Always show error lifetime arguments as `'_`
Fixes#17947
Changed error lifetime argument presentation in non-test environment to `'_` and now showing them even if all of args are error lifetimes.
This also influenced some of the other tests like `extract_function.rs`, `predicate.rs` and `type_pos.rs`. Not sure whether I need to refrain from adding lifetimes args there. Happy to fix if needed
The search is used by IDE features such as rename and find all references.
The search is slow because we need to verify each candidate, and that requires analyzing it; the key to speeding it up is to avoid the analysis where possible.
I did that with a bunch of tricks that exploits knowledge about the language and its possibilities. The first key insight is that associated methods may only be referenced in the form `ContainerName::func_name` (parentheses are not necessary!) (Rust doesn't include a way to `use Container::func_name`, and even if it will in the future most usages are likely to stay in that form.
Searching for `::` will help only a bit, but searching for `Container` can help considerably, since it is very rare that there will be two identical instances of both a container and a method of it.
However, things are not as simple as they sound. In Rust a container can be aliased in multiple ways, and even aliased from different files/modules. If we will try to resolve the alias, we will lose any gain from the textual search (although very common method names such as `new` will still benefit, most will suffer because there are more instances of a container name than its associated item).
This is where the key trick enters the picture. The key insight is that there is still a textual property: a container namer cannot be aliased, unless its name is mentioned in the alias declaration, or a name of alias of it is mentioned in the alias declaration.
This becomes a fixpoint algorithm: we expand our list of aliases as we collect more and more (possible) aliases, until we eventually reach a fixpoint. A fixpoint is not guaranteed (and we do have guards for the rare cases where it does not happen), but it is almost so: most types have very few aliases, if at all.
We do use some semantic information while analyzing aliases. It's a balance: too much semantic analysis, and the search will become slow. But too few of it, and we will bring many incorrect aliases to our list, and risk it expands and expands and never reach a fixpoint. At the end, based on benchmarks, it seems worth to do a lot to avoid adding an alias (but not too much), while it is worth to do a lot to avoid the need to semantically analyze func_name matches (but again, not too much).
After we collected our list of aliases, we filter matches based on this list. Only if a match can be real, we do semantic analysis for it.
The results are promising: searching for all references on `new()` in `base-db` in the rust-analyzer repository, which previously took around 60 seconds, now takes as least as two seconds and a half (roughly), while searching for `Vec::new()`, almost an upper bound to how much a symbol can be used, that used to take 7-9 minutes(!) now completes in 100-120 seconds, and with less than half of non-verified results (aka. false positives).
This is the less strictly correct (but faster) of this patch; it can miss some (rare) cases (there is a test for that - `goto_ref_on_short_associated_function_complicated_type_magic_can_confuse_our_logic()`). There is another branch that have no false negatives but is slower to search (`Vec::new()` never reaches a fixpoint in aliases collection there). I believe it is possible to create a strategy that will have the best of both worlds, but it will involve significant complexity and I didn't bother, especially considering that in the vast majority of the searches the other branch will be more than enough. But all in all, I decided to bring this branch (of course if the maintainers will agree), since our search is already not 100% accurate (it misses macros), and I believe there is value in the additional perf.
Add scip/lsif flag to exclude vendored libaries
#17809 changed StaticIndex to include vendored libraries. This PR adds a flag to disable that behavior.
At work, our monorepo has too many rust targets to index all at once, so we split them up into several shards. Since all of our libraries are vendored, if rust-analyzer includes them, sharding no longer has much benefit, because every shard will have to index the entire transitive dependency graphs of all of its targets. We get around the issue presented in #17809 because some other shard will index the libraries directly.
This PR touches a lot of parts. But the main changes are changing
`hir_expand::Name` to be raw edition-dependently and only when necessary
(unrelated to how the user originally wrote the identifier),
and changing `is_keyword()` and `is_raw_identifier()` to be edition-aware
(this was done in #17896, but the FIXMEs were fixed here).
