Modularization and Micro Front-Ends #1606
Replies: 2 comments 1 reply
-
In my experience, the burden of individually-versioned components is much higher for both maintainers and consumers. Instead of upgrading the library as a whole, we now need to consider what version(s) of Component A work with what version(s) of Component B, C, D, etc. This is hard for users, but it's also hard for us to keep track of as maintainers. It will also cause problems with the autoloader and auto-imported component dependencies. For example, button imports icon and spinner, so now we'll need a runtime check to make sure minimum versions are met for every component that has similar dependencies. Lastly, it would make support very challenging for us, as an infinite number of version combinations will lead to weird edge cases that are hard to reproduce. I went down that road years ago and I'd prefer not to do it again. |
Beta Was this translation helpful? Give feedback.
-
https://github.com/vaadin/web-components/releases also does library versioning. Anything else would be an insane workload that I think you simply wouldn't find customers to pay for. |
Beta Was this translation helpful? Give feedback.
-
Shoelace currently bundles web components in a single repository, necessitating full-package updates for any component changes. Would it be feasible to break down each component into micro web components and integrate them using micro front-end techniques? This approach would allow us to avoid full package updates and instead apply updates only to the relevant component packages.
Beta Was this translation helpful? Give feedback.
All reactions