Does hatch run linter first?

[slavos1]

Hello @ofek, I've started using hatch recently and am so far happy with it!

One thing that perplexed me today: hatch fmt seems to run the linter first which, when it fails on a formatting issue ("line too long"), actually does not continue to the formatter (I have set my line length to 100):

$ hatch --version
Hatch, version 1.9.1

$ hatch fmt
my/file.py:36:101: E501 Line too long (119 > 100)
Found 1 error.

If I run the sequence:

$ hatch fmt -f
1 file reformatted, 5 files left unchanged
$ hatch fmt -l
# no errors

all is hunky-dory as the formatter has already wrapped the too-long line.

Checking the code, the linter will be started first:

    commands: list[list[str]] = []
    if not formatter:
        commands.append(environment.get_linter_command(*args, check=check, preview=preview)) # linter first

    if not linter:
        commands.append(environment.get_formatter_command(*args, check=check, preview=preview))

    with app.project.location.as_cwd(), internal_env.command_context():
        if not environment.config_path or sync:
            environment.write_config_file(preview=preview)

        for command in commands:
            process = app.platform.run_command(command)
            if process.returncode:
                app.abort(code=process.returncode)

On the master branch, the code changed but linter/formatter order was kept.

I have always run black first, then ruff. Is there a reason why hatch does it in a reverse order? Thanks!

[ofek]

I do this deliberately because fixing linting errors has no guarantee of adhering to formatting settings. The line too long is unfortunately the only case that I know of where you are forced to use the formatting flag to fix.

[slavos1]

Ok, thanks for a prompt reply @ofek. Will live with this.