querify layout and move param env out of the infcx #42189

nikomatsakis · 2017-05-24T14:06:45Z

The main goal of this PR is to move the parameter environment out of the inference context. This is because the inference environment will soon be changing over the course of inference --- for example, when we enter into a for<'a> fn(...) type, we will push a new environment with an increasing universe index, rather than skolemizing the 'a references. Similarly, each obligation will soon be able to have a distinct parameter environment, and therefore the Obligation struct is extended to carry a ParamEnv<'tcx>. (I debated about putting it into the cause; seems plausible, but also weird.)

Along the way, I also reworked how layout works, moving the layout cache into a proper query along the lines of needs-drop and friends.

Finally, tweaks the inference context API. It seemed to be accumulating parameters at an alarming rate. The main way to e.g. make a subtype or equality relationship is to do the following:

infcx.at(cause, param_env).sub(a, b)
infcx.at(cause, param_env).eq(a, b)

In both cases, a is considered the "expected" type (this used to be specified by a boolean). I tried hard to preserve the existing notion of what was "expected", although in some cases I'm not convinced it was being set on purpose one way or the other. This is why in some cases you will see me do sup(b, a), which is otherwise equivalent to sub(a, b), but sets the "expected type" differently.

r? @eddyb
cc @arielb1

eddyb · 2017-05-25T07:08:35Z

src/librustc/ty/layout.rs

+            let type_desc = format!("{:?}", ty);
+            let overall_size = layout.size(tcx);
+            let align = layout.align(tcx);
+            tcx.sess.code_stats.borrow_mut().record_type_size(kind,


This is a bit of a hack but I can live with impure stats just like we have diagnostics.

eddyb · 2017-05-25T07:11:07Z

@bors r+

bors · 2017-05-25T07:11:07Z

📌 Commit 629812d has been approved by eddyb

bors · 2017-05-26T18:14:11Z

☔ The latest upstream changes (presumably #42245) made this pull request unmergeable. Please resolve the merge conflicts.

nikomatsakis · 2017-05-26T19:16:44Z

@bors r=eddyb

bors · 2017-05-26T19:16:45Z

💡 This pull request was already approved, no need to approve it again.

There's another pull request that is currently being tested, blocking this pull request: Use the improved submodule handling #42081

bors · 2017-05-26T19:16:45Z

📌 Commit 629812d has been approved by eddyb

Mark-Simulacrum · 2017-05-26T21:15:05Z

@bors r-

Should be a quick fix, this is just the UI tests updating due to no longer being counted:

[01:08:42] diff of stderr:
[01:08:42] 
[01:08:42]  error[E0391]: unsupported cyclic reference between types/traits detected
[01:08:42]    |
[01:08:42]  note: the cycle begins when computing layout of `S`...
[01:08:42]  note: ...which then requires computing layout of `std::option::Option<<S as Mirror>::It>`...
[01:08:42]  note: ...which then requires computing layout of `<S as Mirror>::It`...
[01:08:42]    = note: ...which then again requires computing layout of `S`, completing the cycle.
[01:08:42]  
[01:08:42] -error: aborting due to previous error
[01:08:42] +error: aborting due to previous error(s)
[01:08:42]  
[01:08:42] 
[01:08:42] The actual stderr differed from the expected stderr.
[01:08:42] 
[01:08:42] ------------------------------------------
[01:08:42] stderr:
[01:08:42] ------------------------------------------
[01:08:42] error[E0391]: unsupported cyclic reference between types/traits detected
[01:08:42]   |
[01:08:42] note: the cycle begins when computing layout of `S`...
[01:08:42] note: ...which then requires computing layout of `std::option::Option<<S as Mirror>::It>`...
[01:08:42] note: ...which then requires computing layout of `<S as Mirror>::It`...
[01:08:42]   = note: ...which then again requires computing layout of `S`, completing the cycle.
[01:08:42] 
[01:08:42] error: aborting due to previous error(s)
[01:08:42] 
[01:08:42] ------------------------------------------
[01:08:42] 
[01:08:42] thread '[ui] ui/issue-26548.rs' panicked at 'explicit panic', /checkout/src/tools/compiletest/src/runtest.rs:2472
[01:08:42] failures:
[01:08:42]     [ui] ui/issue-26548.rs

nikomatsakis · 2017-05-27T00:26:48Z

@bors r=eddyb

bors · 2017-05-27T00:26:49Z

📌 Commit 9b6baa5 has been approved by eddyb

bors · 2017-05-28T12:05:56Z

☔ The latest upstream changes (presumably #42276) made this pull request unmergeable. Please resolve the merge conflicts.

