You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
We noticed strong performance issues after upgrading from Prototype 1.7 to 1.7.1, and finally found that it was due to the new getUniqueElementID_IE function, which replaces getUniqueElementID by using the IE's uniqueID property.
It looks like a good idea at first sight except when updating an element containing thousands of elements, on which purgeElement is called, which calls stopObserving, etc finally accessing the uniqueID property which seems to be created when accessed, so it's a lot slower than a simple UID++.
Keeping the original code from getUniqueElementID changes nearly everything, by going drom 80s of purge to 1.2s.
I see no regression to do this, but would like to know if anybody noticed this problem before me.
The text was updated successfully, but these errors were encountered:
This is a tough one. We switched to uniqueId because giving an element an expando property in old IEs had the same performance overhead as adding an attribute to that element, and would (at least in some cases) trigger re-layout of that element. If we do anything about this I'd want to run some benchmarks on each approach.
Hello, thank you for your feedback, I made changes to this and got pretty good results without sacrifying the feature. If you want so I'll send you my modifications.
previous lighthouse ticket #3574
by Fabien Ménager
We noticed strong performance issues after upgrading from Prototype 1.7 to 1.7.1, and finally found that it was due to the new getUniqueElementID_IE function, which replaces getUniqueElementID by using the IE's uniqueID property.
It looks like a good idea at first sight except when updating an element containing thousands of elements, on which purgeElement is called, which calls stopObserving, etc finally accessing the uniqueID property which seems to be created when accessed, so it's a lot slower than a simple UID++.
Keeping the original code from getUniqueElementID changes nearly everything, by going drom 80s of purge to 1.2s.
I see no regression to do this, but would like to know if anybody noticed this problem before me.
The text was updated successfully, but these errors were encountered: