Don't use lift to detect local types #61871

Zoxc · 2019-06-15T19:20:02Z

This overlaps with #61392.

rust-highfive · 2019-06-15T19:34:30Z

The job x86_64-gnu-llvm-6.0 of your PR failed on Travis (raw log). Through arcane magic we have determined that the following fragments from the build log may contain information about the problem.

Click to expand the log.

travis_time:end:05217f82:start=1560626469432056280,finish=1560626564629490044,duration=95197433764
$ git checkout -qf FETCH_HEAD
travis_fold:end:git.checkout

Encrypted environment variables have been removed for security reasons.
See https://docs.travis-ci.com/user/pull-requests/#pull-requests-and-security-restrictions
$ export SCCACHE_BUCKET=rust-lang-ci-sccache2
$ export SCCACHE_REGION=us-west-1
$ export GCP_CACHE_BUCKET=rust-lang-ci-cache
$ export AWS_ACCESS_KEY_ID=AKIA46X5W6CZEJZ6XT55
---

[00:03:49] travis_fold:start:tidy
travis_time:start:tidy
tidy check
[00:03:49] tidy error: /checkout/src/librustc/ty/context.rs: too many lines (3068) (add `// ignore-tidy-filelength` to the file to suppress this error)
[00:03:54] some tidy checks failed
[00:03:54] 
[00:03:54] 
[00:03:54] command did not execute successfully: "/checkout/obj/build/x86_64-unknown-linux-gnu/stage0-tools-bin/tidy" "/checkout/src" "/checkout/obj/build/x86_64-unknown-linux-gnu/stage0/bin/cargo" "--no-vendor" "--quiet"
[00:03:54] 
[00:03:54] 
[00:03:54] failed to run: /checkout/obj/build/bootstrap/debug/bootstrap test src/tools/tidy
[00:03:54] Build completed unsuccessfully in 0:01:11
---
travis_time:end:150ef786:start=1560626809115496126,finish=1560626809120038297,duration=4542171
travis_fold:end:after_failure.3
travis_fold:start:after_failure.4
travis_time:start:18c8ef24
$ ln -s . checkout && for CORE in obj/cores/core.*; do EXE=$(echo $CORE | sed 's|obj/cores/core\.[0-9]*\.!checkout!\(.*\)|\1|;y|!|/|'); if [ -f "$EXE" ]; then printf travis_fold":start:crashlog\n\033[31;1m%s\033[0m\n" "$CORE"; gdb --batch -q -c "$CORE" "$EXE" -iex 'set auto-load off' -iex 'dir src/' -iex 'set sysroot .' -ex bt -ex q; echo travis_fold":"end:crashlog; fi; done || true
travis_fold:end:after_failure.4
travis_fold:start:after_failure.5
travis_time:start:0e9e7588
travis_time:start:0e9e7588
$ cat ./obj/build/x86_64-unknown-linux-gnu/native/asan/build/lib/asan/clang_rt.asan-dynamic-i386.vers || true
cat: ./obj/build/x86_64-unknown-linux-gnu/native/asan/build/lib/asan/clang_rt.asan-dynamic-i386.vers: No such file or directory
travis_fold:end:after_failure.5
travis_fold:start:after_failure.6
travis_time:start:000f04a8
$ dmesg | grep -i kill

I'm a bot! I can only do what humans tell me to, so if this was not helpful or you have suggestions for improvements, please ping or otherwise contact @TimNN. (Feature Requests)

rust-highfive · 2019-06-15T19:56:42Z

The job x86_64-gnu-llvm-6.0 of your PR failed on Travis (raw log). Through arcane magic we have determined that the following fragments from the build log may contain information about the problem.

