`autofix` option for lint messages?

In addition to the existing column, line, ruleId, message, severity, etc. properties defined for a lint message, would you be open to supporting an autofix property (and updating many of the built-in rules to use it)?

For example, a JavaScript linter that tells you that you are missing a semicolon should provide the user the option to just add the semicolon in the right spot. This is something that could easily have an autofix option.

I was thinking the value of autofix could be either a string diff to apply to the file to fix the lint or a JavaScript function that would take the source as input and return the “fixed” source as output (or it could return the diff as a string). The latter is not JSON-serializable, but might be useful for lazily generating the value from autofix. In an IDE that runs the linter repeatedly, it’s important that the linter runs quickly, so I want to avoid a situation where autofix support slows things down.

Arcanist Lint supports the autofix property, though it appears to be a severity level and the patches are a separate property. There isn’t a code example there, but it sounds like there can be an array of patches, which seems sensible.

See also https://github.com/AtomLinter/Linter/issues/418

<bountysource-plugin>

Want to back this issue? Post a bounty on it! We accept bounties via Bountysource. </bountysource-plugin>

[bolinfest]

It always seems a little lame when a computer can tell me that I’m missing a semicolon yet can’t add it for me.

[gyandeeps]

I am going to close this issue as this is one of the items on this issue: https://github.com/eslint/eslint/issues/2203 For further discussions lets talk on the above mentioned issue.