Adapter architecture

[fjorgemota]

Hey.

First, thanks for the project, it seems to work very well!

Second, and more related to the issues, I have a suggestion for this project: What about move this project from the vanilla to the adapter architecture, as suggested by MDC-Web.

I know that preact is pretty focused on DOM/browsers, but this change would allow a few things I think are cool for the future and maintainability of the project: With the MDC-Web migrating to Typescript in the near future, this change would allow to detect breaking changes easily when upgrading MDC-Web components.

To finalize, there are a few other advantages too, as seen on this discussion on the RMWC project.

What do you think?

[cromefire]

I had in mind to try that in 2.0, but had very little time. I will try to migrate them to that architecture when we start migrating the components

I know that preact is pretty focused on DOM/browsers

It could help with prerendering

[prateekbh]

Let's check it out.

The only thing I ask skeptical about is that if we should undertake the task of rewrite plus reach together.

Preferably I'd want to do re-arch in 3.0 instead

[cromefire]

I mean, it not a breaking-change we can do it in 2.1

[fjorgemota]

Hey.

I throught about suggesting the re-arch to migrate to 3.0, but...to what version of the MDC-Web this project will migrate in 2.0? 0.44? depending on the quantity of the breaking changes we will need to fix, I think it would be useful to already migrate to the adapter interface, which I think should be quite easy to do with the Preact features.

Also, as @cromefire noted, it's not a breaking change: There shouldn't be any user-facing breaking changes, because the adapter is basically internal to the DOM changes of the component...

IMO, my suggestion is: Migrate a few components to experiment, and then decide what todo with the next components. I think the adapter architecture will really help on the long-term, and will avoid a ton of bugs in the future (by using Typescript), and will really help with cool things like pre-rendering and, who knows? better performance. :)

[prateekbh]

Well my only reason to suggest 3.0 was not to block the community from getting 2.0 pmc, may be we should do it in 2.1.0 or something or may be a gradual shift if components one by one

[fjorgemota]

Oh, I see. I wouldn't like to delay or even block the community from getting PMC 2.0, because of this I suggested a gradual shift of components. It shouldn't be that harder according to MDC-Web documentation anyway.

[prateekbh]

Yeah so let's just do them one by one after 2.0

[cromefire]

But I would still like to try it on one or two components

[prateekbh]

Sure, let's do it