## Summary
This PR fixes the way NoQA range is inserted to the `NoqaMapping`.
Previously, the way the mapping insertion logic worked was as follows:
1. If the range which is to be inserted _touched_ the previous range, meaning
that the end of the previous range was the same as the start of the new
range, then the new range was added in addition to the previous range.
2. Else, if the new range intersected the previous range, then the previous
range was replaced with the new _intersection_ of the two ranges.
The problem with this logic is that it does not work for the following case:
```python
assert foo, \
"""multi-line
string"""
```
Now, the comments cannot be added to the same line which ends with a continuation
character. So, the `NoQA` directive has to be added to the next line. But, the
next line is also a triple-quoted string, so the `NoQA` directive for that line
needs to be added to the next line. This creates a **union** pattern instead of an
**intersection** pattern.
But, only union doesn't suffice because (1) means that for the edge case where
the range touch only at the end, the union won't take place.
### Solution
1. Replace '<=' with '<' to have a _strict_ insertion case
2. Use union instead of intersection
## Test Plan
Add a new test case. Run the test suite to ensure that nothing is broken.
### Integration
1. Make a `test.py` file with the following contents:
```python
assert foo, \
"""multi-line
string"""
```
2. Run the following command:
```console
$ cargo run --bin ruff -- check --isolated --no-cache --select=F821 test.py
/Users/dhruv/playground/ruff/fstring.py:1:8: F821 Undefined name `foo`
Found 1 error.
```
3. Use `--add-noqa`:
```console
$ cargo run --bin ruff -- check --isolated --no-cache --select=F821 --add-noqa test.py
Added 1 noqa directive.
```
4. Check that the NoQA directive was added in the correct position:
```python
assert foo, \
"""multi-line
string""" # noqa: F821
```
5. Run the `check` command to ensure that the NoQA directive is respected:
```console
$ cargo run --bin ruff -- check --isolated --no-cache --select=F821 test.py
```
fixes: #7530
## Summary
Implement
[`no-ignored-enumerate-items`](https://github.com/dosisod/refurb/blob/master/refurb/checks/builtin/no_ignored_enumerate.py)
as `unnecessary-enumerate` (`FURB148`).
The auto-fix considers if a `start` argument is passed to the
`enumerate()` function. If only the index is used, then the suggested
fix is to pass the `start` value to the `range()` function. So,
```python
for i, _ in enumerate(xs, 1):
...
```
becomes
```python
for i in range(1, len(xs)):
...
```
If the index is ignored and only the value is ignored, and if a start
value greater than zero is passed to `enumerate()`, the rule doesn't
produce a suggestion. I couldn't find a unanimously accepted best way to
iterate over a collection whilst skipping the first n elements. The rule
still triggers, however.
Related to #1348.
## Test Plan
`cargo test`
---------
Co-authored-by: Dhruv Manilawala <dhruvmanila@gmail.com>
Co-authored-by: Charlie Marsh <charlie.r.marsh@gmail.com>
**Summary** Instead of emitting a bogus token per char, we now only emit
on single last bogus token. This leads to much more concise output.
**Test Plan** Updated fixtures
Closes#5497
Needs MkDocs 1.5 to be released.
- [x] https://github.com/mkdocs/mkdocs/milestone/15
## Summary
Uses MkDocs' `not_in_nav` config to hide spam about files in
`docs/rules/` not being in nav.
Close#7479
The `@override` was already implemented
## Test Plan
Tested the code in the issue. After removing all the noqa's, only one
occurrence of `BadName()` raised a violation.
Added a fixture
## Summary
Fix CI (broken in #7496).
The code snippet was formatted as Black would format a stub file, but
the CI script doesn't know that (it assumes all code snippets are
non-stub files). Easier to ignore.
Sorry for breaking CI!
## Test Plan
`python scripts/check_docs_formatted.py`
## Summary
This PR implements a new rule for `flake8-logging` plugin that checks
for uses of `logging.exception()` with `exc_info` set to `False` or a
falsy value. It suggests using `logging.error` in these cases instead.
I am unsure about the name. Open to suggestions there, went with the
most explicit name I could think of in the meantime.
Refer https://github.com/astral-sh/ruff/issues/7248
## Test Plan
Added a new fixture cases and ran `cargo test`
## Summary
This PR updates the `W191` (`tab-indentation`) rule from a line-based to
a token-based rule.
Earlier, the rule used the `triple_quoted_string_ranges` from the
indexer to skip over any lines _inside_ a triple-quoted string. This was the only
use of the ranges. These ranges were extracted through the tokens, so instead
we can directly use the newline tokens to perform the check.
This would also mean that we can remove the `triple_quoted_string_ranges` from
the indexer but I'll hold that off until we have a better idea on #7326
but I don't think it would be a problem to remove it.
This will also fix#7379 once PEP 701 changes are merged.
## Test Plan
`cargo test`
Micha Reiser
Dhruv Manilawala
dependabot[bot]
Charlie Marsh
Mathieu Kniewallner
Tom Kuson
konsti
Valeriy Savchenko
Micael Jarniac
Jonathan Plasse
qdegraaf
Jaap Roes