1238: Macro queries r=edwin0cheng a=matklad
In https://github.com/rust-analyzer/rust-analyzer/pull/1231, I've added aggressive clean up of `ast_id_to_node` query.
The result of this query is a `SyntaxTree`, and we don't want to retain syntax trees in memory unless absolutely necessary.
Moreover, `SyntaxTree` has identity equality semantics, meaning that we'll get a diffferent syntax tree for a file after every reparse. That means that `ast_id_to_node` query should not genereally be used in HIR, unless it is behind some kind of salsa firewall, like the `raw` module of name resoulution.
However, that PR resulted in the abysmal performance: turns out we were using ast_id_to_node quite heavily in hir when expanding macros!
So this PR installs the more incremental-friendly query structure:
* converting source to token tree is now a query; changing source without affecting token-trees will now preserve macro expansions
* expand macro (tt -> tt) is now a query as well, so we cache macro expansions *before* parsing them into item lists or expressions, which is nice: we can cache expansion without knowing the calling context!
r? @edwin0cheng
Co-authored-by: Aleksey Kladov <aleksey.kladov@gmail.com>
Currently, when expanding macros, we look at the source code
directly (we invoke ast_id_to_node query via to_node method).
This is less then ideal, because it make us re-expand macros after
every source change.
This commit establishes a salsa-firewall: a query to get macro call's
token tree. Unlike the syntax tree, token tree changes only if we
actually modify the macro itself.
1230: Desugar doc comments to `#[doc = "...."]` attributes in `syntax node` to tt conversion r=matklad a=edwin0cheng
As discussed in [Zulip](https://rust-lang.zulipchat.com/#narrow/stream/185405-t-compiler.2Fwg-rls-2.2E0/topic/MBE.20discussion/near/164446835), this PR desugar doc comments to `#[doc = "...."]` in `syntax node` to tt conversion.
Note that after this PR, all obvious mbe bugs in dogfooding are fixed. (i.e. No parsing or expanding mbe error in `env RUST_LOG=ra_hir=WARN target\release\ra_cli.exe analysis-stats`) 🎉
Co-authored-by: Edwin Cheng <edwin0cheng@gmail.com>
1227: Add `default_type` method in `TypeParam` Node r=matklad a=edwin0cheng
This PR add a `default_type` method in `TypeParam` Node which allow future PR to handle #1099 case.
Co-authored-by: Edwin Cheng <edwin0cheng@gmail.com>
1222: Skip Dollars when bump raw token r=matklad a=edwin0cheng
This PR fixed a bug while parsing token_tree, it should skip all L_DOLLAR AND R_DOLLAR.
Co-authored-by: Edwin Cheng <edwin0cheng@gmail.com>
1220: Add macro pat parsing r=matklad a=edwin0cheng
This PR add support to parsing macro call in pattern , e.g :
```
let m!(x) = 0;
```
Co-authored-by: Edwin Cheng <edwin0cheng@gmail.com>
1213: Make lexer produce only single character puncts r=matklad a=edwin0cheng
As discussed in Zulip, this PR change `lexer` to produce only single char punct.
* Remove producing `DOTDOTDOT, DOTDOTEQ, DOTDOT, COLONCOLON, EQEQ, FAT_ARROW, NEQ, THIN_ARROW` in lexer.
* Add required code in parser to make sure everythings works fine.
* Change some tests (Mainly because the `ast::token_tree` is different)
Note: i think the use of `COLON` in rust is too overloaded :)
Co-authored-by: Edwin Cheng <edwin0cheng@gmail.com>
E.g. in
```
let foo = 1u32;
if true {
<|>foo;
}
```
the hover shows `()`, the type of the whole if expression, instead of the more
sensible `u32`. The reason for this was that the search for an expression was
slightly left-biased: When on the edge between two tokens, it first looked at
all ancestors of the left token and then of the right token. Instead merge the
ancestors in ascending order, so that we get the smaller of the two possible
expressions.
1200: Allows searching for case-equivalent symbols (fixes#1151) r=matklad a=jrvidal
I couldn't find a nice, functional way of calculating the ranges in one pass so I resorted to a plain old `for` loop.
Co-authored-by: Roberto Vidal <vidal.roberto.j@gmail.com>
1147: Handle macros in type checking / HIR r=matklad a=Lapz
An other attempt at #1102. I will need to flatten the nested if statements and im also not sure if the way that i get the resolver and module will always work
Co-authored-by: Lenard Pratt <l3np27@gmail.com>
We really shouldn't be looking at the identifier at point. Instead,
all filtering and sorting should be implemented at the layer above.
This layer should probably be home for auto-import completions as
well, but, since that is not yet implemented, let's just stick this
into complete_scope.
1193: Add a test for #1178 case r=edwin0cheng a=edwin0cheng
A little PR to add a test case for #1178
Co-authored-by: Edwin Cheng <edwin0cheng@gmail.com>
1184: Start structured editing API r=matklad a=matklad
I think I finally understand how to provide nice, mutable structured editing API on top of red-green trees.
The problem I am trying to solve is that any modification to a particular `SyntaxNode` returns an independent new file. So, if you are editing a struct literal, and add a field, you get back a SourceFile, and you have to find the struct literal inside it yourself! This happens because our trees are immutable, but have parent pointers.
