

kennytm · 2017-11-14T08:35:00Z

@nikomatsakis Looks like Travis passed this time.

nikomatsakis · 2017-11-14T19:54:23Z

Huh. Interesting.

arielb1 · 2017-11-16T16:38:24Z

src/librustc/infer/error_reporting/mod.rs

        self.check_and_note_conflicting_crates(diag, terr, span);
        self.tcx.note_and_explain_type_err(diag, terr, span);
+        self.check_and_note_conflicting_crates(diag, terr, span);


duplicate call to check_and_note_conflicting_crates?

nikomatsakis · 2017-11-16T16:39:16Z

@arielb1 argh

nikomatsakis · 2017-11-16T16:44:10Z

Fixed, added more comments.

nikomatsakis · 2017-11-16T16:44:16Z

@bors r=arielb1

bors · 2017-11-16T16:44:17Z

📌 Commit 5e9469c has been approved by arielb1

nikomatsakis · 2017-11-16T16:44:43Z

@bors r-

arielb1 · 2017-11-16T16:44:58Z

I think it's a good idea to have #45825 land first

nikomatsakis · 2017-11-16T16:45:16Z

Agreed. Waiting for #45825 to land.

bors · 2017-11-16T23:28:42Z

☔ The latest upstream changes (presumably #45825) made this pull request unmergeable. Please resolve the merge conflicts.

Good riddance.

The ordering can affect error msg, and this map is not a high performance pathway.

nikomatsakis · 2017-11-17T15:44:11Z

@bors r=arielb1

bors · 2017-11-17T15:44:11Z

📌 Commit 9877fa0 has been approved by arielb1

bors · 2017-11-17T15:59:24Z

⌛ Testing commit 9877fa0 with merge c86d99d...

@arielb1

Simplify higher-ranked LUB/GLB This is a better version of #44211. It still makes higher-ranked LUB/GLB into a hard equality test, however, it does try to identify that something changed and issue a notice to the user. I wroteup #45852 as a tracking issue for this change. Currently, this moves straight to a hard-error, on the basis that the crater run in #44211 saw no impact. It might be good to retest -- or perhaps to try for a warning period. Trying to do the latter in a precise way would be somewhat painful, but an imprecise way might suffice -- that is, we could issue warning *whenever* a LUB/GLB operation succeeds that will later fail, even if it doesn't ultimately impact the type check. I could experiment with this. ~~I am *mildly* wary about landing this independently of other code that moves to a universe-based system. In particular, I was nervous that this change would make coherence accepts new pairs of impls that will later be errors. I have the code for the universe-based approach available, I hope to open an PR and run some tests on its impact very shortly.~~ @arielb1 points out that I was being silly. r? @arielb1

bors · 2017-11-17T19:00:14Z

💔 Test failed - status-appveyor

kennytm · 2017-11-17T19:04:42Z

@bors retry

check i686-pc-windows-msvc 3-hour timeout

Time breakdown

step	1.0.5329	1.0.5390
make prepare	51.32	56.5
stage0-std	46.15	73.53
stage0-tidy	7.99	9.69
bootstrap	17.54	25.72
stage0-test	15.39	16.74
llvm	448.4	693.57
stage0-rustc	1034.28	1133.1
stage1-std	57.4	66.13
stage1-test	20.3	24.27
stage1-rustc	1489.61	1702.34
stage0-compiletest	64.62	86.1
test/ui	32.7	60.9
test/run-pass	1373.26	2526.66
test/compile-fail	313.2	395.92
test/parse-fail	5.65	6.77
test/run-fail	28.51	60.63
test/run-pass-valgrind	3.8	3.77
test/mir-opt	13.58	37.95
test/codegen	6.26	11.35
test/codegen-units	7.19	11.74
test/incremental	66.69	74.59
test/ui-fulldeps	15.18	15.05
test/run-pass-fulldeps	219.09	242.29
test/run-fail-fulldeps	3.99	4.96
test/compile-fail-fulldeps	62.53	77
stage2-rustdoc	266.1	290.77
test/run-make	93.56	137.75
test/rustdoc	109.51	130.62
stage1-std(test)	190.87	264.18
test/libstd	807.12	1283.24
stage1-test(test)	16.5	16.8
test/libtest	17.41	17.85
stage1-rustc(test)	341.36	336.32
test/librustc	345.77	345.37
stage2-rustdoc(test)	29.8	41.15
test/librustdoc	30.37	48.84
stage0-unstable-book-gen	11.15	19.6
stage0-rustbook	241.87	203.27
doc/libstd	24.7	42.79
doc/libproc_macro	1.39	3.44
stage0-error_index_gen	5.67	7.12
stage0-linkchecker	3.2	3.23
test/linkchecker	35.87	37.21
doc tests	261.95
error-index tests	61.61

Everything, even LLVM build went slower than usual. I think it's just AppVeyor's problem.

bors · 2017-11-17T22:16:17Z

⌛ Testing commit 9877fa0 with merge 18d8acf...

