[WIP] Experiment: stabilize never type #79470

nikomatsakis · 2020-11-27T15:39:10Z

Follow-up to #79366 for the purposes of running crater

We used to avoid doing this because we didn't want to make coercion depend on the state of inference. For better or worse, we have moved away from this position over time. Therefore, I am going to go ahead and resolve the `b` target type early on so that it is done uniformly. (The older technique for managing this was always something of a hack regardless; if we really wanted to avoid integrating coercion and inference we needed to be more disciplined about it.)

I didn't like the sub-unify code executing when a predicate was ENQUEUED, that felt fragile. I would have preferred to move the sub-unify code so that it only occurred during generalization, but that impacted diagnostics, so having it also occur when we process subtype predicates felt pretty reasonable. (I guess we only need one or the other, but I kind of prefer both, since the generalizer ultimately feels like the *right* place to guarantee the properties we want.)

Motivation: in upcoming commits, we are going to create a graph of the coercion relationships between variables. We want to distinguish *coercion* specifically from other sorts of subtyping, as it indicates values flowing from one place to another via assignment.

Along the way, simplify and document the logic more clearly.

We now fallback type variables using the following rules: * Construct a coercion graph `A -> B` where `A` and `B` are unresolved type variables or the `!` type. * Let D be those variables that are reachable from `!`. * Let N be those variables that are reachable from a variable not in D. * All variables in (D \ N) fallback to `!`. * All variables in (D & N) fallback to `()`.

The comment seems incorrect. Testing revealed that the examples in question still work (as well as some variants) even without the special casing here.

Instead, we now record those type variables that are the target of a `NeverToAny` adjustment and consider those to be the "diverging" type variables. This allows us to remove the special case logic that creates a type variable for `!` in coercion.

Instead, we now rely on the code that looks for a NeverToAny adjustment.

This could cause us to visit the start node twice.

Extend the `DepthFirstSearch` iterator so that it can be re-used and extended with add'l start nodes. Then replace the FxHashSets of nodes we were using in the fallback analysis with a single iterator. This way we won't re-walk portions of the graph that are reached more than once, and we also do less allocation etc.

rust-highfive · 2020-11-27T15:39:14Z

r? @oli-obk

(rust-highfive has picked a reviewer for you, use r? to override)

nikomatsakis · 2020-11-27T15:39:19Z

@bors try

nikomatsakis · 2020-11-27T15:39:25Z

r? @ghost

bors · 2020-11-27T15:39:30Z

⌛ Trying commit 5ef3eb6 with merge 449cf216550e3ee885a40c684da555a6ab31a065...

bors · 2020-11-27T16:22:53Z

☀️ Try build successful - checks-actions
Build commit: 449cf216550e3ee885a40c684da555a6ab31a065 (449cf216550e3ee885a40c684da555a6ab31a065)

gilescope · 2020-11-28T07:30:22Z

compiler/rustc_codegen_cranelift/example/mini_core_hello_world.rs

+    _FC(X),
+    _FD(X),
+    _FE(X),
+    _FF(X),

    V3,


Should we be fixing rustfmt here or telling rustfmt to skip this?

I think you'll want to post any comments on #79366 as this PR is just for the purpose of running crater

oli-obk · 2020-11-28T08:47:05Z

I would start crater, but I don't know if just a check suffices or if we also need to run all tests

nikomatsakis · 2020-11-28T14:48:03Z

@craterbot build-and-test

check would probably suffice but it is possible for these changes to change the semantics of running code and so forth.

craterbot · 2020-11-28T14:48:04Z

🚨 Error: failed to parse the command

🆘 If you have any trouble with Crater please ping @rust-lang/infra!
ℹ️ Crater is a tool to run experiments across parts of the Rust ecosystem. Learn more

nikomatsakis · 2020-11-30T14:08:16Z

@bors try

bors · 2020-11-30T14:08:43Z

⌛ Trying commit 3ad1e4d with merge b2f80f6008cdafc35b4472bd46afa16cc6b0cd63...

