Panic when typing the names on a tuple: index out of bounds: the len is 0 but the index is 0
#16746
Comments
|
seems relate to thread 'Worker' panicked at /github/home/.cargo/registry/src/index.crates.io-6f17d22bba15001f/ra-ap-rustc_pattern_analysis-0.40.0/src/usefulness.rs:954:18:
index out of bounds: the len is 0 but the index is 0
stack backtrace:
0: rust_begin_unwind
at /rustc/07dca489ac2d933c78d3c5158e3f43beefeb02ce/library/std/src/panicking.rs:645:5
1: core::panicking::panic_fmt
at /rustc/07dca489ac2d933c78d3c5158e3f43beefeb02ce/library/core/src/panicking.rs:72:14
2: core::panicking::panic_bounds_check
at /rustc/07dca489ac2d933c78d3c5158e3f43beefeb02ce/library/core/src/panicking.rs:208:5
3: ra_ap_rustc_pattern_analysis::usefulness::PlaceInfo<Cx>::split_column_ctors
4: ra_ap_rustc_pattern_analysis::usefulness::compute_exhaustiveness_and_usefulness
5: ra_ap_rustc_pattern_analysis::usefulness::compute_exhaustiveness_and_usefulness
6: ra_ap_rustc_pattern_analysis::usefulness::compute_exhaustiveness_and_usefulness
7: ra_ap_rustc_pattern_analysis::usefulness::compute_match_usefulness
8: hir_ty::diagnostics::expr::BodyValidationDiagnostic::collect |
|
😭 I'll have a look tomorrow. Fwiw: it would be really nice if the typechecker could be trusted to skip match checking when there's a type error, cause that's always the source of the pattern_analysis panics |
… r=compiler-errors pattern analysis: abort on arity mismatch This is one more PR replacing panics by `Err()` aborts. I recently audited all the `unwrap()` calls, but I had forgotten about array accesses. (Again [discovered by rust-analyzer](rust-lang/rust-analyzer#16746)). r? `@compiler-errors`
… r=compiler-errors pattern analysis: abort on arity mismatch This is one more PR replacing panics by `Err()` aborts. I recently audited all the `unwrap()` calls, but I had forgotten about array accesses. (Again [discovered by rust-analyzer](rust-lang/rust-analyzer#16746)). r? ``@compiler-errors``
… r=compiler-errors pattern analysis: abort on arity mismatch This is one more PR replacing panics by `Err()` aborts. I recently audited all the `unwrap()` calls, but I had forgotten about array accesses. (Again [discovered by rust-analyzer](rust-lang/rust-analyzer#16746)). r? ```@compiler-errors```
We want as much as possible to work if there are errors in unrelated parts of the function though 😅 Of course if the actual match has type mismatches there's not much point in trying to do further checking. |
|
It should probably suffice to skip it if the scrutinee contains an |
Rollup merge of rust-lang#121987 - Nadrieril:abort-on-arity-mismatch, r=compiler-errors pattern analysis: abort on arity mismatch This is one more PR replacing panics by `Err()` aborts. I recently audited all the `unwrap()` calls, but I had forgotten about array accesses. (Again [discovered by rust-analyzer](rust-lang/rust-analyzer#16746)). r? ```@compiler-errors```
|
I was assuming both of these things were already what we're doing, but I haven't actually checked |
|
Only for error types, not for types that contain errors And exhaustiveness checking for |
|
Oh, that explains things :D |
|
Sorry for poking my nose into discussion, though I'm getting quite similar error on a medium-sized project. Is there some way to determine what piece of source causes error? In case anyone's interested, I'm getting following error trace: |
|
There's no easy way to know which code triggered it unfortunately I think. This has now been fixed on the rustc side, the fix will come to rust-analyzer next update i.e. within a week. @Veykril if you know how to enforce the type check, please do, I haven't looked at the type-checking side of r-a at all |
|
should be a simple - if scrut_ty.is_unknown() {
+ if scrut_ty.contains_unknown() {for And adding the same check for the let checking
|
|
Fixed in #16778. |
|
I forgot to add a regression test. Do you know where I can add on? I haven't figured out r-a's testing situation yet |
|
https://github.com/rust-lang/rust-analyzer/blob/master/crates/ide-diagnostics/src/handlers/missing_match_arms.rs is where we test match checking (let exhaustiveness checking has its own diagnostic file) |
…r=Veykril minor: Skip match diagnostics for partially unknown types CC #16746 (comment)
When deleting characters from the name of one element in a tuple, if you delete the last one a panic will occur.
To reproduce, create a new empty project and add this to your
main.rs:Then place your cursor right after the 'b' character and press backspace, making your code look like this:
This panic message is printed to stderr.
rust-analyzer version: 0.4.1867-standalone (037924c 2024-03-03)
rustc version: rustc 1.78.0-nightly (516b6162a 2024-03-03)
(sidenote: I know this is the third issue about panics from me in the past 5 days, I apologize if I'm getting annoying)
The text was updated successfully, but these errors were encountered: