That MR is entirely additive (it doesn't modify the existing jump behavior), so I don't know how it would break this, but this and that MR are highly related.
I'm interested to see the reproduction instructions for this, because it sounds like an edge case I don't know about relating to that MR above (which sets currentDiscussionId to null to force it to resolve to the first ID when triggering jumpToNext...
That said, I'm not sure how it could ever remainnull in that scenario (since it immediately jumps to a discussion)...
I think it happens when you are on Changes tab and MR has an unresolved discussion, but it is not related to the diff. Clicking the button generates the js error
This issue has regression without specifying which milestone introduced it. We're assuming that it's introduced in the current version 13.3.0-pre so we're labeling regression:13.3.
Please correct it accordingly and keep regression so that we could easily search for all the regressions across version, too.
@pedroms looking at this report... shouldn't we address this scenario with some UX tweak? Potentially we can detect that we do have some unresolved threads but not in the current tab. Should we display a warning asking the user whether they want to switch tabs? Any other ideas? :D
@andr3 if the action is broken, there are not many options other than:
Disable the button and explain to the user that they have to switch tabs
Hide the button when there are no threads in the active tab
Improve the behavior so that it switches tabs to navigate across the threads
The first two are “hacks” in that they are compensating for the unexpected behavior. If possible, we should address the root cause for this by tackling option 3.