New support for subtasks by powerjungle · Pull Request #594 · pimutils/todoman

powerjungle · 2025-09-08T17:47:17Z

Add support for RELATED-TO property and RELTYPE parameter.

Taken from RFC 5545, section 3.8.4.5. If present in a file, these values are now read and stored in the cache.

Tests are modified to reflect the change.

The SCHEMA_VERSION is incremented to allow the cache to be recreated after the update.

Nothing is required from the user to be done for the changes in this PR to work.

Before:

After:

Closes: #396

balejk · 2025-10-01T07:38:42Z

If by cache you mean the cache database, it should be enough to just increment the schema version to force a rebuild.

powerjungle · 2025-10-01T14:55:59Z

If by cache you mean the cache database, it should be enough to just increment the schema version to force a rebuild.

Ah thank you, I totally missed it! I rebased my changes to make use of this and to describe it properly in comments and in the documentation. I edited the first comment of the PR to remove the "breaking changes" part since it's not true anymore.

jlkDE · 2026-04-10T12:10:29Z

Hey, I just rebased this to the current main for myself and as far as I can see everything works as intended, thanks for that!
What is needed for this to progress, is there some way to support this?

WhyNotHugo

Rebased this applying ruff format on the way.

Looks good, appreciate all the attention to detail.

Mostly minor feedback. There are two things which I think can be improved but can be done separate later (I'd rather keep this PR lean and not block on further changes):

Tree-building. This seems rather inefficient right now, but I won't block merging on that; we can iterate later gradually.
Style. I think the output would be easier to parse keeping the left [ ] $ID fields all aligned, maybe even something like:

[ ] 2 Some task
[ ] 4 ├─ Some subtask
[ ] 3 └─ Another subtask

Again, this seems like something that's best done separately.

WhyNotHugo · 2026-04-13T23:47:27Z

+            "related_to": "",
+            "related_to_reltype": "",


We should exclude the fields when absent, instead of including empty fields.

(same applies to other instances below)

WhyNotHugo · 2026-04-13T23:47:46Z

+If the database is changed in a breaking way, the ``SCHEMA_VERSION`` variable
+in the class ``Cache`` has to be incremented to allow the cache to be recreated
+after todoman has been updated. An example would be adding new fields or
+removing old unnecessary fields and etc.
+


Nice addition!

WhyNotHugo · 2026-04-13T23:53:06Z

+    # Merges all different property arguments into one dictionary
+    # `todo_properties` argument, so that it can be directly looped through
+    # easily for directly setting the `todoman.model.Todo` class attributes
+    # from within a command function.
+    #
+    # The names of the options are the same as the
+    # `todoman.model.Todo` class attributes.
    @functools.wraps(command)
    def command_wrap(*a, **kw) -> click.Command:


Nitpick: this should be a docstring rather than comment.

WhyNotHugo · 2026-04-13T23:54:19Z

 )

+_subtask_option = click.option(
+    "--subtask-for",


What do you think of calling this --parent?

Code-wise it's a bit more obvious, but I'm not sure how obvious it is for a user interface.

WhyNotHugo · 2026-04-13T23:55:04Z

        changes = True
        todo.description = sys.stdin.read()

+    if subtask_for is not None and not_subtask is False:


If subtask_for and not_subtask, we should return an error (invalid arguments).

WhyNotHugo · 2026-04-13T23:58:35Z

+            # are specified."
+            # Source:
+            # https://www.rfc-editor.org/rfc/rfc9253#section-9.1
+            if todo.related_to != "":


Check truthiness, so this also matches if it's None.

