At the moment, this moves only a single diagnostic, but the idea is
reafactor the rest to use the same pattern. We are going to have a
single file per diagnostic. This file will define diagnostics code,
rendering range and fixes, if any. It'll also have all of the tests.
This is similar to how we deal with assists.
After we refactor all diagnostics to follow this pattern, we'll probably
move them to a new `ide_diagnostics` crate.
Not that we intentionally want to test all diagnostics on this layer,
despite the fact that they are generally emitted in the guts on the
compiler. Diagnostics care to much about the end presentation
details/fixes to be worth-while "unit" testing. So, we'll unit-test only
the primary output of compilation process (types and name res tables),
and will use integrated UI tests for diagnostics.
9244: feat: Make block-local trait impls work r=flodiebold a=flodiebold
As long as either the trait or the implementing type are defined in the same block.
CC #8961
Co-authored-by: Florian Diebold <flodiebold@gmail.com>
9239: fix: Fix coercion in match with expected type r=flodiebold a=flodiebold
Plus add infrastructure to test type mismatches without expect.
CC #8961
Co-authored-by: Florian Diebold <flodiebold@gmail.com>
9204: feat: more accurate memory usage info on glibc Linux r=jonas-schievink a=jonas-schievink
This adds support for the new `mallinfo2` API added in glibc 2.33. It addresses a shortcoming in the `mallinfo` API where it was unable to handle memory usage of more than 2 GB, which we sometimes exceed.
Blocked on https://github.com/rust-lang/libc/pull/2228
Co-authored-by: Jonas Schievink <jonasschievink@gmail.com>
9192: internal: Build test-macros in a build script r=jonas-schievink a=jonas-schievink
This build the test-proc-macros in `proc_macro_test` in a build script, and copies the artifact to `OUT_DIR`. This should make it available throughout all of rust-analyzer at no cost other than depending on `proc_macro_test`, fixing https://github.com/rust-analyzer/rust-analyzer/issues/9067.
This hopefully will let us later write inline tests that utilize proc macros, which makes my life fixing proc macro bugs easier.
Opening this as a sort of RFC, because I'm not totally sure this approach is the best.
Co-authored-by: Jonas Schievink <jonasschievink@gmail.com>
9202: feat: Make `MemoryUsage` work on Windows r=jonas-schievink a=jonas-schievink
Unfortunately there is no convenient API for heap statistics, so this instead uses the Commit Charge value, which is the amount of memory that needs to be allocated either in physical RAM or in the page file. This approximation seems to be good enough to find queries that waste a large amount of memory, but it should generally be expected to be off by several MB.
bors r+
Co-authored-by: Jonas Schievink <jonasschievink@gmail.com>
9181: Don't complete values in type position r=jonas-schievink a=Veykril
Will add some proper tests in a bit
9182: fix: don't complete derive macros as fn-like macros r=jonas-schievink a=jonas-schievink
Part of https://github.com/rust-analyzer/rust-analyzer/issues/8518
bors r+
Co-authored-by: Lukas Wirth <lukastw97@gmail.com>
Co-authored-by: Jonas Schievink <jonasschievink@gmail.com>
9169: internal: steps towards attribute macro token mapping r=jonas-schievink a=jonas-schievink
This doesn't work yet, but we seem to be getting a bit further along (for example, we now stop highlighting `use` items inside item with attribute macros as if they were written verbatim).
bors r+
Co-authored-by: Jonas Schievink <jonasschievink@gmail.com>
9144: Extend convert_tuple_struct_to_named_struct to also apply to enum tuple variants r=matklad a=joshuawarner32
This is largely copied and adapted from the existing `convert_tuple_struct_to_named_struct` code.
Not sure if maybe some of this code can/should be shared between those two assists - but the differences are significant enough to make it at least seem like a non-trivial refactor.
Co-authored-by: Joshua Warner <joshuawarner32@gmail.com>
9164: internal: Reduce the number of direct fields in `CompletionContext` some more r=Veykril a=Veykril
Doesn't make the code much simpler yet.
bors r+
Co-authored-by: Lukas Wirth <lukastw97@gmail.com>
e.g. if you have a trait T and `impl T for S` for some struct, if you
goto definition on some function name inside the impl, it will go to the
definition of that function inside the `trait T` block, rather than the
current behaviour of not going anywhere at all.
9148: minor: Prevent renaming of aliases on usages r=Veykril a=Veykril
Otherwise trying to rename a usage that goes through an alias will still rename the aliased item instead, cc https://github.com/rust-analyzer/rust-analyzer/issues/5671
bors r+
Co-authored-by: Lukas Wirth <lukastw97@gmail.com>
9130: Prefix/suffix parameter inlay hint hiding heuristic is more strict r=Veykril a=Veykril
Instead of just plainly checking prefix/suffix of the argument string to the parameter name we only check for prefixes and suffixes if they are split apart via an underscore meaning, with the argument `foo`, it will be hidden for the parameter name `foo_bar` but not for `foobar`.
bors r+
Closes https://github.com/rust-analyzer/rust-analyzer/issues/8878
Co-authored-by: Lukas Wirth <lukastw97@gmail.com>
9127: internal: make variant fields inherit the enum's visibility in the ItemTree r=jonas-schievink a=jonas-schievink
No observable changes from what I can tell, but this is "more correct".
