Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Convert delayed_bugs to bugs. #121208

Merged
merged 2 commits into from
Feb 21, 2024
Merged

Conversation

nnethercote
Copy link
Contributor

I have a suspicion that quite a few delayed bug paths are impossible to reach, so I did an experiment.

I converted every delayed_bug to a bug, ran the full test suite, then converted back every bug that was hit. A surprising number were never hit.

This is too dangerous to merge. Increased coverage (fuzzing or a crater run) would likely hit more cases. But it might be useful for people to look at and think about which paths are genuinely unreachable.

r? @ghost

@rustbot rustbot added S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. T-compiler Relevant to the compiler team, which will review and decide on the PR/issue. labels Feb 16, 2024
@rustbot
Copy link
Collaborator

rustbot commented Feb 16, 2024

Some changes might have occurred in exhaustiveness checking

cc @Nadrieril

Some changes occurred in const_evaluatable.rs

cc @BoxyUwU

Some changes occurred to MIR optimizations

cc @rust-lang/wg-mir-opt

Type relation code was changed

cc @compiler-errors, @lcnr

@nnethercote
Copy link
Contributor Author

cc @compiler-errors @oli-obk @lcnr

@nnethercote nnethercote marked this pull request as draft February 16, 2024 23:40
@nnethercote nnethercote changed the title Convert delayed_bugs to bugs. [Experiment] Convert delayed_bugs to bugs. Feb 16, 2024
@rust-log-analyzer

This comment has been minimized.

Copy link
Contributor

@lcnr lcnr left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I think we should land this with at least a few of these converted to bugs. I think doing so is fine because either:

  1. I am fairly confident that this can be a bug!, the code is familiar and I believe it to be unreachable.
  2. anything triggering that ICE should be self-contained and easy to minimize, meaning that even if it's reachable, we should be able to get a new ui test fairly quickly, which is worth it.

Went through all changes rn and I would like to push this through and merge it after handling all review comments. we still have nearly a month until the next beta cutoff, so we could merge this now.

self.dcx()
.span_delayed_bug(lifetime.ident.span, "no def-id for fresh lifetime");
continue;
self.dcx().span_bug(lifetime.ident.span, "no def-id for fresh lifetime");
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

2

@@ -628,7 +628,7 @@ impl<'tcx> MirBorrowckCtxt<'_, 'tcx> {
) => {
// HIR lowering sometimes doesn't catch this in erroneous
// programs, so we need to use span_delayed_bug here. See #82126.
self.dcx().span_delayed_bug(
self.dcx().span_bug(
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

the original UI test for #82126 does not trigger this anymore, so we should get a new test if it's still reachable. Please update the comment before merge

2

@@ -316,8 +316,7 @@ impl<'tcx> UniversalRegionRelationsBuilder<'_, 'tcx> {
.and(type_op::normalize::Normalize::new(ty))
.fully_perform(self.infcx, span)
else {
tcx.dcx().span_delayed_bug(span, format!("failed to normalize {ty:?}"));
continue;
tcx.dcx().span_bug(span, format!("failed to normalize {ty:?}"));
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

feels like this should be reachable and potentially hard to minimize, please keep this a span_delayed_bug but maybe add a comment "this is currently not reached in any test, so it may be worth it to minimize an example triggering this".

@@ -154,8 +154,7 @@ impl<'a, 'tcx> TypeChecker<'a, 'tcx> {
if argument_index + 1 >= body.local_decls.len() {
self.tcx()
.dcx()
.span_delayed_bug(body.span, "found more normalized_input_ty than local_decls");
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

1

trace!("finalized opaque type {:?} to {:#?}", opaque_type_key, hidden_type.ty.kind());
if hidden_type.has_non_region_infer() {
let reported = infcx.dcx().span_delayed_bug(
infcx.dcx().span_bug(
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

1

not as confident as I maybe should be, but I would really like a test for cases where this is triggered if reachable.

