[`useless_vec`]: add more tests and don't lint inside of macros
Closes#11084.
I realized that the fix I added in #11081 itself also causes an error in a suggestion when inside of a macro. Example:
```rs
macro_rules! x {
() => {
for _ in vec![1, 2] {}
}
}
x!();
```
Here it would suggest replacing `vec![1, 2]` with `[x!()]`, because that's what the source callsite is (reminder: it does this to get the correct span of `x!()` for code like `for _ in vec![x!()]`), but that's wrong when *inside* macros, so I decided to make it not lint if the whole loop construct is inside a macro to avoid this issue.
changelog: [`useless_vec`]: add more tests and don't lint inside of macros
r? `@Alexendoo` since these were your tests, I figured it makes most sense to assign you
lint/ctypes: ext. abi fn-ptr in internal abi fn
Fixes#94223.
- In the improper ctypes lint, instead of skipping functions with internal ABIs, check that the signature doesn't contain any fn-ptr types with external ABIs that aren't FFI-safe.
- When computing the ABI for fn-ptr types, remove an `unwrap` that assumed FFI-safe types in foreign fn-ptr types.
- I'm not certain that this is the correct approach.
Currently, the output of `rustc --explain foo` displays the raw markdown in a
pager. This is acceptable, but using actual formatting makes it easier to
understand.
This patch consists of three major components:
1. A markdown parser. This is an extremely simple non-backtracking recursive
implementation that requires normalization of the final token stream
2. A utility to write the token stream to an output buffer
3. Configuration within rustc_driver_impl to invoke this combination for
`--explain`. Like the current implementation, it first attempts to print to
a pager with a fallback colorized terminal, and standard print as a last
resort.
If color is disabled, or if the output does not support it, or if printing
with color fails, it will write the raw markdown (which matches current
behavior).
Pagers known to support color are: `less` (with `-r`), `bat` (aka `catbat`),
and `delta`.
The markdown parser does not support the entire markdown specification, but
should support the following with reasonable accuracy:
- Headings, including formatting
- Comments
- Code, inline and fenced block (no indented block)
- Strong, emphasis, and strikethrough formatted text
- Links, anchor, inline, and reference-style
- Horizontal rules
- Unordered and ordered list items, including formatting
This parser and writer should be reusable by other systems if ever needed.
Map our diagnostics to rustc and clippy's ones
And control their severity by lint attributes `#[allow]`, `#[deny]` and ... .
It doesn't work with proc macros and I would like to fix that before merge but I don't know how to do it.
internal: Format let-else
As nightly finally got support for it I went ahead and formatted r-a with the latest nightly, then with the latest stable (in case other stuff changed)
Implement selection via new trait solver
Implements selection via the new solver by calling into `assemble_and_evaluate_candidates`, doing a very light-weight "winnow" to choose one candidate over the others, and then re-confirming that candidate outside of the `EvalCtxt` in order to compute the information necessary for the `ImplSource`.
This selection routine is best effort -- if it receives an ambiguous response from the `EvalCtxt`, then it may return ambiguity if the work to re-confirm the goal is not "easy" -- right now, that means everything except for object and impl goals. Since we don't want to reimplement all of the work of the `compute_builtin_*` methods in the solver internals, we bail out if we encounter ambiguity more often than the old solver. This should only barely affect [method selection](f798ada7ba/compiler/rustc_hir_typeck/src/method/probe.rs (L1447)) and not codegen. It can be made more precise later if needed.
r? `@lcnr`
Split out project loading capabilities from rust-analyzer crate
External tools currently depend on the entire lsp infra for no good reason so let's lift that out so those tools have something better to depend on
Revert "Suggest `x build library` for a custom toolchain that fails to load `core`"
This reverts commit b913f5593d.
CI builds with profile=nightly, causing different test output.
Making the output depend on the release channel was not a great idea and we have to be more careful with this.
I did not realize that the change could have such effects on test output but it's kinda obvious with hindsight >.<.
Clean up `ImportMap`
There are several things in `hir_def::import_map` that are never used. This PR removes them and restructures the code. Namely:
- Removes `Query::name_only`, because it's *always* true.
- Because of this, we never took advantage of storing items' full path. This PR removes `ImportPath` and changes `ImportInfo` to only store items' name, which should reduce the memory consumption to some extent.
- Removes `SearchMode::Contains` for `Query` because it's never used.
- Merges `Query::assoc_items_only` and `Query::exclude_import_kinds` into `Query::assoc_mode`, because the latter is never used besides filtering associated items out.
Best reviewed one commit at a time. I made sure each commit passes full test suite. I can squash the first three commits if needed.
Don't lint manual_let_else in cases where ? would work
Don't lint `manual_let_else` where the question mark operator `?` would be sufficient, that is, mostly in cases like:
```Rust
let v = if let Some(v) = ex { v } else { return None };
```
Also, this PR emits the `question_mark` lint for `let...else` patterns that could be written with `?` (also, only `return None` like cases).
```
changelog: [`manual_let_else`]: don't lint in cases where question_mark already lints
changelog: [`question_mark`]: lint for `let Some(...) = ex else { return None };`
```
Fixes #8755
internal: Add analysis-stats flag to trigger some IDE features
Closes https://github.com/rust-lang/rust-analyzer/issues/15131
Running this on r-a showed an 86mb memory increase, but that was without running it on the deps, will try that later when I don't need to use my pc.
Shuffle some proc_macro_expand query things around
Removes some unnecessary extra work we are doing in proc-macro expansion, and more importantly `Arc` the result of the proc_macro_expand query, that way we can reuse the instance for the `macro_expand` query's result
This reverts commit b913f5593d.
CI builds with profile=nightly, causing different test output.
Making the output depend on the release channel was not a great idea.
Extend previous checks for external ABI fn-ptrs to use in internal
statics, constants, type aliases and algebraic data types.
Signed-off-by: David Wood <david.wood@huawei.com>
Extend previous commit's support for checking for external fn-ptrs in
internal fn types to report errors for multiple found fn-ptrs.
Signed-off-by: David Wood <david.wood@huawei.com>
Instead of skipping functions with internal ABIs, check that the
signature doesn't contain any fn-ptr types with external ABIs that
aren't FFI-safe.
Signed-off-by: David Wood <david.wood@huawei.com>
