- In-code TODOs are often more useful than tasks in project management apps, as they're co-located with the relevant code
- In many projects, they grow out of control (everything is just a "TODO" with nothing indicating priority or category)
- Working on related tasks at the same time makes more sense than context switching between different tasks. However, proximity in code doesn't necessarily correlate to "task proximity". Two TODOs in the same file may be two completely different tasks, but two TODOs across different files may be very similar
- There's a difference between TODOs written as quick notes while context switching between different parts of the code while working on a single task, and long term TODOs that are part of the code
Numbered (priority) todos, indicating what needs to be solved now and in what order.
And category todos grouping tasks into categories.
Syntax: todo{number} {description?}
, number can be:
{1, 2, 3, ...}
— the higher the number the lower the priority (i.e.todo1
is more urgent thantodo2
)0
/00
/000
/... — the more zeroes, the more urgent the task is
As an example:
// todo000 something part of currently written code, needs to be solved first (highest amount of 0s)
// todo00 something that needs to be resolved right after ^ this task
// todo0 something that needs to be resolved after the above task
// todo1 something even lower priority
// todo2 something *even* lower priority
// todo3 etc
In general, all priority todos (todos with numbers) need to be solved before the current work is considered complete.
When making a lot of changes in a single commit, this may mean resolving all priority todos before committing changes.
In practice, the usage may look like this:
- you're working on
x
, and as part of that you need to work ony
- there are still some things unresolved in
x
, but you need to work ony
to move forward. You leave atodo1
in thex
part of the code - while working on
y
, you leave atodo0
for something that needs to be resolved before returning tox
- while working on that, you notice another thing that needs to be solved, even before the
todo0
. You leave atodo00
Syntax: todo@{category} {description?}
A way to group TODOs by category.
Examples:
todo@responsivity Hide this on mobile
todo@darkmode Improve input styling
todo@types
todo@testing test this
Any todos that don't fall into the two categories above (i.e. their syntax isn't todo{number}
or todo@{category}
).
TODO: Fix this
todo refactor
In some larger projects, we also keep track of TODOs in markdown files. This is useful when the task is more abstract and not immediately related to any given piece of code.
We follow this convention:
README.md
# Project name
## Some section
...
## TODOs
- foo
- bar
And in larger projects, we have often have a dedicated file for TODOs:
todo.md
- Generic todo 1
- Generic todo 2
## Category 1
- foo
- bar
## Category 2
- abc
- def
As a general rule, in our code, priority todos may not be pushed into master. They need to be resolved before committing (ideally) or before merging PRs (when working on larger things).
To validate this automatically, you can set up a simple GitHub Action:
on: [push, pull_request]
jobs:
validate:
name: Validate code
runs-on: ubuntu-latest
steps:
- uses: actions/checkout@v2
- name: Check for todo0
run: '! grep -r "todo0" --exclude-dir=workflows .'
if: always()
- name: Check for todo1
run: '! grep -r "todo1" --exclude-dir=workflows .'
if: always()
- name: Check for todo2
run: '! grep -r "todo2" --exclude-dir=workflows .'
if: always()
The benefit of TODOs in code is that they're searchable, and searching todo
makes any kind of todo show up (since it doesn't matter whether it's followed by a number, an at sign, or whitespace).
That said, this repo includes a simple CLI tool written in Rust for getting an easy-to-read, ANSI-colored, Markdown-formatted list of all todos in a project.
Usage:
todos --exclude node_modules src/
Output:
# TODOs
<!-- priority todos -->
## todo00
- [ ] foo (/file:123)
- [ ] bar (/file:456)
## todo0
- [ ] abc (/file:123)
- [ ] def (/file:456)
<!-- category todos -->
## testing
- [ ] abc (/file:123)
- [ ] def (/file:456)
## responsivity
- [ ] abc (/file:123)
- [ ] def (/file:456)
<!-- generic todos -->
## Other
- [ ] abc (/file:123)
- [ ] def (/file:456)
(without the HTML comments).
Notes:
node_modules/
(for npm) andvendor/
(for composer) are excluded by default- paths starting with
.
are always excluded --exclude
s are relative to the current working directory, not passed paths (including default excludes mentioned above). If you're running the script for another folder and want to exclude folders there, type out the path in--exclude
- Passing any excludes overrides the default excludes, so if you want to add to the list of excludes, you need to re-define the default ones as well (e.g.
-e node_modules
)
The tool also scans a todo.md file (path can be provided using --todos
):
- all TODOs have to be list items (
- foo
or- [ ] foo
) - any TODOs above the first heading are considered generic TODOs
- any TODOs under a heading are considered category TODOs, with the heading being the category name
- any TODOs with numbers are added to the list of priority TODOs
Scanning TODOs in a README.md file is also supported:
- all TODOs have to be list items (
- foo
or- [ ] foo
) - they have to be directly under a
TODO[s:]
(lower or uppercase) heading
See the samples/
folder for examples.
To omit ANSI formatting and get raw markdown output, set NO_COLOR=1
or TERM=dumb
.
There are no downloadable builds at the moment. To compile the tool manually:
- Set up Rust locally https://www.rust-lang.org/tools/install
cargo install --git https://github.com/archtechx/todo-system.git
- The tool will be added to your
$PATH
automatically astodos
I personally also like creating an alias that does todos | less
:
alias t="todos | less"