It is possible that I missed some cases, but most IDE parts should properly
escape (or not escape) identifiers now.
The rules of thumb are:
- If we show the identifier to the user, its rawness should be determined
by the edition of the edited crate. This is nice for IDE features,
but really important for changes we insert to the source code.
- For tests, I chose `Edition::CURRENT` (so we only have to (maybe) update
tests when an edition becomes stable, to avoid churn).
- For debugging tools (helper methods and logs), I used `Edition::LATEST`.
Include vendored crates in StaticIndex
`StaticIndex::compute` filters out modules from libraries. This makes an exceptions for vendored libraries, ie libraries actually defined inside the workspace being indexed.
This aims to solve https://bugzilla.mozilla.org/show_bug.cgi?id=1846041 In general StaticIndex is meant for code browsers, which likely want to index all visible source files.
fix: tyck for non-ADT types when searching refs for `Self` kw
See e0276dc5dd (r1389848845)
For ADTs, to handle `{error}` in generic args, we should to convert them to ADT for comparisons; for others, we can directly compare the types.
StaticIndex::compute filters out modules from libraries. This makes an
exceptions for vendored libraries, ie libraries actually defined inside
the workspace being indexed.
This aims to solve https://bugzilla.mozilla.org/show_bug.cgi?id=1846041
In general StaticIndex is meant for code browsers, which likely want to
index all visible source files.
feat: Use spans for builtin and declarative macro expansion errors
This should generally improve some error reporting for macro expansion errors. Especially for `compile_error!` within proc-macros
The issue occurs because in some configurations of traits where one of them has Deref as a supertrait, RA's type inference algorithm fails to resolve the Deref::Target type, and instead uses a TyKind::BoundVar (i.e. an unknown type). This "autoderefed" type then incorrectly acts as if it implements all traits in scope.
The fix is to re-apply the same sanity-check that is done in iterate_method_candidates_with_autoref(), that is: don't try to resolve methods on unknown types. This same sanity-check is now done on each autoderefed type for which trait methods are about to be checked. If the autoderefed type is unknown, then the iterating of the trait methods for that type is skipped.
Includes a unit test that only passes after applying the fixes in this commit.
Includes a change to the assertion count in test syntax_highlighting::tests::benchmark_syntax_highlighting_parser as suggested by Lukas Wirth during review.
Includes a change to the sanity-check code as suggested by Florian Diebold during review.
feat: go-to-def and find-references on control-flow keywords
fix#17517.
This PR implements **go-to-definition** and **find-references** functionalities for control flow keywords, which is similar to the behaviors in the `highlight-related` module. Besides, this PR also fixes some incorrect behaviors in `highlight-related`.
## Changes
1. **Support for go-to-definition on control flow keywords**:
This PR introduces functionality allowing users to navigate on the definition of control flow keywords (`return`, `break`, `continue`).
Commit: 2a3244ee147f898dd828c06352645ae1713c260f..7391e7a608634709db002a4cb09229de4d12c056.
2. **Bug fixes and refactoring in highlight-related**:
- **Handling return/break/continue within try_blocks**:
This PR adjusted the behavior of these keywords when they occur within `try_blocks`. When encounter these keywords, the program should exit the outer function or loop which containing the `try_blocks`, rather than the `try_blocks` itself; while the `?` will cause the program to exit `try_blocks`.
Commit: 59d697e807f0197f59814b37dca1563959da4aa1.
- **Support highlighting keywords in macro expansion for highlight-related**:
Commit: 88df24f01727c23a667a763ee3ee0cec22d5ad52.
- Detailed description for the bug fixes
+ The previous implementation of `preorder_expr` incorrectly treated `try_blocks` as new contexts, thereby r-a will not continue to traverse inner `return` and `break/continue` statements. To resolve this, a new function `preorder_expr_with_ctx_checker` has been added, allowing users to specify which expressions to skip.
