[Bug]: Comment from second-level approver is not shown in the reviewer tools when canceling and re-enqueuing

[wagnerand]

What happened?

When a second-level approval cancel and pending decision and re-enqueues the add-on, the activity that is shown in the reviewer tools for this re-enqueuing does not show the comment the approver left as the reason for canceling/re-enqueuing.

STR:

1. Submit a listed (important because of this bug) version for a promoted add-on
2. As a reviewer, reject that version. The rejection is held in the second level approval queue.
3. As a second-level approver, use "Cancel and enqueue in Reviewer Tools"
4. Go back to the reviewer tools and inspect the event/activity "flagged for human review" event

What did you expect to happen?

The comment provided by the second-level reviewer should be shown as part of the activity.

Is there an existing issue for this?

• I have searched the existing issues

┆Issue is synchronized with this Jira Task

[ioanarusiczki]

The comment is available atm but it's under Held rejection

[wagnerand]

@ioanarusiczki Do you mind double-checking that? I cannot reproduce this on prod. The comment I left for the add-on I pushed back into the queue on prod yesterday does not show up anywhere on the reviewer tools page.

[ioanarusiczki]

@wagnerand you're right. After second ☕and trying on dev/stage -> comments from 2nd level approval queue are not visible indeed. That was the comment made when rejection was held 😅

[wagnerand]

Oh, it is shown as a DSA job further down.

That is super confusing. Can we make that not be a DSA job?

[diox]

The logic behind making it a job was to have the ability to link the decisions together: a decision is made, then overridden by 2nd level approval etc, and to do that you need a job.

However, according to @wagnerand we don't necessarily care about keeping that.

It could be a new NHR of some kind to deal with that.

[wagnerand]

The current behavior creates another issue that we should keep in mind and address as part of this as well:

1. Assume a reviewer rejects different versions of an add-on in separate review actions
2. The high-impact approver notices the mistake, cancels all decisions and asks the reviewer to submit a single rejection.
3. The reviewer issues a single rejection, but needs to select all decision check-boxes in the DSA section on the review page.
4. This seems to issue several separate decisions again, instead of a single one.

[ioanarusiczki]

I tried on -dev with some promoted addons and I see the comment from the 2nd level approval queue displayed in version's history. ✅

I've also tried a classic scenario with an abuse report attached -> where the decision is to proceed. Emails are sent to reporter, developer. ✅

There would be an issue if the same version is held for rejection/disable the 2nd time and it's canceled again -> in this case the comment is not displayed ❓

To give 2 examples:
https://reviewers.addons-dev.allizom.org/en-US/reviewers/review-listed/639781 -> the second time canceled with a decision

https://reviewers.addons-dev.allizom.org/en-US/reviewers/review/639798 -> with a 2nd decision

[ioanarusiczki]

Looking on stage - I see that the issue I described it reproducible when reviewers make the decisions on their own initiative.
With an abuse report attached it's not reproducible

abuse reporting
https://reviewers.addons.allizom.org/en-US/reviewers/review/2252708

own initiative
https://reviewers.addons.allizom.org/en-US/reviewers/review-listed/2252709

Sounds like an edge case - opening a followup.