bors · 2017-05-30T07:19:27Z

🔒 Merge conflict

nikomatsakis · 2017-05-30T14:20:26Z

@bors r=eddyb

bors · 2017-05-30T14:20:27Z

📌 Commit 62e4733 has been approved by eddyb

bors · 2017-05-30T15:10:30Z

⌛ Testing commit 62e4733 with merge 8b63d0b...

@eddyb

querify layout and move param env out of the infcx The main goal of this PR is to move the parameter environment *out* of the inference context. This is because the inference environment will soon be changing over the course of inference --- for example, when we enter into a `for<'a> fn(...)` type, we will push a new environment with an increasing universe index, rather than skolemizing the `'a` references. Similarly, each obligation will soon be able to have a distinct parameter environment, and therefore the `Obligation` struct is extended to carry a `ParamEnv<'tcx>`. (I debated about putting it into the cause; seems plausible, but also weird.) Along the way, I also reworked how layout works, moving the layout cache into a proper query along the lines of needs-drop and friends. Finally, tweaks the inference context API. It seemed to be accumulating parameters at an alarming rate. The main way to e.g. make a subtype or equality relationship is to do the following: infcx.at(cause, param_env).sub(a, b) infcx.at(cause, param_env).eq(a, b) In both cases, `a` is considered the "expected" type (this used to be specified by a boolean). I tried hard to preserve the existing notion of what was "expected", although in some cases I'm not convinced it was being set on purpose one way or the other. This is why in some cases you will see me do `sup(b, a)`, which is otherwise equivalent to `sub(a, b)`, but sets the "expected type" differently. r? @eddyb cc @arielb1

bors · 2017-05-30T21:43:44Z

💔 Test failed - status-travis

Mark-Simulacrum · 2017-05-30T21:45:08Z

@bors retry

Travis timed out after 6+ hours due to current problems, I hope this might indicate they've started to be resolved.

bors · 2017-06-01T07:47:13Z

☔ The latest upstream changes (presumably #42348) made this pull request unmergeable. Please resolve the merge conflicts.

nikomatsakis · 2017-06-01T16:56:43Z

tfw you rebase something and, by the time you're done, there's another conflict

now we grow the type-sizes info during execution, rather than walking the cache after the fact

nikomatsakis · 2017-06-01T19:01:53Z

@bors r=eddyb

bors · 2017-06-01T19:01:54Z

📌 Commit 84047db has been approved by eddyb

bors · 2017-06-02T19:23:21Z

⌛ Testing commit 84047db with merge 107bd67...

@eddyb

querify layout and move param env out of the infcx The main goal of this PR is to move the parameter environment *out* of the inference context. This is because the inference environment will soon be changing over the course of inference --- for example, when we enter into a `for<'a> fn(...)` type, we will push a new environment with an increasing universe index, rather than skolemizing the `'a` references. Similarly, each obligation will soon be able to have a distinct parameter environment, and therefore the `Obligation` struct is extended to carry a `ParamEnv<'tcx>`. (I debated about putting it into the cause; seems plausible, but also weird.) Along the way, I also reworked how layout works, moving the layout cache into a proper query along the lines of needs-drop and friends. Finally, tweaks the inference context API. It seemed to be accumulating parameters at an alarming rate. The main way to e.g. make a subtype or equality relationship is to do the following: infcx.at(cause, param_env).sub(a, b) infcx.at(cause, param_env).eq(a, b) In both cases, `a` is considered the "expected" type (this used to be specified by a boolean). I tried hard to preserve the existing notion of what was "expected", although in some cases I'm not convinced it was being set on purpose one way or the other. This is why in some cases you will see me do `sup(b, a)`, which is otherwise equivalent to `sub(a, b)`, but sets the "expected type" differently. r? @eddyb cc @arielb1

bors · 2017-06-02T22:19:05Z

☀️ Test successful - status-appveyor, status-travis
Approved by: eddyb
Pushing 107bd67 to master...

