Commit graph

4 commits

Author SHA1 Message Date
Dylan
a23e489c79
[refurb] Correctly handle lengths of literal strings in slice-to-remove-prefix-or-suffix (FURB188) (#16237)
Fixes false negative when slice bound uses length of string literal.

We were meant to check the following, for example. Given:

```python
  text[:bound] if text.endswith(suffix) else text
```
We want to know whether:
   - `suffix` is a string literal and `bound` is a number literal
   - `suffix` is an expression and `bound` is
       exactly `-len(suffix)` (as AST nodes, prior to evaluation.)
       
The issue is that negative number literals like `-10` are stored as
unary operators applied to a number literal in the AST. So when `suffix`
was a string literal but `bound` was `-len(suffix)` we were getting
caught in the match arm where `bound` needed to be a number. This is now
fixed with a guard.


Closes #16231
2025-02-18 12:52:26 -06:00
Dylan
14ee5dbfde
[refurb] Count codepoints not bytes for slice-to-remove-prefix-or-suffix (FURB188) (#13631) 2024-10-07 16:13:28 +02:00
Dylan
f110d80279
[refurb] Skip slice-to-remove-prefix-or-suffix (FURB188) when nontrivial slice step is present (#13405) 2024-09-19 12:47:17 -04:00
Dylan
b04948fb72
[refurb] Implement slice-to-remove-prefix-or-suffix (FURB188) (#13256) 2024-09-09 15:08:44 +00:00