Micha Reiser 853d8354cb JSON Emitter: Use one indexed column numbers for edits (#4007) ...

I noticed in the byte-offsets refactor that the `JsonEmitter` uses one indexed column numbers for the diagnostic start and end locations but not for `edits`.

This PR changes the `JsonEmitter` to emit one-indexed column numbers for edits, as we already do for `Message::location` and `Message::end_location`.

## Open questions

~We'll need to change the LSP to subtract 1 from the columns in `_parse_fix`~

6e44fadf8a/ruff_lsp/server.py (L129-L150)

~@charliermarsh is there a way to get the ruff version in that method? If not, then I recommend adding a `version` that we increment whenever we make incompatible changes to the serialized message. We can then use it in the LSP to correctly compute the column offset.~

I'll use the presence of the `Fix::applicability` field to detect if the Ruff version uses one or zero-based column indices.

See https://github.com/charliermarsh/ruff-lsp/pull/103

2023-05-10 17:21:02 +02:00

..

black_compatibility_test.rs

refactor: Introduce crates folder (#2088)

2023-02-05 16:47:48 -05:00

integration_test.rs

JSON Emitter: Use one indexed column numbers for edits (#4007)

2023-05-10 17:21:02 +02:00