B905's autofix sets strict=False, but the suggested fix is strict=True

[mdbernard]

The docs for B905 state the suggested fix is to use strict=True, but the autofix introduced here autofixes with strict=False.

Either the docs or the autofix should be updated to match the other.

I would argue that the autofix should be changed to strict=True, since that is the option less likely to cause unexpected runtime behavior for users. It might cause runtime exceptions to be raised, but I personally would prefer those to silently continuing with potentially unintended behavior.

---

Keywords: B905, zip-without-explicit-strict

Current ruff version: 0.6.6

Minimal snippet:

zip(a, b)

Command invoked:
ruff check example.py --target-version=py311 --select=B905

Output:

example.py:1:1: B905 [*] `zip()` without an explicit `strict=` parameter
  |
1 | zip(a, b)
  | ^^^^^^^^^ B905
  |
  = help: Add explicit `strict=False`

Found 1 error.
[*] 1 fixable with the `--fix` option.

Expected output:

...
  = help: Add explicit `strict=True`
...

[zanieb]

I wonder if we use strict=False to avoid changing the semantics / behavior of the code? I think this is probably correct, but we should recommend adding strict not necessarily strict=False in the message?

[mdbernard]

That also works! Just removing the autofix altogether and letting the user decide what they want. Based on personal experience I'd almost always want strict=True, but I see the counter-argument that it changes behavior, so it might not be a desirable autofix.

[zanieb]

I don't think we should remove the autofix entirely though, just change the message.