@@ -2016,11 +2016,10 @@ fn assemble_candidates_from_impls<'cx, 'tcx>(
ImplSource::Builtin(BuiltinImplSource::TraitUpcasting { .. }, _)
| ImplSource::Builtin(BuiltinImplSource::TupleUnsizing, _) => {
// These traits have no associated types.
selcx.tcx().dcx().span_delayed_bug(
selcx.tcx().dcx().span_bug(
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

probably theoretically reachable when manually redefining their lang item 🤷 so we should keep it as a delay span bug

@@ -82,7 +82,7 @@ where
let value = infcx.commit_if_ok(|_| {
let ocx = ObligationCtxt::new(infcx);
let value = op(&ocx).map_err(|_| {
infcx.dcx().span_delayed_bug(span, format!("error performing operation: {name}"))
infcx.dcx().span_bug(span, format!("error performing operation: {name}"))
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

theoretically reachable I think, may make sense to still keep this as is to get a test 🤷 easier to just keep a span_delay_bug here

));
infcx
.dcx()
.span_bug(span, format!("ambiguity processing {obligations:?} from {self:?}"));
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

i think it's currently unreachable, but that may change going forward. Still Making it a bug seems alright to me

@@ -90,8 +90,7 @@ fn univariant_uninterned<'tcx>(
let dl = cx.data_layout();
let pack = repr.pack;
if pack.is_some() && repr.align.is_some() {
cx.tcx.dcx().delayed_bug("struct cannot be packed and aligned");
return Err(cx.tcx.arena.alloc(LayoutError::Unknown(ty)));
cx.tcx.dcx().bug("struct cannot be packed and aligned");
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

🤷 2

@@ -337,7 +337,7 @@ impl<'tcx> TypeVisitor<TyCtxt<'tcx>> for ImplTraitInAssocTypeCollector<'tcx> {
.instantiate(self.0.tcx, impl_args)
.visit_with(self);
} else {
self.0.tcx.dcx().span_delayed_bug(
self.0.tcx.dcx().span_bug(
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

🤷 2

@lcnr
Copy link
Contributor

lcnr commented Feb 19, 2024

r? @lcnr

@nnethercote
Copy link
Contributor Author

Thanks for the amazingly thorough review! I think I have made all the changes you requested.

@matthiaskrgr: would you mind fuzzing this for a bit before we merge? It would be good to get some extra coverage.

@bors try

bors added a commit to rust-lang-ci/rust that referenced this pull request Feb 19, 2024
[Experiment] Convert `delayed_bug`s to `bug`s.

I have a suspicion that quite a few delayed bug paths are impossible to reach, so I did an experiment.

I converted every `delayed_bug` to a `bug`, ran the full test suite, then converted back every `bug` that was hit. A surprising number were never hit.

This is too dangerous to merge. Increased coverage (fuzzing or a crater run) would likely hit more cases. But it might be useful for people to look at and think about which paths are genuinely unreachable.

r? `@ghost`
@bors
Copy link
Contributor

bors commented Feb 19, 2024

⌛ Trying commit 5f00f10 with merge 4d96544...

@nnethercote nnethercote marked this pull request as ready for review February 19, 2024 23:25
@nnethercote nnethercote changed the title [Experiment] Convert delayed_bugs to bugs. Convert delayed_bugs to bugs. Feb 19, 2024
@bors
Copy link
Contributor

bors commented Feb 20, 2024

☀️ Try build successful - checks-actions
Build commit: 4d96544 (4d96544a124b8a36849185bddcfe6f87520c120e)

@matthiaskrgr
Copy link
Member

matthiaskrgr commented Feb 20, 2024

I've run this through roughly 850000 files x 6 different sets of rustcflags on the dev desktop server and didn't see any crashes that would strike me as new 👍
around 650 crashes were found in total, coming down to around 50 "unique" crashes.

