1555409 - D32774 not moved out of Draft state

Status: RESOLVED FIXED

(bugzilla.mozilla.org :: Phabricator Integration, defect, P1)

Description

[:glob ✱]

https://phabricator.services.mozilla.com/D32774 was created with --changes-planned.

BMO set the policy and groups correctly, however it hasn't removed it from the Draft state.

See https://bugzilla.mozilla.org/show_bug.cgi?id=1513116#c10

Comment 1

[:glob ✱]

setting as P1 for investigation

Comment 2

[:glob ✱]

(David Lawrence [:dkl] from bug 1513116comment #15

(In reply to Botond Ballo [:botond] from comment #14)

(In reply to Byron Jones ‹:glob› 🎈 from comment #13)

Are you able to provide the revision where the transition from WIP to GTG didn't remove the draft state?

I just updated the same patch D32774 to not be WIP and it's still in the draft state.

In the action menu, is there an request review choice? when a revision is ready for review, choosing that action should take it out of the draft state. BMO will not move a revision out of the draft state if it is set to "changes planned", etc.

given we have a different meaning of draft from phabricator's usual meaning - not yet processed by bugzilla - this seems to be a bug. at the least it's confusing.

Comment 3

[David Lawrence [:dkl]]

BMO Is incorrectly looking at revision status and instead needs to look at the isdraft flag when deciding to remove the draft state. Will investigate how this should be implemented.

Comment 4

[David Lawrence [:dkl]]

Attachment: GitHub Pull Request

Comment 5

[David Lawrence [:dkl]]

Merged to master.

Comment 6

[Botond Ballo [:botond]]

Patches still end up in the draft state after resubmitting without --wip. Example: https://phabricator.services.mozilla.com/D34256. Should I file a new bug for that?

Comment 7

[David Lawrence [:dkl]]

(In reply to Botond Ballo [:botond] from comment #6)

Patches still end up in the draft state after resubmitting without --wip. Example: https://phabricator.services.mozilla.com/D34256. Should I file a new bug for that?

With the latest changes, now when you submit with --wip, you need to manually choose the request review action from the bottom of the revision. If you are the author, it should be in the action list. BMO will not move it out of draft state anymore if --wip is used. Does this work?

Comment 8

[David Lawrence [:dkl]]

Wrong needinfo.

Comment 9

[Botond Ballo [:botond]]

(In reply to David Lawrence [:dkl] from comment #7)

With the latest changes, now when you submit with --wip, you need to manually choose the request review action from the bottom of the revision. If you are the author, it should be in the action list. BMO will not move it out of draft state anymore if --wip is used. Does this work?

It works, but it's pretty unfortunate. As I mentioned here, if I have a 20-patch patch series that I've posted as a WIP and now want to post for review, I really don't want to perform 20 manual UI actions. I should just be able to re-submit without the --wip flag and that should result in the patches being posted for review.

Comment 10

[David Lawrence [:dkl]]

(In reply to Botond Ballo [:botond] from comment #9)

(In reply to David Lawrence [:dkl] from comment #7)

With the latest changes, now when you submit with --wip, you need to manually choose the request review action from the bottom of the revision. If you are the author, it should be in the action list. BMO will not move it out of draft state anymore if --wip is used. Does this work?

It works, but it's pretty unfortunate. As I mentioned here, if I have a 20-patch patch series that I've posted as a WIP and now want to post for review, I really don't want to perform 20 manual UI actions. I should just be able to re-submit without the --wip flag and that should result in the patches being posted for review.

Ah yeah. Sorry I did not remember your earlier comment. And yes it would not be ideal if you have a large stack of revisions. glob, is this something that could be done with moz-phab without the --wip option on the same stack? Is this something in the works?

Comment 11

[:glob ✱]

this is a bug in moz-phab; after submitting without --wip the revision should be in a reviewable state.
filed as bug 1564166.