5637: SSR: Matching trait associated constants, types and functions r=matklad a=davidlattimore
This fixes matching of things like `HashMap::default()` by resolving `HashMap` instead of `default` (which resolves to `Default::default`).
Same for associated constants and types that are part of a trait implementation.
However, we still don't support matching calls to trait methods.
Co-authored-by: David Lattimore <dml@google.com>
5553: Add fix ranges for diagnostics r=matklad a=SomeoneToIgnore
A follow-up of https://rust-lang.zulipchat.com/#narrow/stream/185405-t-compiler.2Fwg-rls-2.2E0/topic/Less.20red.20in.20the.20code
Now diagnostics can apply fixes in a range that's different from the range used to highlight the diagnostics.
Previous logic did not consider the fix range, having both ranges equal, which could cause a lot of red noise in the editor.
Now, the fix range gets used with the fix, the diagnostics range is used for everything else which allows to improve the error highlighting.
before:
<img width="191" alt="image" src="https://user-images.githubusercontent.com/2690773/88590727-df9a6a00-d063-11ea-97ed-9809c1c5e6e6.png">
after:
<img width="222" alt="image" src="https://user-images.githubusercontent.com/2690773/88590734-e1fcc400-d063-11ea-9b7c-25701cbd5352.png">
`MissingFields` and `MissingPatFields` diagnostics now have the fix range as `ast::RecordFieldList` of the expression with an error (as it was before this PR), and the diagnostics range as a `ast::Path` of the expression, if it's present (do you have any example of `ast::Expr::RecordLit` that has no path btw?).
The rest of the diagnostics have both ranges equal, same as it was before this PR.
Co-authored-by: Kirill Bulatov <mail4score@gmail.com>
4743: Add tracking of packed repr, use it to highlight unsafe refs r=matklad a=Nashenas88
Taking a reference to a misaligned field on a packed struct is an
unsafe operation. Highlight that behavior. Currently, the misaligned
part isn't tracked, so this highlight is a bit too aggressive.
Fixes#4600
Co-authored-by: Paul Daniel Faria <Nashenas88@users.noreply.github.com>
Co-authored-by: Paul Daniel Faria <nashenas88@users.noreply.github.com>
Co-authored-by: Paul Daniel Faria <paulf@pop-os.localdomain>
5699: Fix clippy warnings r=matklad a=popzxc
Currently clippy spawns a bunch of warnings on the `rust-analyzer` project. Nothing critical, but easy to fix, so I guess it won't harm.
Co-authored-by: Igor Aleksanov <popzxc@yandex.ru>
Without this users have no clue why flycheck fails to run.
This is what is printed to the output channel:
```
[ERROR rust_analyzer::main_loop] cargo check failed: Cargo watcher failed,the command produced no valid metadata (exit code: ExitStatus(ExitStatus(25856)))
```
I stumbled with this figuring out that rust-analyzer adds `--all-features` which is not intended
for some crates in the workspace (e.g. they have mutually-exclusive features.
Having the command rust-analyzer ran should help a lot
Taking a reference to a misaligned field on a packed struct is an
unsafe operation. Highlight that behavior. Currently, the misaligned
part isn't tracked, so this highlight is a bit too aggressive.
5692: Add support for extern crate r=jonas-schievink a=Nashenas88
This adds syntax highlighting, hover and goto def functionality for extern crate.
Fixes#5690
Co-authored-by: Paul Daniel Faria <Nashenas88@users.noreply.github.com>
5693: Fix no inlay hints / unresolved tokens until manual edit to refresh r=jonas-schievink a=Veetaha
Fixes https://github.com/rust-analyzer/rust-analyzer/issues/5349
Now we return ContentModified during the workspace loading. This signifies the language
client to retry the operation (i.e. the client will
continue polling the server while it returns ContentModified).
I believe that there might be cases of overly big projects where the backoff
logic we have setup in `sendRequestWithRetry` (which we use for inlay hints)
might bail too early (currently the largest retry standby time is 10 seconds).
However, I've tried on one of my project with 500+ dependencies and it is still enough.
Here are the examples before/after the change (the gifs are quite lengthy because they show testing rather large cargo workspace).
<details>
<summary>Before</summary>
Here you can see that the client receives empty array of inlay hints and does nothing more.
Same applies to semantic tokens. The client receives unresolved tokens and does nothing more.
The user needs to do a manual edit to refresh the editor.
</details>
<details>
<summary>After</summary>
Here the server returns ContentModified, so the client periodically retries the requests and eventually receives the wellformed response.
