1548448 - Workflow hole for "rebase a contributor's patch and land it"

Status: RESOLVED DUPLICATE of bug 1538151

(Conduit :: Phabricator, defect)

Description

[Botond Ballo [:botond]]

Today I approved a contributor's patch.

I tried to land it using Lando, but got a rebase error. So, I applied the patch and rebased it locally.

I then tried to update the patch using moz-phab submit, and got this message:

Updating revision D29084:
534061:8c38d57344ff Bug 1477832 -  Acquire AsyncPanZoomController::mRecursiveMutex in more places to make sure it's held whenever Metrics() is accessed.


 Exception 
The program is attempting to read user input, but stdin is being piped from some other source (not a TTY).
(Run with `--trace` for a full exception trace.)
    You don't own revision D29084: "Bug 1477832 -  Acquire
    AsyncPanZoomController::mRecursiveMutex in more places to make sure it's
    held whenever Metrics() is accessed.". Normally, you should only update
    revisions you own. You can "Commandeer" this revision from the web
    interface if you want to become the owner.

    Update this revision anyway? [y/N]
CalledProcessError: Command '['/home/botond/builds/arcanist/bin/arc', 'diff', '--base', 'arc:this', '--allow-untracked', '--no-amend', '--no-ansi', '--message-file', '/tmp/tmpeESra5', '--message', 'Revision updated.', '--update', '29084']' returned non-zero exit status 1

(By the way, there's another bug here: the error message looks like it's intending to let me choose to update the revision anyways by pressing "y", but I never get that chance -- I'm returned to my terminal prompt right away.)

To resolve this, I "commandeered" the revision in Phabricator, and ran moz-phab submit again. However, this had the effect that I was now the patch author and the contributor was the reviewer, and as a result I could not land the patch right away.

There should be a workflow for this use case with Phabricator that allows me to rebase a contributor's patch, and then update and land it right away.

Comment 1

[:glob ✱]

(By the way, there's another bug here: the error message looks like it's intending to let me choose to update the revision anyways by pressing "y", but I never get that chance -- I'm returned to my terminal prompt right away.)

unfortunately there's no way for us to easily bypass these prompts. we're actively working on removing our reliance on arc.

this looks a lot like bug 1538151; phabricator always assumes the user submitting the patch is the patch author instead of pulling the information from the commit's metadata.

Comment 2

[Botond Ballo [:botond]]

(In reply to Botond Ballo [:botond] from comment #0)

To resolve this, I "commandeered" the revision in Phabricator, and ran moz-phab submit again. However, this had the effect that I was now the patch author and the contributor was the reviewer, and as a result I could not land the patch right away.

Actually, not having L3 commit access, my contributor can't approve the patch either. So, I don't actually know how to proceed to get it landed, short of falling back to mozilla-inbound.