qgil wrote on 2014-05-01 14:31:18 (UTC)
If we have a Phabricator instance one day, would we need to migrate the Trello data we have, or would it be enough to close past/ongoing sprints and start the new ones in Phabricator?
swalling wrote on 2014-05-01 18:46:26 (UTC)
Honestly I think we can do this by hand. There is not enough to bother working with the API on.
qgil wrote on 2014-05-06 00:52:42 (UTC)
Honestly I think we can do this by hand.
Alright, let's assume that this is the opinion from the rest of Trello users. Otherwise please reopen this task.
spage wrote on 2014-05-22 22:20:59 (UTC)
I reopened this because people can usefully work on this, but it shouldn't block any deployment.
Looking around, it seems you can
- create tasks in response to e-mails to e.g. email@example.com
- call the Conduit API from a script to create tasks, e.g. http://fab.wmflabs.org/conduit/method/maniphest.createtask/
The latter could be e.g. a Trellaphab script that repeatedly calls curl, or something...
I agree we should do this, though it should not block the main deployment. I am interested in this for migrating the Growth board, among potentially other projects.
Needless to say, you are welcome to work on such Trello - Phabricator
script. @chasemp and @mmodell can help you, but we would really appreciate
if you could still be the owners and main developers.
- Can flimport and/or the scripts to migrate Bugzilla and RT be recycled in
some way, or do you have another starting point?
- Many Trello tasks are in fact cloned Bugzilla reports. I have no idea how
many, which percentage. Is it possible to prevent an excess of duplicates
- These imports would occur project by project, manually by the respective
Trello owners, without needing to switch off Phabricator.
It is also clear that these migrations would only happen after the Bugzilla
I can provide some direction if someone wants to do this import. Things are arranged so it would be just another source.
Trello folks, we need to know which kind of support you are expecting from the Wikimedia Phabricator team (if any) in order to complete the Trello-Phabricator migration.
@Spage @mattflaschen,or being more specific, do you think you will need some of your time officially allocated to create a Trello migration script? Should we ask for some of your time in Q2 (sooner than later if we want everybody to migrate by the end of December)
@Qgil yes, project managers and scrum masters will have to spend time doing this. I suspect the first person who needs to migrate a lot of Trello cards (that aren't just bugzilla bugs duplicated in Trello) and who has some technical chops will hack something together, and others will improve it. Like how bingle evolved.
You're all busy, I don't expect any support except answer some e-mails and maybe some basic guidance. :-)
If some of my time is officially allocated, that could allow me to contribute to this effort more promptly (if I worked on it ad hoc, it would take longer, if I found time to work on it at all).
With Bugzilla-Migration completed and RT-Migration well on its way, it is time to look at the migration of Trello projects. Let's put all the Trello focus on T821: Trello -> Phabricator migration script, where @Spage has started to drive.
Meanwhile, are there any teams using Trello that have started using Phabricator beyond bug management, i.e. copying cards manually for a sprint?
I believe Core Features is going to have to do so for our next sprint (unless the import script is ready).