</details>
Co-authored-by: Veetaha <veetaha2@gmail.com>
5414: Fix test code lens r=jonas-schievink a=avrong
Closes#5217
The implementation is quite similar to #4821. Maybe we should somehow deal with duplicated code.
Also, both of these requests introduce some unclear behavior. I'm not sure how to process this, therefore asking for advice. Examples are below.
<img width="286" alt="image" src="https://user-images.githubusercontent.com/6342851/87713209-83595f80-c7b2-11ea-8c0f-a12e7571e7df.png">
Co-authored-by: Aleksei Trifonov <avrong@avrong.me>
No we return ContentModified during the workspace loading. This signifies the language
client to retry the operation (i.e. the client will
continue polling the server while it returns ContentModified).
I believe that there might be cases of overly big projects where the backoff
logic we have setup in `sendRequestWithRetry` (which we use for inlay hints)
might bail too early (currently the largest retry standby time is 10 seconds).
However, I've tried on one of my project with 500+ dependencies and it is still enough.
5684: Semantic highlighting for unsafe union field access r=jonas-schievink a=Nashenas88
This change adds support for unions in inference and lowering, then extends on that to add the unsafe semantic modifier on field access only. The `is_possibly_unsafe` function in `syntax_highlighting.rs` could be extended to support fns and static muts so that their definitions are not highlighted as unsafe, but only their usage.
Also, each commit of this PR updates the tests. By reviewing the files by commit, it's easy to see how the changes in the code affected the tests.
Co-authored-by: Paul Daniel Faria <Nashenas88@users.noreply.github.com>
5679: Account for static mut in missing unsafe diagnostic r=jonas-schievink a=Nashenas88
Accessing or modifying a static mut is an unsafe operation. The "missing unsafe" diagnostic now tracks this.
Co-authored-by: Paul Daniel Faria <Nashenas88@users.noreply.github.com>
5678: Static mut unsafe semantic highlighting r=jonas-schievink a=Nashenas88
This marks static mutable names as unsafe, since accessing or modifying a static mut is an unsafe operation.
Co-authored-by: Paul Daniel Faria <Nashenas88@users.noreply.github.com>
5526: Handle semantic token deltas r=kjeremy a=kjeremy
This basically takes the naive approach where we always compute the tokens but save space sending over the wire which apparently solves some GC problems with vscode.
This is waiting for https://github.com/gluon-lang/lsp-types/pull/174 to be merged. I am also unsure of the best way to stash the tokens into `DocumentData` in a safe manner.
Co-authored-by: kjeremy <kjeremy@gmail.com>
Co-authored-by: Jeremy Kolb <kjeremy@gmail.com>
5639: SSR: Allow `self` in patterns. r=jonas-schievink a=davidlattimore
It's now consistent with other variables in that if the pattern references self, only the `self` in scope where the rule is invoked will be accepted. Since `self` doesn't work the same as other paths, this is implemented by restricting the search to just the current function. Prior to this change (since path resolution was implemented), having self in a pattern would just result in no matches.
Co-authored-by: David Lattimore <dml@google.com>
5664: Fix renamed self module. r=jonas-schievink a=Nashenas88
Fixes#5663
Now `inner_mod` below is properly marked as a `module`.
```rust
use crate::inner::{self as inner_mod};
mod inner {}
```
Co-authored-by: Paul Daniel Faria <Nashenas88@users.noreply.github.com>
`current_dir` and relative paths to executables works differently on
unix and windows (unix behavior does not make sense), see:
17e30e83a1/src/lib.rs (L295-L324)
The original motivation to set cwd was to make rustfmt read the
correct rustfmt.toml, but that was future proofing, rather than a bug
fix.
So, let's just remove this and see if breaks or fixes more use-cases.
If support for per-file config is needed, we could use `--config-path`
flag.
5658: do not add to `pub use` in assists that insert a use statement r=jonas-schievink a=jbr
closes#5657 , see issue for rationale
Initially I wrote a version of this that changed the signature of `insert_use_statement` to take an `Option<VisibilityKind>` and only add to use statements with the same visibility, but that didn't make sense for any of the current uses of `insert_use_statement` (they all expected private visibility).
Co-authored-by: Jacob Rothstein <hi@jbr.me>
It's now consistent with other variables in that if the pattern
references self, only the `self` in scope where the rule is invoked will
be accepted. Since `self` doesn't work the same as other paths, this is
implemented by restricting the search to just the current function.
Prior to this change (since path resolution was implemented), having
self in a pattern would just result in no matches.
