[Continue Reading...]

I was just reading Ryan Niemeyer’s 2nd post about performance in knockout.js and I saw arrayPushAll. We ended up needing to use it for reasons other than performance, but performance does benefit greatly from its use.

We just had a problem (4/6/2012) where we were adding an array of items, one at a time in a loop, to an observable array. We wanted that notification of the array changing to trigger an ajax request, but we couldn’t because we were getting {n} notifications and we only wanted one request.

Initially I had to wire up the event in an odd way and that approach didn’t feel right since it seemed that we should be able to use the arrays changed notification. I should have thought to look up if ko had a slice or an equivalent, but I didn’t at the time. Now we just use arrayPushAll and this gives us a single notification because its a single change to the array. This is a much better solution and the added performance is great too!

I’ve been going over alot of knockout work lately and I’m new to 2.0. I’d seen material on dependent observables and then some stuff on computed observables and thought they were different things! I’d been meaning to google what their differences were when I ran across this on Knockoutjs:

What happened to dependent observables?

Prior to Knockout 2.0, computed observables were known as dependent observables. With version 2.0, we decided to rename ko.dependentObservable to ko.computed because it is easier to explain, say, and type. But don’t worry: this won’t break any existing code. At runtime, ko.dependentObservable refers to the same function instance as ko.computed — the two are equivalent.

That makes a lot of sense now! Yay one less thing to have to know. You should check out the API, its pretty useful. I’m going to try and write a small blurb about using a computed observable in a separate viewModel to read/write data from cookies.