Click to expand the log.

travis_time:end:047c93c4:start=1560627424950581360,finish=1560627514195892917,duration=89245311557
$ git checkout -qf FETCH_HEAD
travis_fold:end:git.checkout

Encrypted environment variables have been removed for security reasons.
See https://docs.travis-ci.com/user/pull-requests/#pull-requests-and-security-restrictions
$ export SCCACHE_BUCKET=rust-lang-ci-sccache2
$ export SCCACHE_REGION=us-west-1
$ export GCP_CACHE_BUCKET=rust-lang-ci-cache
$ export AWS_ACCESS_KEY_ID=AKIA46X5W6CZEJZ6XT55
---
[00:14:37]    Compiling rustc_mir v0.0.0 (/checkout/src/librustc_mir)
[00:14:52] error: unused variable: `gcx`
[00:14:52]    --> src/librustc_mir/borrow_check/nll/region_infer/mod.rs:812:13
[00:14:52]     |
[00:14:52] 812 |         let gcx = tcx.global_tcx();
[00:14:52]     |             ^^^ help: consider prefixing with an underscore: `_gcx`
[00:14:52]     = note: `-D unused-variables` implied by `-D warnings`
[00:14:52] 
[00:15:01]    Compiling rustc_typeck v0.0.0 (/checkout/src/librustc_typeck)
[00:15:02] error: aborting due to previous error
---
travis_time:end:2f83a7c6:start=1560628597803657351,finish=1560628597809069415,duration=5412064
travis_fold:end:after_failure.3
travis_fold:start:after_failure.4
travis_time:start:12cc995e
$ ln -s . checkout && for CORE in obj/cores/core.*; do EXE=$(echo $CORE | sed 's|obj/cores/core\.[0-9]*\.!checkout!\(.*\)|\1|;y|!|/|'); if [ -f "$EXE" ]; then printf travis_fold":start:crashlog\n\033[31;1m%s\033[0m\n" "$CORE"; gdb --batch -q -c "$CORE" "$EXE" -iex 'set auto-load off' -iex 'dir src/' -iex 'set sysroot .' -ex bt -ex q; echo travis_fold":"end:crashlog; fi; done || true
travis_fold:end:after_failure.4
travis_fold:start:after_failure.5
travis_time:start:049b5600
travis_time:start:049b5600
$ cat ./obj/build/x86_64-unknown-linux-gnu/native/asan/build/lib/asan/clang_rt.asan-dynamic-i386.vers || true
cat: ./obj/build/x86_64-unknown-linux-gnu/native/asan/build/lib/asan/clang_rt.asan-dynamic-i386.vers: No such file or directory
travis_fold:end:after_failure.5
travis_fold:start:after_failure.6
travis_time:start:3241ba93
$ dmesg | grep -i kill

I'm a bot! I can only do what humans tell me to, so if this was not helpful or you have suggestions for improvements, please ping or otherwise contact @TimNN. (Feature Requests)

Zoxc · 2019-06-15T21:43:47Z

@bors try

bors · 2019-06-15T21:43:59Z

⌛ Trying commit 90fbfd7095ad0b022adb968bbc72ce51b55e2467 with merge 3e509e149d6644da28c5e2b642d164845e724047...

eddyb · 2019-06-16T09:53:41Z

src/librustc/infer/mod.rs

@@ -1455,7 +1455,7 @@ impl<'a, 'tcx> InferCtxt<'a, 'tcx> {
        // Even if the type may have no inference variables, during
        // type-checking closure types are in local tables only.
        if !self.in_progress_tables.is_some() || !ty.has_closure_types() {
-            if let Some((param_env, ty)) = self.tcx.lift_to_global(&(param_env, ty)) {
+            if !(param_env, ty).has_local_value() {


cc @rust-lang/compiler I think it might be better if we stop referring to these as "local values" and instead use something more like .has_infer_vars()?

has_infer_types already exist and is a different property.

eddyb · 2019-06-16T09:55:56Z

src/librustc/middle/mem_categorization.rs

-                    ty.is_copy_modulo_regions(self.tcx.global_tcx(), param_env, span)
-                })
+                if (param_env, ty).has_local_value() {
+                    None


Hmm, if self.infcx is None, this case should be impossible - it's plausible this code is unnecessarily defensive and lift_to_global's result could've been unwrap'd all along.

eddyb · 2019-06-16T09:57:02Z

src/librustc/traits/fulfill.rs

+                                promoted: None,
+                            };
+                            match self.selcx.tcx().at(obligation.cause.span)
+                                                    .const_eval(obligation.param_env.and(cid)) {


I can't wait for this to use erasure and canonicalization instead (@oli-obk and @nikomatsakis are working on that, I think?).

eddyb · 2019-06-16T09:59:32Z

src/librustc/traits/mod.rs

-            }
+        if predicates.has_local_value() {
+            // FIXME: shouldn't we, you know, actually report an error here? or an ICE?
+            Err(ErrorReported)


@rust-lang/wg-traits This is scary, can we do a crater run with a delay_span_bug in here?

