8386: Avoid O(n²) when constructing AttrSourceMap r=jonas-schievink a=jonas-schievink
Brings https://github.com/rust-analyzer/rust-analyzer/issues/8377 down to 2.52s on my machine. Not quite back to where it was before, so I'll leave that issue open for now.
bors r+
Co-authored-by: Jonas Schievink <jonasschievink@gmail.com>
8371: Don't use HirDisplayWrapper when displaying SourceCode r=matklad a=Veykril
The issue was basically that when displaying for `DisplayTarget::SourceCode` some `hir_fmt` functions would create `HirDisplayWrapper`s which would then `fmt` these triggering the Display panic since `fmt::Display` can't fail the same way as `HirDisplay`. Simple fix is to just use `hir_fmt` directly. Should probably write that down somewhere in source, looking for a good spot to put that right now.
Fixes#8077, Fixes#8370
Co-authored-by: Lukas Wirth <lukastw97@gmail.com>
8375: feat: show errors from `cargo metadata` and initial `cargo check` in the status bar r=matklad a=matklad
bors r+
🤖
Co-authored-by: Aleksey Kladov <aleksey.kladov@gmail.com>
8364: Memory usage improvements r=jonas-schievink a=alexmaco
These are mostly focused on splitting up enum variants with large size differences between variants by `Box`-ing things up.
In my testing this reduces the memory usage somewhere in the low percentages, even though the measurements are quite noisy.
Co-authored-by: Alexandru Macovei <alexnmaco@gmail.com>
Rationale: only a minority of variants used almost half the size.
By keeping large members (especially in Option) behind a box
the memory cost is only payed when the large variants are needed.
This reduces the size Vec<Expr> needs to allocate.
8355: internal: do not drop errors from cargo metadata/check r=matklad a=matklad
Work towards #3155
Co-authored-by: Aleksey Kladov <aleksey.kladov@gmail.com>
- don't shift in/out for Chalk mapping (we want to have the same
binders now)
- do shift in when creating the signature for a closure (though it
shouldn't matter much)
- do shift in when lowering a `fn()` type
- correctly deal with the implied binder in TypeWalk
8353: Replace hir_ty::Lifetime with chalk equivalent r=flodiebold a=Veykril
Our `Lifetime` isn't really used yet so this is a rather simple change
Co-authored-by: Lukas Wirth <lukastw97@gmail.com>
8352: Remove dead legacy macro expansion code r=lnicola a=brandondong
I was investigating some unrelated macro issue when I noticed this dead code. This legacy macro expansion logic was changed in https://github.com/rust-analyzer/rust-analyzer/pull/8128.
Co-authored-by: Brandon <brandondong604@hotmail.com>
8351: Use more assoc. type aliases in the chalk interner r=flodiebold a=jonas-schievink
Makes it sligthly easier to swap out these types
Co-authored-by: Jonas Schievink <jonasschievink@gmail.com>
8350: internal: prepare to store OpQueue results in the queue itself r=matklad a=matklad
bors r+
🤖
Co-authored-by: Aleksey Kladov <aleksey.kladov@gmail.com>
8348: Make `Binders` more like Chalk r=flodiebold a=flodiebold
Working towards #8313.
- hide `value`
- use `VariableKinds`
- adjust `subst` to be like Chalk's `substitute`
- also clean up some other `TypeWalk` stuff to prepare for it being replaced by Chalk's `Fold`
Co-authored-by: Florian Diebold <flodiebold@gmail.com>
8245: Properly resolve intra doc links in hover and goto_definition r=matklad a=Veykril
Unfortunately involves a bit of weird workarounds due to pulldown_cmark's incorrect lifetimes on `BrokenLinkCallback`... I should probably open an issue there asking for the fixes to be pushed to a release since they already exist in the repo for quite some time it seems.
Fixes#8258, Fixes#8238
Co-authored-by: Lukas Wirth <lukastw97@gmail.com>
8326: Rewrite reorder fields assist to use mutable syntax trees r=matklad a=Veykril
This also instead uses `Either` to use the typed `RecordPat` and `RecordExpr` nodes, this unfortunately gives a bit of code duplication
Co-authored-by: Lukas Wirth <lukastw97@gmail.com>
8328: Move things in hir_ty into submodules r=flodiebold a=flodiebold
- all the types that will be replaced by Chalk go to `types`
- `TypeWalk` impls go to `walk`
- also fix signature of `Substitution::interned`
Co-authored-by: Florian Diebold <flodiebold@gmail.com>
8325: Check if bitflags deps pulls its weight r=jonas-schievink a=matklad
Bitflags is generally a good dependency -- it's lightweight, well
maintained and embraced by the ecosystem.
I wonder, however, do we really need it? Doesn't feel like it adds much
to be honest.
Co-authored-by: Aleksey Kladov <aleksey.kladov@gmail.com>
8295: Add `convert_into_to_from` assist r=Veykril a=obmarg
This adds a "Convert Into to From" assist, useful since clippy has
recently started adding lints on every `Into`.
It covers converting the signature, and converting any `self`/`Self`
references within the body.