* For example, when searching for the `?` in the context, r-a should skip `try_blocks` where the `?` insides just works for `try_blocks`. But when search for the `return` keyword, r-a should collect both the `return` keywords inside and outside the `try_blocks`
+ Thus, this PR added `WalkExpandedExprCtx` (builder pattern). It offers the following improvements: customizable context skipping, maintenance of loop depth (for `break`/`continue`), and handling macro expansion during traversal.
3. **Support for find-references on control flow keywords**:
This PR enables users to find all references to control flow keywords.
Commit: 9202a33f81218fb9c2edb5d42e6b4de85b0323a8.
Trigger VSCode to rename after extract variable assist is applied
When the user applies the "Extract Variable" assist, the cursor is
positioned at the newly inserted variable. This commit adds a command
to the assist that triggers the rename action in VSCode. This way, the
user can quickly rename the variable after applying the assist.
Fixes part of: #17579https://github.com/user-attachments/assets/4cf38740-ab22-4b94-b0f1-eddd51c26c29
I haven't yet looked at the module or function extraction assists yet.
When the user applies the "Extract Variable" assist, the cursor is
positioned at the newly inserted variable. This commit adds a command
to the assist that triggers the rename action in VSCode. This way, the
user can quickly rename the variable after applying the assist.
Fixes part of: #17579
Add `f16` and `f128` support
Adds `f16` and `f128` support, using the `rustc_apfloat` library (also used by `rustc`) for parsing/arithmetic/displaying since the types aren't stable yet so can't be used by rust-analyzer itself.
Issue: #17451
feat: add inlay hints for generic parameters
fixes#11091
By default, only hints for const generic parameters are shown, and this can be configured through `rust-analyzer.inlayHints.genericParameterHints.enable`.
Probably needs more testing.
Filter builtin macro expansion
This PR adds a filter on the types of built in macros that are allowed to be expanded.
Currently, This list of allowed macros contains, `stringify, cfg, core_panic, std_panic, concat, concat_bytes, include, include_str, include_bytes, env` and `option_env`.
Fixes#14177
fix: handle character boundaries for wide chars in extend_selection
fix#17420.
When calling 'extend_selection' within a string, r-a attempts to locate the current word at the cursor. This is done by finding the first char before the cursor which is not a letter, digit, or underscore.
The position of this character is referred to as `start_idx`, and the word is considered to start from `start_idx + 1`. However, for wide characters, `start_idx + 1` is not character boundaries, which leading to panic. We should use `ceil_char_boundary` to ensure that the idx is always on character boundaries.
feat: TOML based config for rust-analyzer
> Important
>
> We don't promise _**any**_ stability with this feature yet, any configs exposed may be removed again, the grouping may change etc.
# TOML Based Config for RA
This PR ( addresses #13529 and this is a follow-up PR on #16639 ) makes rust-analyzer configurable by configuration files called `rust-analyzer.toml`. Files **must** be named `rust-analyzer.toml`. There is not a strict rule regarding where the files should be placed, but it is recommended to put them near a file that triggers server to start (i.e., `Cargo.{toml,lock}`, `rust-project.json`).
## Configuration Types
Previous configuration keys are now split into three different classes.
1. Client keys: These keys only make sense when set by the client (e.g., by setting them in `settings.json` in VSCode). They are but a small portion of this list. One such example is `rust_analyzer.files_watcher`, based on which either the client or the server will be responsible for watching for changes made to project files.
2. Global keys: These keys apply to the entire workspace and can only be set on the very top layers of the hierarchy. The next section gives instructions on which layers these are.
3. Local keys: Keys that can be changed for each crate if desired.
### How Am I Supposed To Know If A Config Is Gl/Loc/Cl ?
#17101
## Configuration Hierarchy
There are 5 levels in the configuration hierarchy. When a key is searched for, it is searched in a bottom-up depth-first fashion.
### Default Configuration
**Scope**: Global, Local, and Client
This is a hard-coded set of configurations. When a configuration key could not be found, then its default value applies.
