Keep track of parsed array changes when parsing===false

[etimberg]

Resolves #9511

[kurkle]

I think this would remove the performance gains from parsing: false
The parsed array should equal the data array In that case and a push should not need to update anything. So I think the real bug is somewhere else.

[etimberg]

This would have a perf impact, though the items inside the array are not changed, just the outer array itself. We could drop the use of syncList when parsing is false and instead make the changes inline. It would mean all of the onData* functions would get a bit more complicated

[kurkle]

Would putting that inside an if be enough?

Chart.js/src/core/core.datasetController.js

Lines 933 to 935 in bc6ef16

	for (const [method, arg1, arg2] of me._syncList) {
	me[method](arg1, arg2);
	}

[etimberg]

I don't think that's right because then changes to the elements would never happen so a single push or pop would never change the number of elements. This specific case causes a problem because when applying syncList, the number of elements temporarily grows to 4 but it does so after the data array has shrunk back down to 3

[kurkle]

I don't think that's right because then changes to the elements would never happen so a single push or pop would never change the number of elements.

The changes would be applied at the end of _resyncElements, but new elements would always be at end etc, so I don't think that would be an acceptable solution.

Chart.js/src/core/core.datasetController.js

Lines 950 to 954 in bc6ef16

	if (numData > numMeta) {
	me._insertElements(numMeta, numData - numMeta, resetNewElements);
	} else if (numData < numMeta) {
	me._removeElements(numData, numMeta - numData);
	}

I did not realize the actual issue is the delayed multiple syncs that are already applied to the _parsed before _syncList is processed. The performance preserving alternative could be something like (partial):

[etimberg]

That solution looks good to me! I will implement that today

[kurkle]

That solution looks good to me! I will implement that today

I think you forgot about this :)

[etimberg]

haha I did, will definitely get to it today 😂

[kurkle]

I kind of hoped you'd pick a better name for the helper, but I don't have any suggestions :)