This fixes matching of things like `HashMap::default()` by resolving
`HashMap` instead of `default` (which resolves to `Default::default`).
Same for associated constants and types that are part of a trait
implementation.
However, we still don't support matching calls to trait methods.
Note that `for` type is rust-analyzer's own invention.
Both the reference and syn allow `for` only for fnptr types, and we
allow them everywhere. This needs to be checked with respect to type
bounds grammar...
The TypeRef name comes from IntelliJ days, where you often have both
type *syntax* as well as *semantical* representation of types in
scope. And naming both Type is confusing.
In rust-analyzer however, we use ast types as `ast::Type`, and have
many more semantic counterparts to ast types, so avoiding name clash
here is just confusing.
5596: Add checkOnSave.noDefaultFeatures and correct, how we handle some cargo flags. r=clemenswasser a=clemenswasser
This PR adds the `rust-analyzer.checkOnSave.noDefaultFeatures` option
and fixes the handling of `cargo.allFeatures`, `cargo.noDefaultFeatures` and `cargo.features`.
Fixes: #5550
Co-authored-by: Clemens Wasser <clemens.wasser@gmail.com>
This commit fixes the handling of user-defined configuration
of some cargo options. Previously you could either specify
`--all-features`, `--no-default-features` or `--features`.
Now you can specify either `--all-features` or `--no-default-features`
and `--features`. This commit also corrects the `--features`
command-line argument creation inside of `load_extern_resources`.
5567: SSR: Wrap placeholder expansions in parenthesis when necessary r=matklad a=davidlattimore
e.g. `foo($a) ==> $a.to_string()` should produce `(1 + 2).to_string()` not `1 + 2.to_string()`
We don't yet try to determine if the whole replacement needs to be wrapped in parenthesis. That's harder and I think perhaps less often an issue.
Co-authored-by: David Lattimore <dml@google.com>
e.g. `foo($a) ==> $a.to_string()` should produce `(1 + 2).to_string()`
not `1 + 2.to_string()`
We don't yet try to determine if the whole replacement needs to be
wrapped in parenthesis. That's harder and I think perhaps less often an
issue.
5554: Fix remove_dbg r=matklad a=petr-tik
Closes#5129
Addresses two issues:
- keep the parens from dbg!() in case the call is chained or there is
semantic difference if parens are excluded
- Exclude the semicolon after the dbg!(); by checking if it was
accidentally included in the macro_call
investigated, but decided against:
fix ast::MacroCall extraction to never include semicolons at the end -
this logic lives in rowan.
Defensively shorten the macro_range if there is a semicolon token.
Deleted unneccessary temp variable macro_args
Renamed macro_content to "paste_instead_of_dbg", because it isn't a
simple extraction of text inside dbg!() anymore
Co-authored-by: petr-tik <petr-tik@users.noreply.github.com>
5563: Check all targets for package-level tasks r=matklad a=SomeoneToIgnore
When invoking "Select Runnable" with the caret on a runnable with a specific target (test, bench, binary), append the corresponding argument for the `cargo check -p` module runnable.
Co-authored-by: Kirill Bulatov <mail4score@gmail.com>
replaced match with let-if variable assignment
removed the unnecessary semicolon_on_end variable
converted all code and expected test variables to raw strings
and inlined them in asserts
The primary advantage of ungrammar is that it (eventually) allows one
to describe concrete syntax tree structure -- with alternatives and
specific sequence of tokens & nodes.
That should be re-usable for:
* generate `make` calls
* Rust reference
* Hypothetical parser's evented API
We loose doc comments for the time being unfortunately. I don't think
we should add support for doc comments to ungrammar -- they'll make
grammar file hard to read. We might supply docs as out-of band info,
or maybe just via a reference, but we'll think about that once things
are no longer in flux
5564: SSR: Restrict to current selection if any r=davidlattimore a=davidlattimore
The selection is also used to avoid unnecessary work, but only to the file level. Further restricting unnecessary work is left for later.
Co-authored-by: David Lattimore <dml@google.com>
5565: SSR: Don't mix non-path-based rules with path-based r=matklad a=davidlattimore
If any rules contain paths, then we reject any rules that don't contain paths. Allowing a mix leads to strange semantics, since the path-based rules only match things where the path refers to semantically the same thing, whereas the non-path-based rules could match anything. Specifically, if we have a rule like `foo ==>> bar` we only want to match the `foo` that is in the current scope, not any `foo`. However "foo" can be parsed as a pattern (BIND_PAT -> NAME -> IDENT). Allowing such a rule through would result in renaming everything called `foo` to `bar`. It'd also be slow, since without a path, we'd have to use the slow-scan search mechanism.
