This attempts to bring better error messages to invalid method calls, by applying some heuristics to identify common mistakes.
The algorithm is inspired by Levenshtein distance and longest common sub-sequence. In essence, we treat the types of the function, and the types of the arguments you provided as two "words" and compute the edits to get from one to the other.
We then modify that algorithm to detect 4 cases:
- A function input is missing
- An extra argument was provided
- The type of an argument is straight up invalid
- Two arguments have been swapped
- A subset of the arguments have been shuffled
(We detect the last two as separate cases so that we can detect two swaps, instead of 4 parameters permuted.)
It helps to understand this argument by paying special attention to terminology: "inputs" refers to the inputs being *expected* by the function, and "arguments" refers to what has been provided at the call site.
The basic sketch of the algorithm is as follows:
- Construct a boolean grid, with a row for each argument, and a column for each input. The cell [i, j] is true if the i'th argument could satisfy the j'th input.
- If we find an argument that could satisfy no inputs, provided for an input that can't be satisfied by any other argument, we consider this an "invalid type".
- Extra arguments are those that can't satisfy any input, provided for an input that *could* be satisfied by another argument.
- Missing inputs are inputs that can't be satisfied by any argument, where the provided argument could satisfy another input
- Swapped / Permuted arguments are identified with a cycle detection algorithm.
As each issue is found, we remove the relevant inputs / arguments and check for more issues. If we find no issues, we match up any "valid" arguments, and start again.
Note that there's a lot of extra complexity:
- We try to stay efficient on the happy path, only computing the diagonal until we find a problem, and then filling in the rest of the matrix.
- Closure arguments are wrapped in a tuple and need to be unwrapped
- We need to resolve closure types after the rest, to allow the most specific type constraints
- We need to handle imported C functions that might be variadic in their inputs.
I tried to document a lot of this in comments in the code and keep the naming clear.
To make this work, the `#[rustc_on_unimplemented]` data needs to be used to
report method resolution errors, which is most of what this commit does.
Fixes#94581
This also reorders the fields to reduce the assembly operations for hashing
and changes two UI tests that depended on the former ordering because of
hashmap iteration order.
When encountering a binop where the types would have been accepted, if
all the predicates had been fulfilled, include information about the
predicates and suggest appropriate `#[derive]`s if possible.
Point at trait(s) that needs to be `impl`emented.
Use larger span for adjustment THIR expressions
Currently, we use a relatively 'small' span for THIR
expressions generated by an 'adjustment' (e.g. an autoderef,
autoborrow, unsizing). As a result, if a borrow generated
by an adustment ends up causing a borrowcheck error, for example:
```rust
let mut my_var = String::new();
let my_ref = &my_var
my_var.push('a');
my_ref;
```
then the span for the mutable borrow may end up referring
to only the base expression (e.g. `my_var`), rather than
the method call which triggered the mutable borrow
(e.g. `my_var.push('a')`)
Due to a quirk of the MIR borrowck implementation,
this doesn't always get exposed in migration mode,
but it does in many cases.
This commit makes THIR building consistently use 'larger'
spans for adjustment expressions. These spans are recoded
when we first create the adjustment during typecheck. For
example, an autoref adjustment triggered by a method call
will record the span of the entire method call.
The intent of this change it make it clearer to users
when it's the specific way in which a variable is
used (for example, in a method call) that produdes
a borrowcheck error. For example, an error message
claiming that a 'mutable borrow occurs here' might
be confusing if it just points at a usage of a variable
(e.g. `my_var`), when no `&mut` is in sight. Pointing
at the entire expression should help to emphasize
that the method call itself is responsible for
the mutable borrow.
In several cases, this makes the `#![feature(nll)]` diagnostic
output match up exactly with the default (migration mode) output.
As a result, several `.nll.stderr` files end up getting removed
entirely.
Fix incorrect disambiguation suggestion for associated items
Fixes#88806. I have not added a new test case, because the erroneous behavior is already present in existing test cases.
