Describes itself as a RubyLinter when it should just be a Linter

[gerph]

The linter declares itself as a RubyLinter, but when you're running tools you shouldn't care what language the tool is implemented in and the interpreter that should be used to run it. That's what the #! lines are for - to say how you run the tool.

Why is this a problem? Well, the tool called mdl in my path isn't a ruby script. It's a shell script that runs the markdownlint/markdownlint docker container on the command that was supplied. This avoids my having to install ruby and the mdl tool into the environment.

So I have to change the linter.py code from:

class Mdl(RubyLinter):

to

class Mdl(Linter):

[kaste]

That's somehow correct. I too think the RubyLinter implementation is outdated that's why we tried a simplified new implementation for rubocop only supporting "rvm" and "bundle". See https://github.com/SublimeLinter/SublimeLinter-rubocop/blob/master/linter.py#L5-L22

Do you think we should use that here as well?

It is otherwise not necessary to fork just for your use case. You can override the default lookup algo by setting "executable" in the settings. Here setting it to just "mdl" (if it's an executable in PATH) would be enough. (http://www.sublimelinter.com/en/stable/linter_settings.html#executable)