Fix the mixed-use of UUID and str for task IDs

[gycsaba96]

Fixes #1077

The cause of the (seemingly) extra empty line was that existing subtasks were not correctly detected and thus moved to the end of the text. This happened because of the mixed use of strings and UUIDs for task IDs.

This PR converts strings to UUIDs when necessary and fixes the related type hints.

[SqAtx]

Thank you for looking at it! I kind of got bogged down in #1083 on my way to add tests for this part of the code ^^'

I see that this fixes the bug, but I don't understand why. How were the subtasks not correctly detected? Weren't they detected by line.lstrip().startswith('{!'):, which would be the same whether we use UUID or str?

[gycsaba96]

Thank you for looking at it! I kind of got bogged down in #1083 on my way to add tests for this part of the code ^^'

Don't give up, QA is very important. :)

I see that this fixes the bug, but I don't understand why. How were the subtasks not correctly detected? Weren't they detected by line.lstrip().startswith('{!'):, which would be the same whether we use UUID or str?

Let me provide more details.

1. When the subtask is created, it is stored in the task store's lookup dictionary with a UUID key.
2. Line 772 and 777 extract and store the subtask's id as a string.
3. Line 791 raises a KeyError because we want to do the lookup using a string key, although the subtask does exist in the store with a UUID key. (Sorry for referencing the wrong line.)
4. As a consequence, it remains in the subtasks list and gets erased from the text by the for loop at line 801. This is how it is removed from the original position.
5. Finally, it gets added back to the end of the text by editor.py#210 when you reopen the editor.

[diegogangl]

Merged, thanks!

[SqAtx]

@gycsaba96 ahhh I see, thank you! I haven't quite wrapped my head around the entire task/editor logic yet