- https://hackmd.io/@mhevery/BkDUxaW84/https%3A%2F%2Fhackmd.io%2Fc%2FBkDUxaW84%2Fedit%3Fedit?type=book
- angular/angular#32103
- https://www.reddit.com/r/angular/comments/bxigkb/zonejs_performance_issues_in_angular_hybrid/
- https://github.com/angular/angular/pull/30533/files#diff-9d549aa403494fd043a146b8db64f503R264-R275
- https://stackblitz.com/edit/angular-zone-rxjs-patch-test?file=src%2Fpolyfills.ts
- https://blog.bitsrc.io/quantum-angular-maximizing-performance-by-removing-zone-e0eefe85b8d8
- https://medium.com/angular-in-depth/boosting-performance-of-angular-applications-with-manual-change-detection-42cb396110fb
- nearly no thoughts on when to update the view
- manual change detection
== TO ORGANIZE ======================
If we output an ES2017 bundle, we will have native async-await
which can't be monkey patched by zone.js.
The ~35-40 KB additional bundle. Additional runtime overhead on all asynchronous tasks.
NgZone: Every event handler triggers a change detection cycle, for example clicks to navigate or open a dialog or sidebar which doesn't change any state (other than URL and router state). Again, a major runtime overhead (performance hit).
If we combine this with the default change detection strategy (CheckAlways
), this comes a bigger and bigger issue as the application grows.
It's maintained by a single man (albeit a very talented man).
Event Coalescing The whole story is,
- there are some performance issue in Angular material, and this issue is discussed in the Angular team.
- Then Misko has an idea to use async change detection to improve the performance. And he write a design doc here https://hackmd.io/wXsFRwBJRSOUu03fdCPi1w
- Because this idea involves Zone and event listener, so Jia just volunteer to enhance the design and help to implement. angular/angular#29610
I made a typo in:
Change
comes
tobecomes
.