### User configuration
**Scope**: Global, Local
If you want your configurations to apply to **every** project you have, you can do so by setting them in your `$CONFIG_DIR/rust-analyzer/rust-analyzer.toml` file, where `$CONFIG_DIR` is :
| Platform | Value | Example |
| ------- | ------------------------------------- | ---------------------------------------- |
| Linux | `$XDG_CONFIG_HOME` or `$HOME`/.config | /home/alice/.config |
| macOS | `$HOME`/Library/Application Support | /Users/Alice/Library/Application Support |
| Windows | `{FOLDERID_RoamingAppData}` | C:\Users\Alice\AppData\Roaming |
### Client configuration
**Scope**: Global, Local, and Client
Previously, the only way to configure rust-analyzer was to configure it from the settings of the Client you are using. This level corresponds to that.
> With this PR, you don't need to port anything to benefit from new features. You can continue to use your old settings as they are.
### Workspace Root Configuration
**Scope**: Global, Local
Rust-analyzer already used the path of the workspace you opened in your Client. We used this information to create a configuration file that won't affect your other projects and define global level configurations at the same time.
### Local Configuration
**Scope**: Local
You can also configure rust-analyzer on a crate level. Although it is not an error to define global ( or client ) level keys in such files, they won't be taken into consideration by the server. Defined local keys will affect the crate in which they are defined and crate's descendants. Internally, a Rust project is split into what we call `SourceRoot`s. This, although with exceptions, is equal to splitting a project into crates.
> You may choose to have more than one `rust-analyzer.toml` files within a `SourceRoot`, but among them, the one closer to the project root will be
fix: do not resolve prelude within block modules
fix#17338 (continuing from #17251).
In #17251, we injected preludes into non-top-level modules, which leading to r-a to directly resolve names in preludes in block modules. This PR fix it by checking whether the module is a pseudo-module introduced by blocks. (similar to what we do for extern preludes)
Feat: hide double underscored symbols from symbol search
Fixes#17272 by changing the default behavior of query to skip results that start with `__` (two underscores).
Not sure if this has any far reaching implications - a review would help to understand if this is the right place to do the filtering, and if it's fine to do it by default on the query.
If you type `__` as your search, then we'll show the matching double unders, just in case you actually need the symbol.
Fix: infer type of async block with tail return expr
Fixes#17106
The `infer_async_block` method calls the `infer_block` method internally, which returns the never type without coercion when `tail_expr` is `None` and `ctx.diverges` is `Diverges::Always`.This is the reason for the bug in this issue.
cfce2bb46d/crates/hir-ty/src/infer/expr.rs (L1411-L1413)
This PR solves the bug by adding a process to coerce after calling `infer_block` method.
This code passes all the tests, including tests I added for this isuue, however, I am not sure if this solution is right. I think that this solution is an ad hoc solution. So, I would appreciate to have your review.
I apologize if I'm off the mark, but `infer_async_block` method should be rewritten to share code with the process of infering type of `expr::Closure` instead of the `infer_block` method. That way it will be closer to the infer process of rustc.
When viewing traces, it's slightly confusing when the span name doesn't
match the function name. Ensure the names are consistent.
(It might be worth moving most of these to use #[tracing::instrument]
so the name can never go stale. @davidbarsky suggested that is marginally
slower, so I've just done the simple change here.)
Support hovering limits for adts
Fix#17009
1. Currently, r-a supports limiting the number of struct fields displayed when hovering. This PR extends it to support enum variants and union fields. Since the display of these three (ADTs) is similar, this PR extends 'hover_show_structFields' to 'hover_show_adtFieldsOrVariants'.
2. This PR also resolved the problem that the layout of ADT was not restricted by display limitations when hovering on the Self type.
3. Additionally, this PR changes the default value of display limitations to `10` (instead of the original `null`), which helps users discover this feature.
It is bitset semantically --- many categorical things can be true about
a reference at the same time.
In parciular, a reference can be a "test" and a "write" at the same
time.
