When discussing the merits of AngularJS, two-way data binding is often touted as a major benefit of Angular over other JS frameworks. Digging deeper, the documentation suggests this process is done through dirty-checking rather than through event-driven measures. At first, it seems that the digest-loop works by having a method fire off in the background at periodic intervals, checking all the $watch
es during each cycle. However, reading further, it seems that the digest-loop is actually triggered by rootScope.digest()
, which in turn is triggered by $.apply
, which is in turn triggered by an event(!), such as an onClick
event called through ng-click
.
But, how can this be? I thought Angular doesn't use change listeners. So how does the digest-loop really operate? Does Angular automatically kick-off the digest-loop internally, or is the digest-loop triggered by events? If the digest-loop is run automatically, how often does it run?
Some clarification points:
- I'm not asking about how the digest-loop runs when manually binding to changes. In this case, if you want to force a digest loop, you can do so by calling
$.apply()
- I'm also not asking about how often digest loop runs in response to user events. For example, if ng-model is on an input box, Angular will kick-off a digest loop when a user starts typing. The confusing part is that in order to know a user was typing, doesn't Angular use an event-based
onKeyUp
somewhere? - I already know that there is a limit of 10 cycles max per digest-loop. My question is less about the number of cycles per digest-loop, but rather the number of digest-loops that run, say, per second.
- Bonus questions: How does the digest-loop relate to the JavaScript event-loop? Does the JS event loop run periodically in the background? Is the digest-loop the same thing as the event loop, but only in the "Angular Context"? Are these totally different ideas?