Silence follow-up errors directly based on error types and regions #125667
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
rustbot
added
S-waiting-on-review
This comment has been minimized.
This comment has been minimized.
|
@bors r+ |
bors
added
S-waiting-on-bors
bors
added a commit
to rust-lang-ci/rust
that referenced
this pull request
Jun 4, 2024
…mpiler-errors Rollup of 8 pull requests Successful merges: - rust-lang#125667 (Silence follow-up errors directly based on error types and regions) - rust-lang#125717 (Refactor `#[diagnostic::do_not_recommend]` support) - rust-lang#125795 (Improve renaming suggestion for names with leading underscores) - rust-lang#125865 (Fix ICE caused by ignoring EffectVars in type inference) - rust-lang#125953 (Streamline `nested` calls.) - rust-lang#125959 (Reduce `pub` exposure in `rustc_mir_build`) - rust-lang#125967 (Split smir `Const` into `TyConst` and `MirConst`) - rust-lang#125968 (Store the types of `ty::Expr` arguments in the `ty::Expr`) r? `@ghost` `@rustbot` modify labels: rollup
rust-timer
added a commit
to rust-lang-ci/rust
that referenced
this pull request
Jun 4, 2024
Rollup merge of rust-lang#125667 - oli-obk:taintify, r=TaKO8Ki Silence follow-up errors directly based on error types and regions During type_of, we used to just return an error type if there were any errors encountered. This is problematic, because it means a struct declared as `struct Foo<'static>` will end up not finding any inherent or trait impls because those impl blocks' `Self` type will be `{type error}` instead of `Foo<'re_error>`. Now it's the latter, silencing nonsensical follow-up errors about `Foo` not having any methods. Unfortunately that now allows for new follow-up errors, because borrowck treats `'re_error` as `'static`, causing nonsensical errors about non-error lifetimes not outliving `'static`. So what I also did was to just strip all outlives bounds that borrowck found, thus never letting it check them. There are probably more nuanced ways to do this, but I worried there would be other nonsensical errors if some outlives bounds were missing. Also from the test changes, it looked like an improvement everywhere.
This comment was marked as spam.
This comment was marked as spam.
Sorry, something went wrong.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
During type_of, we used to just return an error type if there were any errors encountered. This is problematic, because it means a struct declared as
struct Foo<'static>will end up not finding any inherent or trait impls because those impl blocks'Selftype will be{type error}instead ofFoo<'re_error>. Now it's the latter, silencing nonsensical follow-up errors aboutFoonot having any methods.Unfortunately that now allows for new follow-up errors, because borrowck treats
're_erroras'static, causing nonsensical errors about non-error lifetimes not outliving'static. So what I also did was to just strip all outlives bounds that borrowck found, thus never letting it check them. There are probably more nuanced ways to do this, but I worried there would be other nonsensical errors if some outlives bounds were missing. Also from the test changes, it looked like an improvement everywhere.