@@ -48,8 +48,10 @@ impl<'a, 'tcx> FnCtxt<'a, 'tcx> {
let to = normalize(to);
trace!(?from, ?to);
if from.has_non_region_infer() || to.has_non_region_infer() {
// Note: this path is currently not reached in any test, so any
// example that triggers this would be worth minimizing and
// converting into a test.
tcx.dcx().span_delayed_bug(span, "argument to transmute has inference variables");
Copy link
Contributor

@lcnr lcnr Feb 20, 2024

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

should this not be a span_bug? After removing the return we now use the below code even if there are infer vars

} else {
ty.needs_drop(tcx, tcx.param_env(item.owner_id))
}
assert!(!ty.has_infer());
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Suggested change
assert!(!ty.has_infer());

@@ -172,6 +172,7 @@ impl<'cx, 'tcx> VerifyBoundCx<'cx, 'tcx> {
self.bound_from_components(components, visited)
}
Component::UnresolvedInferenceVariable(v) => {
// njn: comment?
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

comment?

format!("unexpected inference var {b:?}",),
);
Ok(a)
self.infcx.dcx().bug(format!("unexpected inference var {b:?}"));
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

bug vs span_bug?

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

You want me to use span_bug? What span would I use?

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

self.delegate.span()? the same we previously used in the span_delayed_bug 🤔

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Oh, I removed that because of this comment. Maybe I misinterpreted what you were asking for.

Copy link
Contributor

@lcnr lcnr Feb 22, 2024

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

ah xd, yeah 😁 i messed that up, didn't look at the context in my second review 😅 sorry

@@ -42,7 +42,7 @@ fn opt_span_bug_fmt<S: Into<MultiSpan>>(
/// delayed bug, so what is the point of this? It exists to help us test the interaction of delayed
/// bugs with the query system and incremental.
pub fn trigger_delayed_bug(tcx: TyCtxt<'_>, key: rustc_hir::def_id::DefId) {
tcx.dcx().span_delayed_bug(
tcx.dcx().span_bug(
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

still needs to get converted back

.size;
let width = tcx
.layout_of(param_ty)
.expect(&format!("couldn't compute width of literal: {:?}", lit_input.lit))
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

now eagerly formats even in the success case, which is unnecessary work and may negatively impact perf

.size;
let width = tcx
.layout_of(param_ty)
.expect(&format!("couldn't compute width of literal: {:?}", lit_input.lit))
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

same here

@@ -3682,13 +3682,13 @@ impl<'a: 'ast, 'b, 'ast, 'tcx> LateResolutionVisitor<'a, 'b, 'ast, 'tcx> {
None
}
Res::SelfCtor(_) => {
// njn: remove comment?
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

still relevant

Comment on lines 65 to 69
// njn: ?
// FIXME(generic_const_exprs): we have a `ConstKind::Expr` which is fully concrete,
// but currently it is not possible to evaluate `ConstKind::Expr` so we are unable
// to tell if it is evaluatable or not. For now we just ICE until this is
// implemented.
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Suggested change
// njn: ?
// FIXME(generic_const_exprs): we have a `ConstKind::Expr` which is fully concrete,
// but currently it is not possible to evaluate `ConstKind::Expr` so we are unable
// to tell if it is evaluatable or not. For now we just ICE until this is
// implemented.
// FIXME(generic_const_exprs): we have a fully concrete `ConstKind::Expr`,
// but haven't implemented evaluating `ConstKind::Expr` yet, so we
// are unable to tell if it is evaluatable or not. As this is
// unreachable for now, we can simple ICE here.

@nnethercote
Copy link
Contributor Author

@lcnr: Ok, I think I've addressed all the new comments, except for this one.

@matthiaskrgr: thank you for the fuzzing! That's very helpful.

@lcnr
Copy link
Contributor

lcnr commented Feb 20, 2024

r=me after addressing the last comment

I have a suspicion that quite a few delayed bug paths are impossible to
reach, so I did an experiment.

I converted every `delayed_bug` to a `bug`, ran the full test suite,
then converted back every `bug` that was hit. A surprising number were
never hit.

