todo-linter

Ensure to-do items in your code have an owner and an issue.

This is inspired by the todo_linter.py I wrote several years ago for ISARA.

todo-linter searches for various "TODO" tags in the given files, and complains if they don't include an owner and a reference to an issue or task describing the deferred work:

// TODO(chrish@pobox.com https://worktree.ca/taffer/todo-linter/issues/1): a short
// description

or:

# TODO(chrish@pobox.com https://worktree.ca/taffer/todo-linter/issues/1): a short
# description

or:

<!--
TODO(chrish@pobox.com https://worktree.ca/taffer/todo-linter/issues/1): a short
description
-->

The idea here is that future you, or someone else, will know where to get more context for the task, so they can actually address it at some point. The linter makes no attempt to verify the user email or the issue URL, so these can be whatever makes sense for your project.

You can run todo-linter against any file, as long as it's valid UTF-8.

To-do items

By default, todo-linter looks lines containing these tags: TODO, FIXME, XXX, followed by a user identifier, space, and URL inside of parentheses.

You can adjust these with the --delimiter, --issue-reference, --todo-tags, and --user-identifier arguments:

• --delimiter - Character to use as a delimiter between the user identifier and the issue reference. Defaults to space.
• --issue-reference - One of these values, such as --issue-reference url:
  • number - An issue number in your repo, such as #14. This is nice and short, but doesn't provide a clickable link like url does.
  • other - Free-form.
  • unsafe-url - Like url but for http://hostname/path type strings.
  • url - Matches https://hostname/path type strings.
• --todo-tags - A comma-separated list of to-do tags, such as --todo-tags To-do or --todo-tags TODO,FIXME,XXX (the default).
• --user-identifier - One of these values, such as --user-identifier email:
  • at-tag - Matches @user type strings, frequently found in Discord, Mastodon, Mattermost, Slack, etc.
  • bare - Matches any user.
  • email - Matches user@hostname type strings. (Default)

Again, todo-linter doesn't validate issue references or user identifiers.

You can also set these in a .todo-linter.toml file in the current directory (or anywhere is you use the --config argument).

Ignoring problems

You can tell todo-linter to ignore an invalid to-do item by adding an inline comment with todo-linter: ignore:

// I don't want to correct my linter errors:
//
// TODO: I hate writing real TODO items. todo-linter: ignore

# This one isn't actually an error:
#
# This is a big TODO. todo-linter: ignore

You can disable (and re-enable) the linter for larger blocks (or entire files) using todo-linter: disable and todo-linter: enable comments:

# todo-linter: disable
#
# Any TODO, FIXME, XXX, etc. in this block will be ignored by the linter.
# You can put todo-linter: disable at the top of a file to exclude it entirely.
#
# Use this to reactivate the linter:
#
# todo-linter: enable

Because todo-linter isn't parsing your code at all, these can appear anywhere in your comments (or even your code if that's your thing).

If it comes up, todo-linter: ignore takes precedent over disable/enable.

Installation

If you're using todo-linter with pre-commit, you can skip this.

To install todo-linter, use pipx from your clone of the repo:

$ pipx install .
  installed package todo-linter 1.0.0, installed using Python 3.14.5
  These apps are now available
    - todo-linter
done! ✨ 🌟 ✨

Using it with pre-commit

pre-commit is handy, you should try it!

To use todo-linter with your project, add this - repo stanza to your .pre-commit-config.yaml's repos: section:

repos:
  - repo: https://worktree.ca/taffer/todo-linter
    rev: {release}
    hooks:
      - id: todo-linter

Where {release} is the version tag you want to use, such as v1.0 or a commit hash.

Credits

Repo icon by Delapoutie on game-icons.net, licensed under the Creative Commons BY 3.0 license.

License

todo-linter is Creative Commons BY-NC-SA 4.0; see LICENSE.md for details.

No, you cannot train your LLM on this repo. Yes, I know you're going to ignore that, you slop-generating scum.