Meteor 1.5 with code splitting is coming! I’m very excited about that. We also talked about the Meteor team’s longer-term vision, and about migrating Atmosphere packages to npm.
The audio-only podcast version is available here and it should be available in your favourite podcast app and on iTunes very soon.
Exciting, code splitting is something I’m really looking forward to!
One aspect I wanted more clarity on. With code splitting, I saw mentioned in the GitHub issue that the code can be lazy-evaluated. Going through flame graphs of my Meteor apps, I can definitely see how this in itself offers a considerable performance benefit. Right now, my app bundle sits close to 2MB. Will this provide any benefit in that regard? Will each route (Using Blaze Layout with dynamic templates) download its own required files, rather than bundling them all together upfront? Or is that going to stay the same?
You can lazy evaluate modules starting in Meteor 1.3, so you should get to improving that right away!
Code splitting/dynamic imports is specifically about reducing the download size.
Would be awesome if you could provide some examples. Also, if I’m already using the imports style introduced with 1.3, will I need to change my code to get the reduced download size benefit with 1.5?
We’re not at the examples stage yet - this is still an early technical beta.
Hmm. How about code changes, are they necessary to gain benefit?
Yes - in the current release nothing changes in your app until you replace some import statements with dynamic import() calls. We think it’s important to control how and when different parts of the code are loaded.
How do you create flame graphs for your meteor apps?