There is some confusion about task categories and repository queries.
Right now, when creating task/issue you don't know if it will appear in
query or not, because query runs at the server side. as a work around,
task is being added to the root of task list or into the Archive
category. So, user does not have complete control on the structure of
the task list and Archive category is getting severely cluttered.
I would like to suggest some changes in the architecture of the task
list that would resolve these issues and will give more flexibility on
task management and repositories integration.
The idea is to always create local tasks. Each task can be
(optionally) linked to one (or more) repository. In this case, local
changes will be promoted remotely, as well as remote changes will be
brought back locally.
Then any category could also have one (or more) linked query that
would bring in new issues created in remote repository. For these query
hits user could create local tasks, either explicitly, or by scheduling,
assigning or adding cc'ing self to the task. After that these tasks will
be showing new updates. Note that user won't need query to see changes
on monitored tasks.
This approach would allow nice mix of local and remote tasks within