rust-highfive assigned eddyb May 24, 2017

aidanhs added the S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. label May 24, 2017

eddyb reviewed May 25, 2017

View reviewed changes

nikomatsakis force-pushed the chalk-trait-env-param-env branch from 629812d to 5977dc3 Compare May 26, 2017 19:16

Mark-Simulacrum 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 28, 2017

carols10cents 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 May 29, 2017

nikomatsakis force-pushed the chalk-trait-env-param-env branch from 9b6baa5 to 62e4733 Compare May 30, 2017 14:10

nikomatsakis 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 May 31, 2017

nikomatsakis force-pushed the chalk-trait-env-param-env branch from 62e4733 to 952bc4b Compare June 1, 2017 15:18

nikomatsakis added 2 commits June 1, 2017 12:56

move projection mode into parameter environment

6c4b961

rewrite layout to take a (param-env, ty) pair instead of infcx

541523d

nikomatsakis added 6 commits June 1, 2017 13:31

strip param-env from infcx

5fb0f0d

remove layout_cache and fix -Z print-type-sizes

c1e895d

now we grow the type-sizes info during execution, rather than walking the cache after the fact

ergonomic improvements to the methods in infcx

c7a2e32

pacify the mercilous tidy

af223fa

fix librustc_driver tests

4c599bc

fix ui reference

84047db

nikomatsakis force-pushed the chalk-trait-env-param-env branch from 952bc4b to 84047db Compare June 1, 2017 17:31

bors merged commit 84047db into rust-lang:master Jun 2, 2017

This was referenced Jun 2, 2017

Print the two types in the span label for transmute errors. #42304

Merged

Downgrade ProjectionTy's TraitRef to its substs #42388

Merged

dwrensha mentioned this pull request Jun 3, 2017

update for upstream ParamEnv changes rust-lang/miri#179

Merged

kyren mentioned this pull request Mar 11, 2019

Fix #54822 and associated faulty tests #59083

Merged

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

querify layout and move param env out of the infcx #42189

querify layout and move param env out of the infcx #42189

nikomatsakis commented May 24, 2017

eddyb May 25, 2017

eddyb commented May 25, 2017

bors commented May 25, 2017

bors commented May 26, 2017

nikomatsakis commented May 26, 2017

bors commented May 26, 2017

bors commented May 26, 2017

Mark-Simulacrum commented May 26, 2017

nikomatsakis commented May 27, 2017

bors commented May 27, 2017

bors commented May 28, 2017

bors commented May 30, 2017

nikomatsakis commented May 30, 2017

bors commented May 30, 2017

bors commented May 30, 2017

bors commented May 30, 2017

Mark-Simulacrum commented May 30, 2017

bors commented Jun 1, 2017

nikomatsakis commented Jun 1, 2017

nikomatsakis commented Jun 1, 2017

bors commented Jun 1, 2017

bors commented Jun 2, 2017

bors commented Jun 2, 2017

querify layout and move param env out of the infcx #42189

querify layout and move param env out of the infcx #42189

Conversation

nikomatsakis commented May 24, 2017

eddyb May 25, 2017

Choose a reason for hiding this comment

eddyb commented May 25, 2017

bors commented May 25, 2017

bors commented May 26, 2017

nikomatsakis commented May 26, 2017

bors commented May 26, 2017

bors commented May 26, 2017

Mark-Simulacrum commented May 26, 2017

nikomatsakis commented May 27, 2017

bors commented May 27, 2017

bors commented May 28, 2017

bors commented May 30, 2017

nikomatsakis commented May 30, 2017

bors commented May 30, 2017

bors commented May 30, 2017

bors commented May 30, 2017

Mark-Simulacrum commented May 30, 2017

bors commented Jun 1, 2017

nikomatsakis commented Jun 1, 2017

nikomatsakis commented Jun 1, 2017

bors commented Jun 1, 2017

bors commented Jun 2, 2017

bors commented Jun 2, 2017