Co-authored-by: David Lattimore <dml@google.com>
If any rules contain paths, then we reject any rules that don't contain paths. Allowing a mix leads to strange semantics, since the path-based rules only match things where the path refers to semantically the same thing, whereas the non-path-based rules could match anything. Specifically, if we have a rule like `foo ==>> bar` we only want to match the `foo` that is in the current scope, not any `foo`. However "foo" can be parsed as a pattern (BIND_PAT -> NAME -> IDENT). Allowing such a rule through would result in renaming everything called `foo` to `bar`. It'd also be slow, since without a path, we'd have to use the slow-scan search mechanism.
Addresses two issues:
- keep the parens from dbg!() in case the call is chained or there is
semantic difference if parens are excluded
- Exclude the semicolon after the dbg!(); by checking if it was
accidentally included in the macro_call
investigated, but decided against:
fix ast::MacroCall extraction to never include semicolons at the end -
this logic lives in rowan.
Defensively shorten the macro_range if there is a semicolon token.
Deleted unneccessary temp variable macro_args
Renamed macro_content to "paste_instead_of_dbg", because it isn't a
simple extraction of text inside dbg!() anymore
It seems that Semantics::scope, if given a statement node, won't resolve
locals that were defined in the current scope, only in parent scopes.
Not sure if this is intended / expected behavior, but we work around it
for now by finding another nearby node to use as the scope (e.g. the
expression inside the EXPR_STMT).
5520: Add DocumentData to represent in-memory document with LSP info r=matklad a=kjeremy
At the moment this only holds document version information but in the near-future it will hold other things like semantic token delta info.
Co-authored-by: kjeremy <kjeremy@gmail.com>
5522: Increace tracing-tree version from 0.1.3 to 0.1.4 r=kjeremy a=vandenheuvel
Co-authored-by: Bram van den Heuvel <b.vandenheuvel@student.tudelft.nl>
5516: Better LSP conformance r=matklad a=vsrs
At the moment rust-analyzer does not fully conform to the LSP. This PR fixes two LSP related issues:
1) rust-analyzer sends predefined server capabilities and does not take supplied client capabilities in mind.
2) rust-analyzer uses dynamic `textDocument/didSave` registration even if the client does not support it.
Co-authored-by: vsrs <vit@conrlab.com>
5518: Use resolved paths in SSR rules r=matklad a=davidlattimore
The main user-visible changes are:
* SSR now matches paths based on whether they resolve to the same thing instead of whether they're written the same.
* So `foo()` won't match `foo()` if it's a different function `foo()`, but will match `bar::foo()` if it's the same `foo`.
* Paths in the replacement will now be rendered with appropriate qualification for their context.
* For example `foo::Bar` will render as just `Bar` inside the module `foo`, but might render as `baz::foo::Bar` from elsewhere.
* This means that all paths in the search pattern and replacement template must be able to be resolved.
* It now also matters where you invoke SSR from, since paths are resolved relative to wherever that is.
* Search now uses find-uses on paths to locate places to try matching. This means that when a path is present in the pattern, search will generally be pretty fast.
* Function calls can now match method calls again, but this time only if they resolve to the same function.
Co-authored-by: David Lattimore <dml@google.com>
This differs from how this used to work before I removed it in that:
a) It's only one direction. Function calls in the pattern can match
method calls in the code, but not the other way around.
b) We now check that the function call in the pattern resolves to the
same function as the method call in the code.
The lack of (b) was the reason I felt the need to remove the feature
before.
Previously, submatches were handled simply by searching in placeholders
for more matches. That only works if we search all nodes in the tree
recursively. In a subsequent commit, I intend to make search not always
be recursive recursive. This commit prepares for that by finding all
matches, even if they overlap, then nesting them and removing
overlapping matches.
In a later commit, paths in templates will be resolved. This allows us
to render the path with appropriate qualifiers for its context. Here we
prepare for that change by updating existing tests where I'd previously
not bothered to define the items that the template referred to.
The methods `edits_for_file` and `find_matches_in_file` are replaced with just `edits` and `matches`. This simplifies the API a bit, but more importantly it makes it possible in a subsequent commit for SSR to decide to not search all files.
Also renamed find_matches to slow_scan_node to reflect that it's a slow
way to do things. Actually the name came from a later commit and
probably makes more sense once there's an alternative.
