Optional review feature
Categories
(Conduit :: Phabricator, enhancement, P5)
Tracking
(Not tracked)
People
(Reporter: kvark, Unassigned)
Details
(Keywords: conduit-triaged, conduit-upstream)
For assigning reviewers today we have 2 options:
- regular reviewer. The problem here is that whenever there are 2+ reviewers, as soon as one of them gives r+, the whole change disappears from "Needs Review" section of other reviewers. This is very annoying.
- so devs try to use the blocking review, which enforces that each review is independent. The problem, however, is that sometimes a review should not be required.
Consider a situation where I want to potentially review some of the changes in a particular component. I can set up a rule to add myself as a reviewer to this component. If the rule makes me a regular reviewer, and I give r+, then I screw up other reviewers, as explained in (1). If the rule makes me a blocking reviewer, than I'd be unnecessarily blocking changes that only tangentially touch the component I'm interested in. Neither option works well for this.
Would you consider adding a review mode that is "optional"? That is, giving r+ would not make the change considered "reviewed" for the purpose of landing. At the same time, change would be landable without that review.
Updated•4 years ago
|
Comment 1•4 years ago
|
||
:kvark -- just wondering if in this situation, as a work around, you could just leave comments on the revision without changing the state of it. You could also time your approval so that it happens after other reviewers had a chance to approve. Not ideal, but works around interfering with the regular workflow.
Updated•4 years ago
|
Reporter | ||
Comment 2•4 years ago
|
||
I guess that would work, too. What happens, generally, is that if you CC somebody, they may not necessarily pay attention. The CC, conceptually, is inviting to observe, not act.
Description
•