bors r+
Co-authored-by: Jonas Schievink <jonasschievink@gmail.com>
9119: fix: some minor "extract type alias" fixes r=jonas-schievink a=jonas-schievink
It now correctly works inside traits, and no longer messes up the indentation of the original node
bors r+
Co-authored-by: Jonas Schievink <jonasschievink@gmail.com>
9117: Allow expand-macro to be invoked anywhere inside a macro call r=Veykril a=Veykril
I don't really see a reason to only limit this to the name-ref of a macro.
bors r+
Closes#4606
Co-authored-by: Lukas Wirth <lukastw97@gmail.com>
9108: Don't show extract into variable assist for unit expressions r=jonas-schievink a=brandondong
**Reproduction:**
```rust
fn main() {
let mut i = 3;
$0if i >= 0 {
i += 1;
} else {
i -= 1;
}$0
}
```
1. Select the snippet of code between the $0's.
2. The extract into variable assist shows up, pushing down the more useful extract into function assist.
3. The resulting output of selecting the extract into variable assist is valid but with the extracted variable having the unit type:
```rust
fn main() {
let mut i = 3;
let var_name = if i >= 0 {
i += 1;
} else {
i -= 1;
};
var_name
}
```
**Fix:**
- Don't show the extract into variable assist for unit expressions. I could not think of any scenarios where such a variable extraction would be desired.
Co-authored-by: Brandon <brandondong604@hotmail.com>
9112: Fix some bugs in `extract_struct_from_enum_variant` assist r=Veykril a=Veykril
bors r+
Fixes#9100Fixes#9099
Kind of fixes #9109, it now copies all the generics might be incorrect if the variant doesn't use all of them)
Co-authored-by: Lukas Wirth <lukastw97@gmail.com>
9111: fix: make "extract type alias" place extracted type alias outside of impl r=jonas-schievink a=jonas-schievink
bors r+
Co-authored-by: Jonas Schievink <jonasschievink@gmail.com>
8717: Update match checking algorithm r=iDawer a=iDawer
I've recently got interest in the match checking to extend the current algo to support reporting witnesses of non-exhaustiveness.
It appears the algo is outdated from rustc's implementation. I decided to rewrite it based on the latest rustc's version. It is a diff-based port to ra codebase. That means you can diff-compare these files to rustc.
I'm striving to keep minimal ra-related changes in the algo to make it easier to backport future changes from the upstream.
Based on upstream algorithm of version rust-lang/rust 1.52.0-nightly (25c15cdbe 2021-04-22)
https://github.com/rust-lang/rust/blob/25c15cdbe/compiler/rustc_mir_build/src/thir/pattern/usefulness.rs
The goal of this PR is to cover the current `missing-match-arm` diagnostic.
What is remaining to do:
- [x] Error handling. The errors that are unrelated to match checking will be handled before the check. Just like how it made in rustc.
- [x] Lowering `hir_def::expr::Pat` to `hir_ty::diagnostics::match_check::Pat`. rustc's match checking works on top of `rustc_mir_build::thir::Pat`, which is lowered from `hir::Pat` and carries some extra semantics used by the check. All unrelated checks are done there. RA could use this to rule out running the check on unimplemented cases (`Pat::ConstBlock`, etc).
- [x] ~~Proper~~Loose typecheck of match arm patterns (https://github.com/rust-analyzer/rust-analyzer/pull/8840, https://github.com/rust-analyzer/rust-analyzer/pull/8875).
- [x] Tests from `hir_ty::diagnostics::match_check::tests`.
- [x] Clean up `todo`s
- [x] Test run on real repos https://github.com/rust-analyzer/rust-analyzer/pull/8717#issuecomment-847120265.
Co-authored-by: Dawer <7803845+iDawer@users.noreply.github.com>
fn is_useful , more skeletons
Specify a lifetime on pattern references
impl PatStack
fill impl Matrix
PatStack::pop_head_constructor
Index-based approach
struct PatCtxt
fields construction fn Fields::wildcards
split wildcard
fn Constructor::is_covered_by_any(..)
fn Matrix::specialize_constructor(..)
impl Usefulness
Initial work on witness construction
Reorganize files
Replace match checking diagnostic
Handle types of expanded patterns
unit match checking go brrr
8952: add support of impl block for doctest into runnables r=matklad a=bnjjj
close#6356
Co-authored-by: Benjamin Coenen <5719034+bnjjj@users.noreply.github.com>
8866: Update salsa r=matklad a=jonas-schievink
This updates salsa to include https://github.com/salsa-rs/salsa/pull/265, and removes all cancellation-related code from rust-analyzer
Co-authored-by: Jonas Schievink <jonasschievink@gmail.com>
9060: feat: Diagnose unimplemented built-in macros r=matklad a=jonas-schievink
A number of built-in attribute macros are unsupported, I thought it might be useful to put a diagnostic on their definition in libcore. Not sure.
Co-authored-by: Jonas Schievink <jonasschievink@gmail.com>
9051: Fix incorrect setting descriptions r=lnicola a=sclu1034
Descriptions for diagnostic warning hint and info display were swapped.
Fixes#8485.
Co-authored-by: Lucas Schwiderski <lucas@lschwiderski.de>
9027: feat: Attribute completion is context aware r=Veykril a=Veykril
This splits off the `lint` and `derive` completions into their own submodules of `attribute`.
The idea is to create a lazy global hashmap that maps `SyntaxKind` to attribute names(`&[&str]`) in which we index with the syntax kind of the "thing" we are attributing giving us the attributes back that are valid for this kind. Then we use this name to do a binary search on the attribute list to fetch and build the corresponding completion item.
Co-authored-by: Lukas Wirth <lukastw97@gmail.com>