@eddyb What happened to this?

src/librustc/traits/project.rs

src/librustc/traits/query/normalize.rs

src/librustc_typeck/check/writeback.rs

eddyb · 2019-06-16T10:02:16Z

src/librustc_typeck/check/writeback.rs

-        if let Some(lifted) = self.tcx().lift_to_global(&x) {
-            lifted
-        } else {
+        if cfg!(debug_assertions) && x.has_local_value() {


Is it really that expensive? Anyway it should be impossible, given a correct Resolver.

I don't know how expensive these are. They're not free though, so I'd prefer to not have them for release builds.

And yes, it should be impossible, that's why it's an assertion =P

eddyb · 2019-06-16T10:04:07Z

src/librustc_typeck/check/writeback.rs

-            let c_sig = if let Some(c_sig) = self.tcx().lift_to_global(c_sig) {
-                c_sig
-            } else {
+            if cfg!(debug_assertions) && c_sig.has_local_value() {


Hmm, can we do a crater run with all of these assertions turned on by default, to confirm they're very likely never triggered, and then remove them?

They're effectively turned on now, and they have caught at least one bug #46184.

I do think we should have these assertions somewhere though, to ensure that no infer types leaks out from typeck. I could just use debug_assert here instead of span_bug to make them more slim.

eddyb · 2019-06-16T10:07:56Z

src/librustc/traits/select.rs

@@ -808,26 +825,24 @@ impl<'cx, 'tcx> SelectionContext<'cx, 'tcx> {

            ty::Predicate::ConstEvaluatable(def_id, substs) => {
                let tcx = self.tcx();
-                match tcx.lift_to_global(&(obligation.param_env, substs)) {
-                    Some((param_env, substs)) => {


I'd prefer if one of two things happened here (and in other places match-ing over lift_to_global results):

land a PR on master replacing the matches with if lets (you can just r=me that if you want)

keep the match and either use Option::filter or just match on the bool predicate

That is, I would prefer if indentation didn't change (and it only does because of the match -> if change).

Why do you not want indention to change? I like less rightward drift, and I definitely don't like matching on bools.

It's fine to change, I just want to be able to read the diff without mentally pretending nothing changed without checking, where indentation is involved.

I changed the indentation to reduce the diffs in this PR.

@eddyb you can append ?w=1 to the diff url to have whitespace insensitive diffs https://github.com/rust-lang/rust/pull/61871/files?w=1

Nice. I also reordered some branches which that can't deal with though =P

eddyb · 2019-06-16T10:08:17Z

src/librustc/traits/fulfill.rs

@@ -461,40 +461,34 @@ impl<'a, 'b, 'tcx> ObligationProcessor for FulfillProcessor<'a, 'b, 'tcx> {
            }

            ty::Predicate::ConstEvaluatable(def_id, substs) => {
-                match self.selcx.tcx().lift_to_global(&obligation.param_env) {
-                    None => {
+                if obligation.param_env.has_local_value() {


Same as the other match -> if change.

bors · 2019-06-19T09:58:37Z

☔ The latest upstream changes (presumably #61945) made this pull request unmergeable. Please resolve the merge conflicts.

rust-highfive · 2019-06-19T18:45:22Z

The job x86_64-gnu-llvm-6.0 of your PR failed on Travis (raw log). Through arcane magic we have determined that the following fragments from the build log may contain information about the problem.

Click to expand the log.

travis_time:end:1df9c270:start=1560968986542163132,finish=1560968987405713224,duration=863550092
$ git checkout -qf FETCH_HEAD
travis_fold:end:git.checkout

Encrypted environment variables have been removed for security reasons.
See https://docs.travis-ci.com/user/pull-requests/#pull-requests-and-security-restrictions
$ export SCCACHE_BUCKET=rust-lang-ci-sccache2
$ export SCCACHE_REGION=us-west-1
$ export GCP_CACHE_BUCKET=rust-lang-ci-cache
$ export AWS_ACCESS_KEY_ID=AKIA46X5W6CZEJZ6XT55
---

[00:04:52] travis_fold:start:tidy
travis_time:start:tidy
tidy check
[00:04:52] tidy error: /checkout/src/librustc/traits/fulfill.rs:480: line longer than 100 chars
[00:04:57] some tidy checks failed
[00:04:57] 
[00:04:57] 
[00:04:57] command did not execute successfully: "/checkout/obj/build/x86_64-unknown-linux-gnu/stage0-tools-bin/tidy" "/checkout/src" "/checkout/obj/build/x86_64-unknown-linux-gnu/stage0/bin/cargo" "--no-vendor" "--quiet"
[00:04:57] 
[00:04:57] 
[00:04:57] failed to run: /checkout/obj/build/bootstrap/debug/bootstrap test src/tools/tidy
[00:04:57] Build completed unsuccessfully in 0:01:15
---
travis_time:end:24c77cd0:start=1560969296547077810,finish=1560969296551855472,duration=4777662
travis_fold:end:after_failure.3
travis_fold:start:after_failure.4
travis_time:start:0e6f8794
$ ln -s . checkout && for CORE in obj/cores/core.*; do EXE=$(echo $CORE | sed 's|obj/cores/core\.[0-9]*\.!checkout!\(.*\)|\1|;y|!|/|'); if [ -f "$EXE" ]; then printf travis_fold":start:crashlog\n\033[31;1m%s\033[0m\n" "$CORE"; gdb --batch -q -c "$CORE" "$EXE" -iex 'set auto-load off' -iex 'dir src/' -iex 'set sysroot .' -ex bt -ex q; echo travis_fold":"end:crashlog; fi; done || true
travis_fold:end:after_failure.4
travis_fold:start:after_failure.5
travis_time:start:07441038
travis_time:start:07441038
$ cat ./obj/build/x86_64-unknown-linux-gnu/native/asan/build/lib/asan/clang_rt.asan-dynamic-i386.vers || true
cat: ./obj/build/x86_64-unknown-linux-gnu/native/asan/build/lib/asan/clang_rt.asan-dynamic-i386.vers: No such file or directory
travis_fold:end:after_failure.5
travis_fold:start:after_failure.6
travis_time:start:1d1a7b9f
$ dmesg | grep -i kill

I'm a bot! I can only do what humans tell me to, so if this was not helpful or you have suggestions for improvements, please ping or otherwise contact @TimNN. (Feature Requests)

eddyb · 2019-07-01T12:40:40Z

@bors r+

bors · 2019-07-01T12:40:42Z

📌 Commit 8465daf has been approved by eddyb

bors · 2019-07-01T13:32:28Z

⌛ Testing commit 8465daf with merge ebe126b40db5acb422774453b0d9a2d821bb3755...

pietroalbini · 2019-07-01T15:14:12Z

@bors retry

Yielding priority to the Azure migration.

@eddyb

Don't use lift to detect local types This overlaps with rust-lang#61392. r? @eddyb

bors · 2019-07-02T08:09:23Z

⌛ Testing commit 8465daf with merge ef064d2...