@arielb1

Simplify higher-ranked LUB/GLB This is a better version of #44211. It still makes higher-ranked LUB/GLB into a hard equality test, however, it does try to identify that something changed and issue a notice to the user. I wroteup #45852 as a tracking issue for this change. Currently, this moves straight to a hard-error, on the basis that the crater run in #44211 saw no impact. It might be good to retest -- or perhaps to try for a warning period. Trying to do the latter in a precise way would be somewhat painful, but an imprecise way might suffice -- that is, we could issue warning *whenever* a LUB/GLB operation succeeds that will later fail, even if it doesn't ultimately impact the type check. I could experiment with this. ~~I am *mildly* wary about landing this independently of other code that moves to a universe-based system. In particular, I was nervous that this change would make coherence accepts new pairs of impls that will later be errors. I have the code for the universe-based approach available, I hope to open an PR and run some tests on its impact very shortly.~~ @arielb1 points out that I was being silly. r? @arielb1

bors · 2017-11-18T00:44:53Z

☀️ Test successful - status-appveyor, status-travis
Approved by: arielb1
Pushing 18d8acf to master...

rust-highfive assigned arielb1 Nov 7, 2017

nikomatsakis force-pushed the chalk-simplify-hr-lub-glb branch from 0948c82 to 77ca1b7 Compare November 7, 2017 22:44

Mark-Simulacrum added the relnotes Marks issues that should be documented in the release notes of the next release. label Nov 7, 2017

kennytm added the S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. label Nov 8, 2017

nikomatsakis force-pushed the chalk-simplify-hr-lub-glb branch from c924176 to e782f33 Compare November 8, 2017 09:20

kennytm 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 Nov 10, 2017

nikomatsakis force-pushed the chalk-simplify-hr-lub-glb branch 2 times, most recently from e782f33 to 0bbbfaf Compare November 13, 2017 10:28

kennytm 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-bors Status: Waiting on bors to run and complete tests. Bors will change the label on completion. labels Nov 13, 2017

nikomatsakis force-pushed the chalk-simplify-hr-lub-glb branch from 7985ac6 to bb0e756 Compare November 14, 2017 09:51

kennytm 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-bors Status: Waiting on bors to run and complete tests. Bors will change the label on completion. labels Nov 14, 2017

arielb1 reviewed Nov 16, 2017

View reviewed changes

nikomatsakis force-pushed the chalk-simplify-hr-lub-glb branch from bd9167e to 5e9469c Compare November 16, 2017 16:43

nikomatsakis added 4 commits November 17, 2017 10:32

make LUB/GLB of higher-ranked things actually do EQ

397973b

issue better error message when LUB/GLB diverge under new behavior

b224397

strip old lub-glb tests from librustc_driver

ff8cd2e

Good riddance.

use a BTreeMap instead of an FxHasMap for the skol regions

9877fa0

The ordering can affect error msg, and this map is not a high performance pathway.

nikomatsakis force-pushed the chalk-simplify-hr-lub-glb branch from 5e9469c to 9877fa0 Compare November 17, 2017 15:33

kennytm 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 Nov 17, 2017

bors merged commit 9877fa0 into rust-lang:master Nov 18, 2017

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Simplify higher-ranked LUB/GLB #45853

Simplify higher-ranked LUB/GLB #45853

nikomatsakis commented Nov 7, 2017 •

edited

Loading

arielb1 commented Nov 8, 2017

arielb1 commented Nov 8, 2017

arielb1 commented Nov 8, 2017

bors commented Nov 8, 2017

bors commented Nov 12, 2017

bors commented Nov 12, 2017

kennytm commented Nov 12, 2017

nikomatsakis commented Nov 12, 2017

arielb1 commented Nov 12, 2017

nikomatsakis commented Nov 13, 2017

bors commented Nov 13, 2017

bors commented Nov 13, 2017

bors commented Nov 13, 2017

kennytm commented Nov 13, 2017 •

edited

Loading

nikomatsakis commented Nov 13, 2017

kennytm commented Nov 14, 2017

nikomatsakis commented Nov 14, 2017

arielb1 Nov 16, 2017

nikomatsakis commented Nov 16, 2017

nikomatsakis commented Nov 16, 2017

nikomatsakis commented Nov 16, 2017

bors commented Nov 16, 2017

nikomatsakis commented Nov 16, 2017

arielb1 commented Nov 16, 2017

nikomatsakis commented Nov 16, 2017

bors commented Nov 16, 2017

nikomatsakis commented Nov 17, 2017

bors commented Nov 17, 2017

bors commented Nov 17, 2017

bors commented Nov 17, 2017

kennytm commented Nov 17, 2017 •

edited

Loading

bors commented Nov 17, 2017

bors commented Nov 18, 2017

Simplify higher-ranked LUB/GLB #45853

Simplify higher-ranked LUB/GLB #45853

Conversation

nikomatsakis commented Nov 7, 2017 • edited Loading

arielb1 commented Nov 8, 2017

arielb1 commented Nov 8, 2017

arielb1 commented Nov 8, 2017

bors commented Nov 8, 2017

bors commented Nov 12, 2017

bors commented Nov 12, 2017

kennytm commented Nov 12, 2017

nikomatsakis commented Nov 12, 2017

arielb1 commented Nov 12, 2017

nikomatsakis commented Nov 13, 2017

bors commented Nov 13, 2017

bors commented Nov 13, 2017

bors commented Nov 13, 2017

kennytm commented Nov 13, 2017 • edited Loading

nikomatsakis commented Nov 13, 2017

kennytm commented Nov 14, 2017

nikomatsakis commented Nov 14, 2017

arielb1 Nov 16, 2017

Choose a reason for hiding this comment

nikomatsakis commented Nov 16, 2017

nikomatsakis commented Nov 16, 2017

nikomatsakis commented Nov 16, 2017

bors commented Nov 16, 2017

nikomatsakis commented Nov 16, 2017

arielb1 commented Nov 16, 2017

nikomatsakis commented Nov 16, 2017

bors commented Nov 16, 2017

nikomatsakis commented Nov 17, 2017

bors commented Nov 17, 2017

bors commented Nov 17, 2017

bors commented Nov 17, 2017

kennytm commented Nov 17, 2017 • edited Loading

bors commented Nov 17, 2017

bors commented Nov 18, 2017

nikomatsakis commented Nov 7, 2017 •

edited

Loading

kennytm commented Nov 13, 2017 •

edited

Loading

kennytm commented Nov 17, 2017 •

edited

Loading