bors · 2020-11-30T14:55:54Z

☀️ Try build successful - checks-actions
Build commit: b2f80f6008cdafc35b4472bd46afa16cc6b0cd63 (b2f80f6008cdafc35b4472bd46afa16cc6b0cd63)

nikomatsakis · 2020-11-30T22:56:44Z

@craterbot check

I'm going to start with a check build and we can consider a run-and-test check later.

craterbot · 2020-11-30T22:56:54Z

👌 Experiment pr-79470 created and queued.
🤖 Automatically detected try build b2f80f6008cdafc35b4472bd46afa16cc6b0cd63
⚠️ Try build based on commit b7ebc6b, but latest commit is 3ad1e4d. Did you forget to make a new try build?
🔍 You can check out the queue and this experiment's details.

ℹ️ Crater is a tool to run experiments across parts of the Rust ecosystem. Learn more

nikomatsakis · 2020-12-01T14:08:48Z

Seems like it'll be a while :)

craterbot · 2020-12-11T01:25:16Z

🚧 Experiment pr-79470 is now running

ℹ️ Crater is a tool to run experiments across parts of the Rust ecosystem. Learn more

craterbot · 2020-12-14T11:49:49Z

🎉 Experiment pr-79470 is completed!
📊 119 regressed and 8 fixed (132594 total)
📰 Open the full report.

⚠️ If you notice any spurious failure please add them to the blacklist!
ℹ️ Crater is a tool to run experiments across parts of the Rust ecosystem. Learn more

nikomatsakis · 2020-12-14T15:12:43Z

Relatively few root failures, but definitely some impact. I did a bit of digging but will have to read in more carefully.

One sample error (chmlib v1.0.0):

[INFO] [stdout] error[E0277]: the trait bound `Continuation: From<!>` is not satisfied
[INFO] [stdout]    --> src/lib.rs:485:13
[INFO] [stdout]     |
[INFO] [stdout] 485 |         chm.for_each(Filter::all(), |_, _| panic!("Oops..."))
[INFO] [stdout]     |             ^^^^^^^^ the trait `From<!>` is not implemented for `Continuation`
[INFO] [stdout]     |
[INFO] [stdout]     = help: the following implementations were found:
[INFO] [stdout]               <Continuation as From<()>>
[INFO] [stdout]               <Continuation as From<std::result::Result<(), E>>>
[INFO] [stdout]     = note: required because of the requirements on the impl of `Into<Continuation>` for `!`

tema3210 · 2020-12-29T01:02:48Z

AFAIK there are three root failures: "trait is not implemented for !", "never type was propageted as a generic param to smth." and type inference failures.

The first point was discussed before
Second is the major problem: tainting generic parameters. For example: here the anonymous future type got a never return type (and caused error). The point is that these error are caused by proper usage of never type: regressed futures used a loop, which is an expression of type never. I guess to fix this we have either to invent some crazy schema or to break part of a world in the next edition.
An inference failure only happened here.

nikomatsakis added 17 commits November 19, 2020 18:24

factor fallback code into its own module

1cbde85

convert to a method on FnCtxt

0005746

introduce a Diverging enum instead of a bool

d8a881b

cleanup: extract a helper for coercion from inference variables

b3f8d74

introduce a Coerce predicate

448eb37

make TyVid implement the Idx trait

de3bf35

move fallback_if_possible and friends to fallback.rs

73b743b

Along the way, simplify and document the logic more clearly.

remove diverging type variables from fn check

8eda90e

The comment seems incorrect. Testing revealed that the examples in question still work (as well as some variants) even without the special casing here.

stop categorizing inference variables as diverging when created

fc02f0f

Instead, we now rely on the code that looks for a NeverToAny adjustment.

fix bug in DepthFirstSearch where start node was not visited

128de40

This could cause us to visit the start node twice.

stabilize feature(never_type)

5ef3eb6

rust-highfive assigned oli-obk Nov 27, 2020

rust-highfive added the S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. label Nov 27, 2020

nikomatsakis mentioned this pull request Nov 27, 2020

conditional fallback for the ! type #79366

Closed

gilescope reviewed Nov 28, 2020

View reviewed changes

craterbot added S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. and removed S-waiting-on-crater Status: Waiting on a crater run to be completed. labels Nov 29, 2020

WIP: make Infallible an alias for !

