Hi all,
I took apart the Horizon paging mechanism while working on [1] and have a
few of findings:
- Paging is unimplemented/turned off for many (if not most) panels, not just Routers and Networks
- Currently, single page data loads could potentially bump up against API hard limits
- Sorting is also broken in places where paging is enabled (Old images...), see [2]
- The Networks table loads data via three API calls due to neutron API limitations, which makes the marker based mechanism unusable
- There is at least one more minor bug which breaks pagination, there may be more
While some of these things may be fixable in different hacky and/or inefficient ways,
we already have Angular implementations which fix many of them and make improving
and fixing the rest easier.
Since Angular ports would help with other unrelated issues as well and allow us to
start deprecating old code, I was wondering two things:
1) What would it take to increase the priority of Angularization in general?
2) Can the Code Review process be modified/improved to increase the chance for
Angularization changes to be code reviewed and merged if they do happen?
My previous attempts in this area have failed because of lack of code reviewers...
Since full Angularization is still the goal for Horizon as far as I know, I'd rather
spend time doing that than hacking solutions to different problems in legacy code
which is slated deprecation.
Best Regards,
Marek