@eddyb

Don't use lift to detect local types This overlaps with #61392. r? @eddyb

bors · 2019-07-02T11:38:49Z

☀️ Test successful - checks-azure, checks-travis, status-appveyor
Approved by: eddyb
Pushing ef064d2 to master...

rust-highfive assigned eddyb Jun 15, 2019

rust-highfive added the S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. label Jun 15, 2019

Zoxc force-pushed the no-lift-branch branch from 0f340b3 to 2d23d26 Compare June 15, 2019 19:35

Zoxc force-pushed the no-lift-branch branch from 2d23d26 to 90fbfd7 Compare June 15, 2019 20:05

Zoxc force-pushed the no-lift-branch branch from 90fbfd7 to cc53706 Compare June 15, 2019 23:47

Zoxc mentioned this pull request Jun 15, 2019

Use a single CtxtInterners #61392

Merged

eddyb reviewed Jun 16, 2019

View reviewed changes

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

eddyb reviewed Jun 16, 2019

View reviewed changes

src/librustc/traits/query/normalize.rs Outdated Show resolved Hide resolved

eddyb reviewed Jun 16, 2019

View reviewed changes

src/librustc_typeck/check/writeback.rs Outdated Show resolved Hide resolved

eddyb reviewed Jun 16, 2019

View reviewed changes

Zoxc force-pushed the no-lift-branch branch 3 times, most recently from a35c705 to fabe708 Compare June 16, 2019 13:32

Zoxc force-pushed the no-lift-branch branch from fabe708 to da7a27d Compare June 19, 2019 18:28

Zoxc added 2 commits June 26, 2019 13:04

Don't use lift to detect local types

c932518

Check for local types in writeback with debug assertions

8465daf

Zoxc force-pushed the no-lift-branch branch from da7a27d to 8465daf Compare June 26, 2019 11:04

eddyb approved these changes Jul 1, 2019

View reviewed changes

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 Jul 1, 2019

Manishearth added a commit to Manishearth/rust that referenced this pull request Jul 2, 2019

Rollup merge of rust-lang#61871 - Zoxc:no-lift-branch, r=eddyb

e58fbbb

Don't use lift to detect local types This overlaps with rust-lang#61392. r? @eddyb

Manishearth mentioned this pull request Jul 2, 2019

Rollup of 16 pull requests #62294

Closed

bors added a commit that referenced this pull request Jul 2, 2019

Auto merge of #61871 - Zoxc:no-lift-branch, r=eddyb

ef064d2

Don't use lift to detect local types This overlaps with #61392. r? @eddyb

bors added the merged-by-bors This PR was explicitly merged by bors. label Jul 2, 2019

bors merged commit 8465daf into rust-lang:master Jul 2, 2019

Zoxc deleted the no-lift-branch branch July 2, 2019 11:46

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Don't use lift to detect local types #61871

Don't use lift to detect local types #61871

Zoxc commented Jun 15, 2019

rust-highfive commented Jun 15, 2019

rust-highfive commented Jun 15, 2019

Zoxc commented Jun 15, 2019

bors commented Jun 15, 2019

eddyb Jun 16, 2019

Zoxc Jun 16, 2019

eddyb Jun 16, 2019

Zoxc Jun 16, 2019

eddyb Jun 16, 2019

eddyb Jun 16, 2019

Centril Jul 1, 2019

eddyb Jun 16, 2019

Zoxc Jun 16, 2019

eddyb Jun 16, 2019

Zoxc Jun 16, 2019

eddyb Jun 16, 2019

Zoxc Jun 16, 2019

eddyb Jun 16, 2019

Zoxc Jun 19, 2019

estebank Jun 19, 2019

Zoxc Jun 20, 2019

eddyb Jun 16, 2019

bors commented Jun 19, 2019

rust-highfive commented Jun 19, 2019

eddyb commented Jul 1, 2019

bors commented Jul 1, 2019

bors commented Jul 1, 2019

pietroalbini commented Jul 1, 2019

bors commented Jul 2, 2019

bors commented Jul 2, 2019

Don't use lift to detect local types #61871

Don't use lift to detect local types #61871

Conversation

Zoxc commented Jun 15, 2019

rust-highfive commented Jun 15, 2019

rust-highfive commented Jun 15, 2019

Zoxc commented Jun 15, 2019

bors commented Jun 15, 2019

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

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

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

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

bors commented Jun 19, 2019

rust-highfive commented Jun 19, 2019

eddyb commented Jul 1, 2019

bors commented Jul 1, 2019

bors commented Jul 1, 2019

pietroalbini commented Jul 1, 2019

bors commented Jul 2, 2019

bors commented Jul 2, 2019