The main idea here is to introduce `AstEditor<T>` type, which abstracts away that API. So, you create an `AstEditor` for node you want to edit and call various `&mut` taking methods on it. Internally, `AstEditor` stores both the original node and the current node. All edits are applied to the current node, which is replaced by the corresponding node in the new file. In the end, `AstEditor` computes a text edit between old and new nodes.
Note that this also should sole a problem when you create an anchor pointing to a subnode and mutate the parent node, invalidating anchor. Because mutation needs `&mut`, all anchors must be killed before modification.
Co-authored-by: Aleksey Kladov <aleksey.kladov@gmail.com>
1175: Fix bugs and add error log about macro expansion r=matklad a=edwin0cheng
This PR fixed / add following things:
* Add a fused count which stop recursion of macro expansion in name resolution.
* Add some logs when macro expansion fails
* Add `$crate` meta variable support in mbe, which create a `$crate` ident token in token tree.
* Fixed matching a `$REPEAT` pattern inside a subtree, e.g. `(fn $name:ident {$($i:ident)*} ) => {...}`
* Remove composite-able punct token in syntax node to token conversion.
Co-authored-by: Edwin Cheng <edwin0cheng@gmail.com>
1154: Initial support for lang items (and str completion) r=flodiebold a=marcogroppo
This PR adds partial support for lang items.
For now, the only supported lang items are the ones that target an impl block.
Lang items are now resolved during type inference - this means that `str` completion now works.
Fixes#1139.
(thanks Florian Diebold for the help!)
Co-authored-by: Marco Groppo <marco.groppo@gmail.com>
1148: Add token_tree_to_xxx functions r=matklad a=edwin0cheng
<del>As discus in PR #1147 , this PR added a `mbe::MacroKind` .
Currently only 2 kind of macro are supported, `SourceFile` and `Block`.</del>
Added following functions for `tt::TokenTree` and `ast::Node` conversion:
* token_tree_to_expr
* token_tree_to_pat
* token_tree_to_ty
* token_tree_to_macro_stmts
* token_tree_to_macro_items
And added two new syntax kind:
* MACRO_ITEMS
* MACRO_STMTS
Co-authored-by: Edwin Cheng <edwin0cheng@gmail.com>
1145: Fix#1099, plug self type type parameters in infer_path_expr r=flodiebold a=edwin0cheng
As discussed in #1099, this PR try to "plug" a `self type` type parameters in `infer_path_expr`.
All the cases in 1099 was fixed and tested.
And luckily, this PR fixed bug #1030 again and make the test output correct.
Co-authored-by: Edwin Cheng <edwin0cheng@gmail.com>
1153: "Restart server" command r=jrvidal a=jrvidal
The only tricky aspect is that fact that once the `exit` command has been received, we no longer need to join on the reader thread.
Also, I think `terminateProcesses.sh` was not working properly. In fact, the very same script from the vscode language server implementation is not working either! It's because of that I noticed the reader thread issue 😮
Co-authored-by: Roberto Vidal <vidal.roberto.j@gmail.com>
This fixes the order in which candidates are chosen a bit (not completely
though, as the ignored test demonstrates), and makes autoderef work with trait
methods. As a side effect, this also makes completion of trait methods work :)
1137: Adds support for multiple editor workspaces on initialization r=matklad a=jrvidal
OK, so this "simple hack" turned out to be way more contrived than I expected 😂
### What works
This patch only handles multi-folder editor workspaces _on initialization_.
* I've found that modifying the layout of a workspace in VSCode just reloads the extension, so this hack should be enough for now.
* Not sure about how emacs-lsp behaves, but we fallback gracefully to the mono-folder workspace, so it should be fine.
### What doesn't work
* [x] `cargo watch` can only watch a single root folder with a `Cargo.toml`. I've left this part untouched but we could either warn that it's not supported or launch _multiple_ `cargo-watch` processes.
* [x] The `rust-analyzer/runnables` command is not functional, since we don't send the correct `cwd`.
* [x] Should we add some happy path test to `heavy_tests`?
* [ ] Going from a single `root` to multiple `roots` leaves us with a couple of `n * m` loops that smell a bit. The number of folders in the editor workspace is probably low though.
Co-authored-by: Roberto Vidal <vidal.roberto.j@gmail.com>
1138: Add L_DOLLAR and R_DOLLAR r=matklad a=edwin0cheng
As discussion in issue https://github.com/rust-analyzer/rust-analyzer/issues/1132 and PR #1125 , this PR add 2 `Syntax::Kind` : `L_DOLLAR` and `R_DOLLAR` for representing `Delimiter::None` in mbe and proc_marco. By design, It should not affect the final syntax tree, and will be discard in `TreeSink`.
My original idea is handling these 2 tokens case by case, but i found that they will appear in every place in the parser (imagine `tt` matcher). So this PR only handle it in `Parser::do_bump` and `Parser::start`, although It will not fix the `expr` matcher executing order problem in original idea.
Co-authored-by: Edwin Cheng <edwin0cheng@gmail.com>
- make it possible to get parent trait from method
- add 'obligation' machinery for checking that a type implements a
trait (and inferring facts about type variables from that)
- handle type parameters of traits (to a certain degree)
- improve the hacky implements check to cover enough cases to exercise the
handling of traits with type parameters
- basic canonicalization (will probably also be done by Chalk)