The next commit will convert some more back, based on human judgment.
This commit undoes some of the previous commit's mechanical changes,
based on human judgment.
@nnethercote
Copy link
Contributor Author

@bors r=lcnr

@bors
Copy link
Contributor

bors commented Feb 21, 2024

📌 Commit 2903bbb has been approved by lcnr

It is now in the queue for this repository.

@bors bors removed the S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. label Feb 21, 2024
matthiaskrgr added a commit to matthiaskrgr/rust that referenced this pull request Feb 23, 2024
…lcnr

Fix rust-lang#121208 fallout

rust-lang#121208 converted lots of delayed bugs to bugs. Unsurprisingly, there were a few invalid conversion found via fuzzing.

r? `@lcnr`
matthiaskrgr added a commit to matthiaskrgr/rust that referenced this pull request Feb 23, 2024
…t, r=lcnr

Fix more rust-lang#121208 fallout

rust-lang#121208 converted lots of delayed bugs to bugs. Unsurprisingly, there were a few invalid conversion found via fuzzing.

r? `@lcnr`
bors added a commit to rust-lang-ci/rust that referenced this pull request Feb 23, 2024
…iaskrgr

Rollup of 8 pull requests

Successful merges:

 - rust-lang#121434 (Fix rust-lang#121208 fallout)
 - rust-lang#121471 (When encountering `<&T as Clone>::clone(x)` because `T: Clone`, suggest `#[derive(Clone)]`)
 - rust-lang#121476 (remove `llvm.assertions=true` in compiler profile)
 - rust-lang#121479 (fix generalizer unsoundness)
 - rust-lang#121480 (Fix more rust-lang#121208 fallout)
 - rust-lang#121482 (Allow for a missing `adt_def` in `NamePrivacyVisitor`.)
 - rust-lang#121484 (coverage: Use variable name `this` in `CoverageGraph::from_mir`)
 - rust-lang#121487 (Explicitly call `emit_stashed_diagnostics`.)

r? `@ghost`
`@rustbot` modify labels: rollup
bors added a commit to rust-lang-ci/rust that referenced this pull request Feb 23, 2024
…iaskrgr

Rollup of 8 pull requests

Successful merges:

 - rust-lang#121434 (Fix rust-lang#121208 fallout)
 - rust-lang#121471 (When encountering `<&T as Clone>::clone(x)` because `T: Clone`, suggest `#[derive(Clone)]`)
 - rust-lang#121476 (remove `llvm.assertions=true` in compiler profile)
 - rust-lang#121479 (fix generalizer unsoundness)
 - rust-lang#121480 (Fix more rust-lang#121208 fallout)
 - rust-lang#121482 (Allow for a missing `adt_def` in `NamePrivacyVisitor`.)
 - rust-lang#121484 (coverage: Use variable name `this` in `CoverageGraph::from_mir`)
 - rust-lang#121487 (Explicitly call `emit_stashed_diagnostics`.)

r? `@ghost`
`@rustbot` modify labels: rollup
rust-timer added a commit to rust-lang-ci/rust that referenced this pull request Feb 23, 2024
Rollup merge of rust-lang#121480 - nnethercote:fix-more-121208-fallout, r=lcnr

Fix more rust-lang#121208 fallout

rust-lang#121208 converted lots of delayed bugs to bugs. Unsurprisingly, there were a few invalid conversion found via fuzzing.

r? `@lcnr`
rust-timer added a commit to rust-lang-ci/rust that referenced this pull request Feb 23, 2024
Rollup merge of rust-lang#121434 - nnethercote:fix-121208-fallout, r=lcnr

Fix rust-lang#121208 fallout

rust-lang#121208 converted lots of delayed bugs to bugs. Unsurprisingly, there were a few invalid conversion found via fuzzing.

r? `@lcnr`
GuillaumeGomez added a commit to GuillaumeGomez/rust that referenced this pull request Feb 26, 2024
…allout, r=lcnr

Fix more rust-lang#121208 fallout (round 3)

rust-lang#121208 converted lots of delayed bugs to bugs. Unsurprisingly, there were a few invalid conversion found via fuzzing.