This is in preparation for a subsequent commit where we add special
handling for paths in the template, allowing them to be qualified
differently in different contexts.
Previously we had:
- Multiple rules
- Each rule had its pattern parsed as an expression, path etc
This meant that there were two levels at which there could be multiple
rules.
Now we just have multiple rules. If a pattern can parse as more than one
kind of thing, then they get stored as multiple separate rules.
We also now don't have separate fields for the different kinds of things
that a pattern can parse as. This makes adding new kinds of things
simpler.
Previously, add_search_pattern would construct a rule with a dummy
replacement. Now the replacement is an Option. This is slightly cleaner
and also opens the way for parsing the replacement template as the same
kind of thing as the search pattern.
5498: assists: change_return_type_to_result: clarify assist description r=matklad a=matthiaskrgr
I had a -> Option<PathBuf> fn, which I wanted to change to Result<PathBuf, _>, but despite advertising to do so, the assist did not change the result type to Result<PathBuf, _> but instead just wrapped it in a Result: <Result<Option<PathBuf>, _>.
I changed the assist description to "Wrap return type in Result" to clarify that the assist only wraps the preexisting type and does not do any actual Option-to-Result refactoring.
Co-authored-by: Matthias Krüger <matthias.krueger@famsik.de>
5497: Store macro invocation parameters as text instead of tt r=jonas-schievink a=lnicola
We don't want to expand macros on every source change because it can be arbitrarily slow, but the token trees can be rather large. So instead we can cache the invocation parameters (as text).
Co-authored-by: Laurențiu Nicola <lnicola@dend.ro>
I had a -> Option<PathBuf> fn, which I wanted to change to Result<PathBuf, _>, but despite advertising to do so, the assist did not
change the result type to Result<PathBuf, _> but instead just wrapped it in a Result: <Result<Option<PathBuf>, _>.
I changed the assist description to "Wrap return type in Result" to clarify that the assist only wraps the preexisting type and does
not do any deep Option-to-Result refactoring.
5481: Track document versions in the server r=kjeremy a=kjeremy
This also pushes diagnostics for the correct file version on close so that when it is reopened stale diagnostics are not shown.
Closes#5452
Co-authored-by: kjeremy <kjeremy@gmail.com>
Co-authored-by: Jeremy Kolb <kjeremy@gmail.com>
5478: Replace existing visibility modifier in fix_visibility r=matklad a=TimoFreiberg
Fixes#4636
I would have liked to do something about the `// FIXME: this really should be a fix for diagnostic, rather than an assist.`, but that would take a while and there's no reason not to fix this immediately.
Co-authored-by: Timo Freiberg <timo.freiberg@gmail.com>
5451: Highlight more cases of SyntaxKind when it is a punctuation r=matklad a=GrayJack
This maybe closes#5406
Closes #5453
Separate what one expect to be a punctuation semantic token (like `,`, `;`, `(`, etc), and what is not (`&`, `::`, `+`, etc)
5463: Bump lexer r=matklad a=kjeremy
Since we're now on rust 1.45
5465: Bump chalk r=matklad a=kjeremy
5466: Do not show default types in function and closure return values r=matklad a=SomeoneToIgnore
Avoid things like
<img width="522" alt="image" src="https://user-images.githubusercontent.com/2690773/87985936-1bbe4f80-cae5-11ea-9b8a-5383d896c296.png">
Co-authored-by: GrayJack <gr41.j4ck@gmail.com>
Co-authored-by: kjeremy <kjeremy@gmail.com>
Co-authored-by: Kirill Bulatov <mail4score@gmail.com>
5467: Allow null or empty values for configuration r=matklad a=kjeremy
Allow the client to respond to `workspace/configuration` with `null` values. This is allowed per the spec if the client doesn't know about the configuration we've requested.
This also protects against `null` or `{}` during initialize. I'm not sure if we want to interpret `{}` as "don't change anything" but I think that's a reasonable approach to take.
This should help with LSP clients working out of the box.
Fixes#5464
Co-authored-by: kjeremy <kjeremy@gmail.com>
It is currently unused, but, in the future, it will be used to:
* drive certain UX (symbols search by default will look only in the
members)
* improve performance (rust-analyzer will assume that non-members
change rarely)
If not specified, is_workspace member is inferred from the path
5423: Correctly resolve assoc. types in path bindings r=matklad a=jonas-schievink
Previously invoking goto def on `impl Iterator<Item<|> = ()>` would go to `Iterator`, not `Item`. This fixes that.
Co-authored-by: Jonas Schievink <jonasschievink@gmail.com>
