#157 New repo request rejected because of bugzilla assignment?????
Closed: Invalid 5 years ago by kevin. Opened 5 years ago by gbcox.

I'm following the package review process here:
https://fedoraproject.org/wiki/Package_Review_Process

Here is the bugzilla ticket:
https://bugzilla.redhat.com/show_bug.cgi?id=1758925

The package passed review. The next step says:
When your package pass the review, you should use the fedpkg tool to request a git repository for it. Before doing that you will need a pagure_api_token configured (one with "Create a new ticket" ACL) and added into ~/.config/rpkg/fedpkg.conf

I did that and created the pagure tickets. Then the tickets were closed as invalid because the bugzilla ticket is not assigned? I didn't see in the procedure where I needed to do something else to the bugzilla ticket.

Please advise.


It's the reviewers responsiblity... in the reviewer section:

"if you want to formally review the package, set the fedora-review flag to ? and assign the bug to yourself."

Please just ask in the bug for the reviewer to assign the bug to themselves, or just do so yourself...

This is so the tool that adds packages can make sure the person who reviewed the package is in the packager group.

Metadata Update from @kevin:
- Issue close_status updated to: Invalid
- Issue status updated to: Closed (was: Open)

5 years ago

Kevin, I don't understand your response.

I'm creating new packages. The package was reviewed by another person. That person indicated that it was approved and set the flag.

The procedure says after the flag has been set and the package has been approved for you to request via fedpkg for the pagure tickets be created.

The tickets were created and then rejected as invalid because the bugzilla ticket after it was approved by the reviewer was left in an unassigned state.

No where in the process does it say that after the package has passed review does something else need to be done to the bugzilla ticket.

Metadata Update from @gbcox:
- Issue status updated to: Open (was: Closed)

5 years ago

So, I assigned the ticket to myself and resubmitted. I then got a response that the assignee has to match the reviewer - so I re-assigned the ticket to the person who did the review, resubmitted and it worked.

The documentation needs to be revised because it does not reflect this requirement. First of all, the reviewer isn't the one that is tasked to create the git repo via fedpkg - that according to the procedure is me. When the reviewer approves the package he is no longer assigned the ticket, he should be done and if the ticket were need to be re-assigned, it would go back to the initiator if anybody.

I was quoting from: https://fedoraproject.org/wiki/Package_Review_Process

🔗 Reviewer

The Reviewer is the person who chooses to review a package.
Note.png
fedora-review tool
fedora-review is a very useful tool for handling some grunt work in the review process and it is highly recommended that you take advantage of it. dnf install fedora-review and refer to the man page for more details. Note, however that it is not a replacement for human input and you still need to understand the Packaging Guidelines thoroughly.
Note.png
Comments by other people
Other people are encouraged to comment on the review request as well. Especially people searching for sponsorship should comment other review requests to show, that they know the Packaging Guidelines.

The Reviewer can be any Fedora account holder who is a member of the packager group. (If the Contributor is not yet sponsored, the review can still proceed to completion but they will need to find a sponsor at some point.)

    Search for a review request that needs a reviewer: http://fedoraproject.org/PackageReviewStatus/ (fedora-review flag is blank or the bug is assigned to nobody@fedoraproject.org)
    If you notice some issues that need to be solved before you want to start a formal review, add these issues in a comment and set the Whiteboard of the bug to contain NotReady. This helps other possible reviewers to notice that the review request is not yet ready for further review action.
    if you want to formally review the package, set the fedora-review flag to ? and assign the bug to yourself.

Note the last sentence there. The reviewer is the one who assigns the bug to themselves. If they approve the package, then they are still the reviewer, the one who reviewed the package. The tool that processes new package requests checks this to make sure a valid packager reviewed your package and not just someone random.

I'm happy to see the documentation improved, but not sure how in this case. In any event, it's a wiki, so feel free to edit the page and try and improve it...

Kevin,

I'm not the reviewer. According to the process, I am the Contributor. I created the package and asked it be reviewed. It was then reviewed by another person and approved. When that was completed, I the "contributor" was tasked by the process to "When your package pass the review, you should use the fedpkg tool to request a git repository for it. "

When I did that, my request was rejected TWICE. First I was told that I as the "contributor", needed to assign the bugzilla request to myself. I did, resubmitted and then the request was rejected again. I was then told that I need to re-assign the bugzilla ticket to the "REVIEWERS" email address.

Why is the system looking at the "Assigned" field to match the person who completed the "REVIEW" flag. You know who reviewed the the package... their name and email address is associated with the REVIEW flag. What value does it add to require the current ASSIGNMENT of the ticket to match the REVIEW flag. If anything, I would think at the time the git was requested it would want the ticket to be assigned to the "CONTRIBUTOR" so you could check that address to make sure that person was a member of the packagers group. At least that would make some sense.

And from the process:

"Once a package is flagged as fedora-review + (or -), the Reviewer's job is done although they may be called upon to assist the Contributor with the import/build/update process and to ensure that the Contributor closes the ticket out when the process is complete."

So at that point, as the process says, the "Reviewers" job is done. Why is the system insisting that the ticket remain with the "Reviewer".

Kevin,
I'm not the reviewer. According to the process, I am the Contributor. I created the package and asked it be reviewed. It was then reviewed by another person and approved. When that was completed, I the "contributor" was tasked by the process to "When your package pass the review, you should use the fedpkg tool to request a git repository for it. "

Right

When I did that, my request was rejected TWICE. First I was told that I as the "contributor", needed to assign the bugzilla request to myself. I did, resubmitted and then the request was rejected again. I was then told that I need to re-assign the bugzilla ticket to the "REVIEWERS" email address.

Yes, it should have been your reviewer @rathann that assigned the bug to them when they started reviewing it. If they had, you would have not had any problems there.

Why is the system looking at the "Assigned" field to match the person who completed the "REVIEW" flag. You know who reviewed the the package... their name and email address is associated with the REVIEW flag. What value does it add to require the current ASSIGNMENT of the ticket to match the REVIEW flag. If anything, I would think at the time the git was requested it would want the ticket to be assigned to the "CONTRIBUTOR" so you could check that address to make sure that person was a member of the packagers group. At least that would make some sense.

It's easier to check assigned and submitted than who set a flag?

In any case, you're welcome to improve the docs if you like.

This project is the upstream for the 'fedora-packager' package, and has nothing itself to do with the review process. Changes there would likely be discussed on the devel list or in a fesco ticket.

Metadata Update from @kevin:
- Issue close_status updated to: Invalid
- Issue status updated to: Closed (was: Open)

5 years ago

Kevin,

Thanks for taking the time to reply. I do believe it would make more sense to check the flag rather than to who currently has the assignment of the bug. IMO the assignment reflects who currently has responsibility for taking further action. The reviewer is out of the picture as soon as he sets the flag - and the assignment field should then reflect the "contributor" - since that person is the one who finishes the process.

If you're saying that there are technical limitations that prevent you from using the flag field, then it is what it is and I suppose we are stuck with the current implementation.

I am more than happy to make modifications to documentation - but at this point it appears to be a quirk in the system where you are relying on the "assignment" field rather than the information contained in the "flag".

Log in to comment on this ticket.

Metadata