r? `@lcnr`
bors added a commit to rust-lang-ci/rust that referenced this pull request Feb 26, 2024
…llaumeGomez

Rollup of 5 pull requests

Successful merges:

 - rust-lang#120656 (Allow tests to specify a `//@ filecheck-flags:` header)
 - rust-lang#120840 (Split Diagnostics for Uncommon Codepoints: Add Individual Identifier Types)
 - rust-lang#121554 (Don't unnecessarily change `SIGPIPE` disposition in `unix_sigpipe` tests)
 - rust-lang#121590 (Correctly handle if rustdoc JS script hash changed)
 - rust-lang#121620 (Fix more rust-lang#121208 fallout (round 3))

r? `@ghost`
`@rustbot` modify labels: rollup
rust-timer added a commit to rust-lang-ci/rust that referenced this pull request Feb 26, 2024
Rollup merge of rust-lang#121620 - nnethercote:fix-even-more-121208-fallout, r=lcnr

Fix more rust-lang#121208 fallout (round 3)

rust-lang#121208 converted lots of delayed bugs to bugs. Unsurprisingly, there were a few invalid conversion found via fuzzing.

r? `@lcnr`
nnethercote added a commit to nnethercote/rust that referenced this pull request Jun 17, 2024
PR rust-lang#121208 converted this from a `span_delayed_bug` to a `span_bug`
because nothing in the test suite caused execution to hit this path. But
now fuzzing has found a test case that does hit it. So this commit
converts it back to `span_delayed_bug` and adds the relevant test.
nnethercote added a commit to nnethercote/rust that referenced this pull request Jun 17, 2024
PR rust-lang#121208 converted this from a `span_delayed_bug` to a `span_bug`
because nothing in the test suite caused execution to hit this path. But
now fuzzing has found a test case that does hit it. So this commit
converts it back to `span_delayed_bug` and adds the relevant test.

Fixes rust-lang#126385.
nnethercote added a commit to nnethercote/rust that referenced this pull request Jun 17, 2024
PR rust-lang#121208 converted this from a `span_delayed_bug` to a `span_bug`
because nothing in the test suite caused execution to hit this path. But
now fuzzing has found a test case that does hit it. So this commit
converts it back to `span_delayed_bug` and adds the relevant test.

Fixes rust-lang#126385.
nnethercote added a commit to nnethercote/rust that referenced this pull request Jun 17, 2024
PR rust-lang#121208 converted this from a `span_delayed_bug` to a `span_bug`
because nothing in the test suite caused execution to hit this path. But
now fuzzing has found a test case that does hit it. So this commit
converts it back to `span_delayed_bug` and adds the relevant test.

Fixes rust-lang#126385.
GuillaumeGomez added a commit to GuillaumeGomez/rust that referenced this pull request Jun 17, 2024
Convert a `span_bug` to a `span_delayed_bug`.

PR rust-lang#121208 converted this from a `span_delayed_bug` to a `span_bug` because nothing in the test suite caused execution to hit this path. But now fuzzing has found a test case that does hit it. So this commit converts it back to `span_delayed_bug` and adds the relevant test.

Fixes rust-lang#126385.

r? `@lcnr`
rust-timer added a commit to rust-lang-ci/rust that referenced this pull request Jun 17, 2024
Rollup merge of rust-lang#126570 - nnethercote:fix-126385, r=lcnr

Convert a `span_bug` to a `span_delayed_bug`.

PR rust-lang#121208 converted this from a `span_delayed_bug` to a `span_bug` because nothing in the test suite caused execution to hit this path. But now fuzzing has found a test case that does hit it. So this commit converts it back to `span_delayed_bug` and adds the relevant test.

Fixes rust-lang#126385.

r? `@lcnr`
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
S-waiting-on-bors Status: Waiting on bors to run and complete tests. Bors will change the label on completion. T-compiler Relevant to the compiler team, which will review and decide on the PR/issue.
Projects
None yet
Development

Successfully merging this pull request may close these issues.

7 participants