This PR includes an experimental change to evaluate the impact of #36, which proposes not rendering parts that update their own options.
First impressions are positive so far- the SlickGrid in particular seems more stable and performant. It seems like the wrapper was written for this execution model, which makes sense considering it's history in the enterprise platform.
We'll need to test a variety of UDPs and dashboards to ensure that this doesn't cause unworkable breakage, and going through the process of fixing UDPs that are broken by this change should tell us more about the real-world, practical impact of this.
Post-experiment To Do
After that's done, there's still some followup before this can be merged onto mainline:
[ ] Review Parts and UDPs
Various UDPs, and possibly a built-in part or three, rely on this behavior.
[ ] Review SlickGrid
This needs careful study:
the grid shouldn't be relying on this...
...but might be as a matter of course.
Review:
[ ] Broken Column formatting
[ ] Conditional formatting
[ ] Heatmaps/progress bars
[ ] Sparklines
[ ] Dashboard links
[ ] Change highlighting
[x] Tick Performance
No major regressions in Fund Summary tick-to-render time
(Appears to have actually improved, which wasn't expected)
This PR includes an experimental change to evaluate the impact of #36, which proposes not rendering parts that update their own options.
First impressions are positive so far- the SlickGrid in particular seems more stable and performant. It seems like the wrapper was written for this execution model, which makes sense considering it's history in the enterprise platform.
We'll need to test a variety of UDPs and dashboards to ensure that this doesn't cause unworkable breakage, and going through the process of fixing UDPs that are broken by this change should tell us more about the real-world, practical impact of this.
Post-experiment To Do
After that's done, there's still some followup before this can be merged onto mainline:
Will fix #36.
Known breakages: