Incorrectly showing entries for "To Check In" when bugs were reopened

[victorporof]

I have "To Check In 3" bugs, all of them oranges which had a shot-in-the-dark patch at some point. The patches were actually landed, however the bugs got reopened since the oranges persisted. I'm not assigned to them anymore, but they still appear in that tab.

[harthur]

Hm, this is a tough one. This query for this list is non-obsolete patches by you, with an r+, on bugs that are new, unconfirmed, reopened, or assigned that also don't contain the word "fixed" in the whiteboard.

I can see a case where a reopened bug contains a patch with an r+ that needs checkin. So I don't think we can filter out reopened bugs.

Maybe an option to hide an item and the decision is saved in localStorage?

[victorporof]

I initially thought that an easy fix would be changing the query to force "bug status == assigned", but I guess that having a r+ed patch in a bug (that's not specifically marked as assigned) doesn't necessarily mean that I don't need to check it in. People don't always change the bug status from * to ASSIGNED.

Just having the ability to hide an item would be superb.

[mounirlamouri]

If you use "checkin-" (which intent is to be used on backed out patches), we could filter out those patches. Actually, I just sent a pull request to filter out "checkin+" patches, it would be easy to extend this to checkin{+,-}.

[campd]

Maybe you could filter out REOPENED bugs whose r+'ed patch is from before the REOPEN?

[Mossop]

This doesn't just happen for reopened bugs though, sometimes you land the patch and wait and see if it fixed it before closing. Maybe filter out bugs who don't have you as the assignee?

[Mossop]

Or check for hg links in the mail since the patch was reviewed?

[edmorley]

I guess the other options are to either obsolete the patch (not ideal, since it did land), or perhaps best of all: just clear the review+. If the bug isn't assigned to you, then it won't show up at all in Bugzilla-todos.