Bump Ruff to 0.12.9, update output parsing for new rendering

[hramezani]

Ruff 0.12.9 changed the full output rendering (astral-sh/ruff#19966): the diagnostic header is split so the location is emitted on a separate --> file:line:col line instead of the old file:line:col: CODE message single line. This broke ruff_check's offset rewriting and the associated tests.

This bumps the minimum Ruff to >=0.12.9 and updates the parsing to the new format.

Relationship to #65

This builds directly on @ntBre's work in #65 — same root cause and same fix approach. I opened a fresh PR for two reasons:

1. Bump Ruff to 0.12.9, update regexes for new output rendering #65 now conflicts with main (after Add support for Python 3.14, drop EOL 3.8-3.9 #67 dropped EOL 3.8/3.9 and touched the same requires-python line). This branch is cut from current main, so it merges cleanly.
2. It additionally handles the case where the lint error is on a multi-digit line number (see below).

Happy to fold these changes into #65 instead if that's preferable — I just wanted to get a mergeable, rebased version up.

Multi-digit line numbers

Ruff indents the --> arrow to align with the source gutter, so the leading whitespace widens once the line number reaches two digits:

 --> -:1:7      # 1 leading space
  --> -:10:7    # 2 leading spaces

To keep the offset substitution working regardless of line-number width, this matches variable leading whitespace (r'^( *)--> -:(\d+)') and preserves it in the replacement. It also adds test_ruff_offset_multiline, which puts the error on line 10 (and again with start_line=10) so this stays covered.

All existing tests pass, plus ruff format --check, ruff check, and pyright are clean.

[ntBre]

I'm also happy to rebase #65. I wasn't aware that that was the blocker to moving forward with the PR.

[hramezani]

I'm also happy to rebase #65. I wasn't aware that that was the blocker to moving forward with the PR.

would be great if you could rebase your PR. Please consider the fix that I did for multi-digit line number and include the test

[ntBre]

Actually, I'll just close mine in favor of this given your additional improvements. Thanks for picking it up!

[hramezani]

Actually, I'll just close mine in favor of this given your additional improvements. Thanks for picking it up!

Sorry @ntBre. I created the new PR because I didn't want to bother you after a long time for the rebase and review!

[ntBre]

No worries at all, I'm very happy to leave it in your hands! Let me know if you end up needing anything from me :)