Hi,
My simple app (angular2) is noticeable slower (loading?/rendering? not sure) after updating to 2.5.0. Anyone experiencing the same issue?
I'll try to put together a jsfiddle.
@activeperception please put together a fiddle. I don't think we've done anything to make it slower. Which version did you upgrade from?
Yes this is true, it even crashing browser. But it's okay after downgrading to 2.4.0
@frdteknikelektro do you have a test case we can debug?
Here, , , I think the problem is when you try to render Chart in parent that display style is none.
https://jsfiddle.net/frdteknikelektro/9pgs9qo9/ <-- Using 2.4
Try to change it to 2.5
Looks like what's happening is that the chart renders even though it has no size. That shouldn't happen
So what do you think? It's okay from until 2.5
Yes, I'm not sure what changed in 2.5 to change this. I'm looking through the history to see
@simonbrunel thoughts on this?
Seems to be caused by the time scale (loop over scaleSizeInUnits == 2674800000 at construction time - update). The chart has no size, but still need to be updated so I guess this issue needs to be fixed at the time scale level.
Maybe @tredston can have a look since the time scale implementation changes a lot in #3914 and might have fixed this issue as well.
Looks like me.width has changed in this example case from 0 to be negative, which results in the tickUnit staying on milliseconds instead of bumping up to larger intervals.
Unfortunately it looks like the negative width causes #3914 to get stuck in an infinite loop.
Edit: Seems like this change causes the y-axis to have a positive width, while before it was set to me.maxWidth = 0. This scale's width ends up making the chart's width negative when calculating the layout.
Seems indeed that this problem persists only when using the time scale . I had to search everywhere and remove all display:none props from the chart container in order to solve the freezing problem.
Resolved in #4105
Most helpful comment
Seems to be caused by the time scale (loop over
scaleSizeInUnits == 2674800000at construction time - update). The chart has no size, but still need to be updated so I guess this issue needs to be fixed at the time scale level.Maybe @tredston can have a look since the time scale implementation changes a lot in #3914 and might have fixed this issue as well.