RUF025 is remapped to C420 but now it's also a new rule

[oprypin]

• #12533 remapped the rule RUF025 as C420.

• #15104 added a different new rule under the code RUF025.

The immediate issue is that even though RUF025 has just appeared in documentation, if you enable it, the rule doesn't actually work, and you just get the following warning:

warning: `RUF025` has been remapped to `C420`.

So I suppose (luckily?) the new rule was not actually added yet because it's impossible to use it, it only appears in docs and in the codebase.

Rule codes that were remapped from are supposed to remain reserved forever, right?

[MichaReiser]

Rule codes that were remapped from are supposed to remain reserved forever, right?

Argh, thanks. This is such a foot gun, especially with the many holes in the Ruff code section. I assumed that the redirects are at least visible in the codes section because it's the case for deprecated rules. I'll look into this on Monday, but we should have a static check that asserts that codes and redirects don't overlap.

[InSyncWithFoo]

Current RUF "holes":

• RUF004: Redirects to B026 (RUF004 appears to be a duplicate of B026 #3290, Redirect RUF004 to B026 #3293)
• RUF014: Would be a duplicate of PLW0101 (Add unreachable code rule #5384, [pylint] Re-implement unreachable (PLW0101) #10891)
• RUF037: One rejected ([ruff] Unformatted special comments (RUF037) #14111)
• RUF042: One rejected ([ruff] d.update({k: v}) (RUF042) #14564) and one pending ([ruff] Unnecessary intermediate representation (RUF042) #14831)
• RUF044: One pending ([ruff] Abstract method defined in normal class (RUF044) #14688)
• RUF045: One pending ([ruff] Implicit class variable in dataclass (RUF045) #14349)
• RUF047: One rejected ([ruff] Unnecessary re.escape() call (RUF047) #14746) and one pending ([ruff] Needless else clause (RUF047) #15051)
• RUF049: Empty
• RUF050: Two rejected ([ruff] Implemented wrong-class-body-content (RUF050) #14607, [ruff] Empty branches (RUF050) #14763)
• RUF053: Empty
• RUF054: Empty

It should perhaps be emphasized that I opened most of the aforementioned PRs...

[MichaReiser]

So I suppose (luckily?) the new rule was not actually added yet because it's impossible to use it, it only appears in docs and in the codebase.

I think it was possible to use the rule using the ALL selector but yes, most users weren't able to use it.

Thank you @InSyncWithFoo for volunteering to fix this issue so quickly! I also want to make clear that this was entirely my fault and by no means yours, just because we skipped a few rule codes.