Currently, we use a relatively 'small' span for THIR
expressions generated by an 'adjustment' (e.g. an autoderef,
autoborrow, unsizing). As a result, if a borrow generated
by an adustment ends up causing a borrowcheck error, for example:
```rust
let mut my_var = String::new();
let my_ref = &my_var
my_var.push('a');
my_ref;
```
then the span for the mutable borrow may end up referring
to only the base expression (e.g. `my_var`), rather than
the method call which triggered the mutable borrow
(e.g. `my_var.push('a')`)
Due to a quirk of the MIR borrowck implementation,
this doesn't always get exposed in migration mode,
but it does in many cases.
This commit makes THIR building consistently use 'larger'
spans for adjustment expressions
The intent of this change it make it clearer to users
when it's the specific way in which a variable is
used (for example, in a method call) that produdes
a borrowcheck error. For example, an error message
claiming that a 'mutable borrow occurs here' might
be confusing if it just points at a usage of a variable
(e.g. `my_var`), when no `&mut` is in sight. Pointing
at the entire expression should help to emphasize
that the method call itself is responsible for
the mutable borrow.
In several cases, this makes the `#![feature(nll)]` diagnostic
output match up exactly with the default (migration mode) output.
As a result, several `.nll.stderr` files end up getting removed
entirely.
* On suggestions that include deletions, use a diff inspired output format
* When suggesting addition, use `+` as underline
* Color highlight modified span
E0599 suggestions and elision of generic argument if no canditate is found
fixes#81576
changes: In error E0599 (method not found) generic argument are eluded if the method was not found anywhere. If the method was found in another inherent implementation suggest that it was found elsewhere.
Example
```rust
struct Wrapper<T>(T);
struct Wrapper2<T> {
x: T,
}
impl Wrapper2<i8> {
fn method(&self) {}
}
fn main() {
let wrapper = Wrapper(i32);
wrapper.method();
let wrapper2 = Wrapper2{x: i32};
wrapper2.method();
}
```
```
Error[E0599]: no method named `method` found for struct `Wrapper<_>` in the current scope
....
error[E0599]: no method named `method` found for struct `Wrapper2<i32>` in the current scope
...
= note: The method was found for Wrapper2<i8>.
```
I am not very happy with the ```no method named `test` found for struct `Vec<_, _>` in the current scope```. I think it might be better to show only one generic argument `Vec<_>` if there is a default one. But I haven't yet found a way to do that,
This is step 2 towards fixing #77548.
In the codegen and codegen-units test suites, the `//` comment markers
were kept in order not to affect any source locations. This is because
these tests cannot be automatically `--bless`ed.
The different impls are all guarded by cfg-flags, and the revisions could be
used to cover the full power-set of combinations.
(I only included 20 of the possible 32 cases here; the null-set is not
interesting, and the remaining 11 all yielded ambiguous method resolution errors
which did not mix well with this testing strategy; I'm not trying to check UI
for the resolution diagnostics; I'm trying to create checkpoint of current
resolution semantics when compilation succeeds.)
Fixes#76267
When there is a single applicable method candidate, but its trait bounds
are not satisfied, we avoid saying that the method is "not found".
Insted, we update the error message to directly mention which bounds are
not satisfied, rather than mentioning them in a note.
If a symbol name can only be imported from one place for a type, and
as long as it was not glob-imported anywhere in the current crate, we
can trim its printed path and print only the name.
This has wide implications on error messages with types, for example,
shortening `std::vec::Vec` to just `Vec`, as long as there is no other
`Vec` importable anywhere.
This adds a new '-Z trim-diagnostic-paths=false' option to control this
feature.
On the good path, with no diagnosis printed, we should try to avoid
issuing this query, so we need to prevent trimmed_def_paths query on
several cases.
This change also relies on a previous commit that differentiates
between `Debug` and `Display` on various rustc types, where the latter
is trimmed and presented to the user and the former is not.