Suggested change

if todo.related_to != "":

if todo.related_to:

WhyNotHugo · 2026-04-14T00:04:37Z

+        return self._join_tree(tree).strip()
+
+    def _tree_reorder_related(
+        self, tree: dict[str, list], related_todos: dict[str, list]


Formatting:

Suggested change

self, tree: dict[str, list], related_todos: dict[str, list]

self,

tree: dict[str, list],

related_todos: dict[str, list],

WhyNotHugo · 2026-04-14T00:08:20Z

+        store_path: list = []
+        for related, related_to in related_todos.items():
+            # Find the root parent todo of the `related_to` todo.
+            related_to_path_tracer: str = related_to[0]
+            related_dict: dict[str, list] = tree
+            while related_to_path_tracer in related_todos:
+                related_to_path_tracer = related_todos[related_to_path_tracer][0]
+                store_path.append(related_to_path_tracer)
+
+            # Walk from the parent root todo, down to the
+            # `related_to` todo itself.
+            store_path.reverse()
+            # If `related_to` is a SIBLING, walk one todo backwards towards the
+            # root parent todo. If there is no path to be walked,
+            # just ignore it.
+            with contextlib.suppress(IndexError):
+                if related_to[1] == "SIBLING":
+                    related_to[0] = store_path.pop()
+            for inwards in store_path:
+                try:
+                    related_dict = related_dict[inwards][1]
+                    if related_dict is None:
+                        break
+                except KeyError:
+                    related_dict = tree
+                    break
+
+            self._tree_move_related(related, related_to[0], tree, related_dict)


This looks quite inefficient, since it walks the entire list multiple times. I think we need a "real" tree when building the whole tree in the first place.

But this is non-blocking at this stage; we can merge it in its current form and iterate on this later.

WhyNotHugo · 2026-04-14T00:09:03Z

+        self.related_to = ""
+        self.related_to_reltype = ""


Default should be None, not empty string. This affects the field's presence in the output icalendar file.

WhyNotHugo · 2026-04-14T00:11:28Z

+        got_related_to_reltype = ""
+        got_related_to = todo.get("related-to", "")
+        if got_related_to != "":
+            got_related_to_reltype = got_related_to.params.get("reltype")
+


As above, default should be None.

Taken from RFC 5545, section `3.8.4.5`. If present in a file, these values are now read and stored in the cache. Nothing in the UI has been changed. Tests are modified to reflect the change. Increment `SCHEMA_VERSION` to allow cache recreation on todoman update.

New names and values added to the JSON output for the `RELATED-TO` property.

List text output is now with a tree structure: task 1 task 2 task 3 task 4 task 5 The table list is changed to a dictionary representing a tree. No changes to the SQL queries are made. RELTYPEs PARENT, CHILD and SIBLING are supported. https://www.rfc-editor.org/rfc/rfc5545#section-3.2.15 Includes output space formatting fixes. Tests adjusted to changes.

Comment for describing the weird code for merging arguments.

Using the newly supported attributes, two new options are added. `--subtask-for`, making a task be a subtask for another and `--not-subtask`, making a task no longer be a subtask. The commands `new` and `edit` are fitted with these new options.

powerjungle force-pushed the powerjungle/subtasks branch from e87552a to b6ee7fa Compare October 1, 2025 14:50

WhyNotHugo force-pushed the powerjungle/subtasks branch 2 times, most recently from 064a26f to 8b78386 Compare April 13, 2026 23:50

WhyNotHugo requested changes Apr 14, 2026

View reviewed changes

WhyNotHugo force-pushed the powerjungle/subtasks branch from 8b78386 to d52e187 Compare April 14, 2026 01:16

powerjungle added 9 commits April 14, 2026 05:34

chore(gitignore): add .mypy_cache/

7259f6b

feat(porcelain): add related_to and related_to_reltype

be317d4

New names and values added to the JSON output for the `RELATED-TO` property.

refactor(cli): description of _todo_property_options wrapper

3e1d526

Comment for describing the weird code for merging arguments.

chore(CHANGELOG): describe new subtasks feature

5c72d46

chore(AUTHORS): add powerjungle

a3b5e9d

docs(Hacking): new description of SCHEMA_VERSION variable

373cdee

WhyNotHugo force-pushed the powerjungle/subtasks branch from d52e187 to 373cdee Compare April 14, 2026 03:34

Conversation

powerjungle commented Sep 8, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

balejk commented Oct 1, 2025

Uh oh!

powerjungle commented Oct 1, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

jlkDE commented Apr 10, 2026

Uh oh!

WhyNotHugo left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

powerjungle commented Sep 8, 2025 •

edited

Loading

powerjungle commented Oct 1, 2025 •

edited

Loading