It does assume that every instance of `Into` can be converted to a
`From`, which I _think_ is the case now. Let me know if there's
something I'm not thinking of and I can try and make it smarter.
Closes#8196
I'm extremely new to this codebase so please let me know if anything needs
changed.
Co-authored-by: Graeme Coupar <grambo@grambo.me.uk>
Bitflags is generally a good dependency -- it's lightweight, well
maintained and embraced by the ecosystem.
I wonder, however, do we really need it? Doesn't feel like it adds much
to be honest.
8322: Access a body's block def maps via a method r=jonas-schievink a=jonas-schievink
bors r+
Co-authored-by: Jonas Schievink <jonasschievink@gmail.com>
This adds a "Convert Into to From" assist, useful since clippy has
recently started adding lints on every `Into`.
It covers converting the signature, and converting any `self`/`Self`
references within the body to the correct types.
It does assume that every instance of `Into` can be converted to a
`From`, which I _think_ is the case now. Let me know if there's
something I'm not thinking of and I can try and make it smarter.
Only one upgradeable read lock can be handed out at the same time, and
we never acquire a non-upgradeable read lock, so this has no benefit
over just using a write lock in the first place.
8284: Reduce memory usage by using global `Arc`-based interning r=jonas-schievink a=jonas-schievink
This saves around 50 mb when running `analysis-stats` on r-a itself. Not a lot, but this infra can be easily reused to intern more stuff.
Co-authored-by: Jonas Schievink <jonasschievink@gmail.com>
8285: Don't recheck obligations if we have learned nothing new r=matklad a=flodiebold
This is just the most trivial check: If no inference variables have been updated, and there are no new obligations, we can just skip trying to solve them again. We could be smarter about it, but this already helps quite a bit, and I don't want to touch this too much before we replace the inference table by Chalk's.
Fixes#8263 (well, improves it quite a bit).
Co-authored-by: Florian Diebold <flodiebold@gmail.com>
8283: Resolve associated types r=flodiebold a=Veykril
Prior we were only resolving paths until the first type was found, then discarding the result if the path wasn't fully consumed. That of course causes associated types to not resolve. Fixes#5003
Co-authored-by: Lukas Wirth <lukastw97@gmail.com>
This is just the most trivial check: If no inference variables have been
updated, and there are no new obligations, we can just skip trying to
solve them again. We could be smarter about it, but this already helps
quite a bit, and I don't want to touch this too much before we replace
the inference table by Chalk's.
Fixes#8263 (well, improves it quite a bit).
8274: Adding a few more gifs and screenshots for features in manual r=Veykril a=MozarellaMan
Related #8267,#6539. Gifs are [here](https://github.com/rust-analyzer/rust-analyzer/issues/6539#issuecomment-809574840)
Finishing up the last PR, for the last two features that didn't have a visual example.
For syntax highlighting, I wasn't able to find a theme that displayed the difference between an enum and struct, but I only tried a few apart from the default so there could be one out there!
e.g., with the default light theme, `Ord` and `Ordering` in `use std::cmp::{Ord, Ordering}` had the same highlight colour. So I just went with displaying `mut` items being underlined.
Co-authored-by: Ayomide Bamidele <48062697+MozarellaMan@users.noreply.github.com>
8267: Adding gifs and screenshots for features in manual r=matklad a=MozarellaMan
For #6539
This includes most of gif or screenshot examples of most items in the "Features" header. With the exceptions of:
- **On Typing Assists** - couldn't get it to work for a demo, I'm probably missing something?
- **Structural search and replace** - looked to be already a visual example of the feature
- **Workspace symbol** - wasn't sure how best to show this, all of the examples maybe? Also wasn't sure of the best code example to show it off
- **Semantic Syntax Highlighting** - seemed obvious enough to not need a screenshot, but I could easily add this
All the gifs/pngs are hosted in this [comment](https://github.com/rust-analyzer/rust-analyzer/issues/6539#issuecomment-809574840). Please let me know if any of them aren't suitable (and why) and I'll improve it! Or if you don't like the theme/font
Co-authored-by: Ayomide Bamidele <48062697+MozarellaMan@users.noreply.github.com>
8266: Fix generic arguments being incorrectly offset in qualified trait casts r=flodiebold a=Veykril
We reverse the segments and generic args of the lowered path after building it, this wasn't accounted for when inserting the self parameter in `Type as Trait` segments.
Fixes#5886
Co-authored-by: Lukas Wirth <lukastw97@gmail.com>
We have a CLI for benchmarking, but no one actually uses it it seems.
Let's try switching to "internal" benchmarks, implemented as rust tests.
They should be easier to "script" to automate tracking of perf
regressions.
8247: internal: ensure that runaway type-inference doesn't block the main loop r=flodiebold a=matklad
We have a bug where type-checking `per_query_memory_usage` takes a
couple of seconds. It also reveals another bug: our type inference is
not cancellable.
Co-authored-by: Aleksey Kladov <aleksey.kladov@gmail.com>
We have a bug where type-checking `per_query_memory_usage` takes a
couple of seconds. It also reveals another bug: our type inference is
not cancellable.