3ad1e4d

craterbot added S-waiting-on-crater Status: Waiting on a crater run to be completed. and removed S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. labels Nov 30, 2020

craterbot added S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. and removed S-waiting-on-crater Status: Waiting on a crater run to be completed. labels Dec 14, 2020

crlf0710 added S-experimental Status: Ongoing experiment that does not require reviewing and won't be merged in its current state. and removed S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. labels Dec 18, 2020

crlf0710 added S-experimental Status: Ongoing experiment that does not require reviewing and won't be merged in its current state. and removed S-experimental Status: Ongoing experiment that does not require reviewing and won't be merged in its current state. labels Jan 15, 2021

bstrie mentioned this pull request Jan 25, 2021

Tracking issue for promoting ! to a type (RFC 1216) #35121

Open

2 tasks

ehuss mentioned this pull request Jan 25, 2021

Remove the never type documentation rust-lang/reference#945

Open

crlf0710 added S-experimental Status: Ongoing experiment that does not require reviewing and won't be merged in its current state. and removed S-experimental Status: Ongoing experiment that does not require reviewing and won't be merged in its current state. labels Mar 5, 2021

Mark-Simulacrum closed this Mar 30, 2021

Mark-Simulacrum mentioned this pull request Apr 26, 2021

[WIP] Never type experiments #84573

Closed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

[WIP] Experiment: stabilize never type #79470

[WIP] Experiment: stabilize never type #79470

nikomatsakis commented Nov 27, 2020

rust-highfive commented Nov 27, 2020

nikomatsakis commented Nov 27, 2020

nikomatsakis commented Nov 27, 2020

bors commented Nov 27, 2020

bors commented Nov 27, 2020

gilescope Nov 28, 2020

oli-obk Nov 28, 2020

oli-obk commented Nov 28, 2020

nikomatsakis commented Nov 28, 2020

craterbot commented Nov 28, 2020

nikomatsakis commented Nov 30, 2020

bors commented Nov 30, 2020

bors commented Nov 30, 2020

nikomatsakis commented Nov 30, 2020

craterbot commented Nov 30, 2020

nikomatsakis commented Dec 1, 2020

craterbot commented Dec 11, 2020

craterbot commented Dec 14, 2020

nikomatsakis commented Dec 14, 2020

tema3210 commented Dec 29, 2020

+                  _FC(X),
+                  _FD(X),
+                  _FE(X),
+                  _FF(X),
                   V3,

[WIP] Experiment: stabilize never type #79470

[WIP] Experiment: stabilize never type #79470

Conversation

nikomatsakis commented Nov 27, 2020

rust-highfive commented Nov 27, 2020

nikomatsakis commented Nov 27, 2020

nikomatsakis commented Nov 27, 2020

bors commented Nov 27, 2020

bors commented Nov 27, 2020

gilescope Nov 28, 2020

Choose a reason for hiding this comment

oli-obk Nov 28, 2020

Choose a reason for hiding this comment

oli-obk commented Nov 28, 2020

nikomatsakis commented Nov 28, 2020

craterbot commented Nov 28, 2020

nikomatsakis commented Nov 30, 2020

bors commented Nov 30, 2020

bors commented Nov 30, 2020

nikomatsakis commented Nov 30, 2020

craterbot commented Nov 30, 2020

nikomatsakis commented Dec 1, 2020

craterbot commented Dec 11, 2020

craterbot commented Dec 14, 2020

nikomatsakis commented Dec 14, 2020

tema3210 commented Dec 29, 2020