PLEASE READ - Future direction of Sapper

[benmccann]

We are currently working on the successor to Sapper called SvelteKit. It contains most of Sapper's client code and adds unbundled development via Vite for a better development experience. There is a migration path from Sapper to SvelteKit

We are not accepting new features to Sapper as a result of our focus on SvelteKit. However, SvelteKit has received many features not present in Sapper and we will continue to add new functionality there

[Mlocik97]

How long will be Sapper supported after SvelteKit release? Will repo / package exist after that, or will be removed at some point, or archived?

[benmccann]

We won't be removing the repo or package, but won't be updating them either.

[lovasoa]

Is sapper going to continue receiving bug fixes ?

[benmccann]

We won't be fixing any bugs. We possibly would consider merging bug fixes that others submit though I prefer time spent reviewing PRs be spent on SvelteKit.

[lovasoa]

So, for someone starting a project today, if sapper is not going to be maintained in the future and sveltekit is still in beta today, what would you recommend ?

[Mlocik97]

@lovasoa You can use Sapper now, and migrate it to SvelteKit later, migration will be relatively easy. https://kit.svelte.dev/migrating

except when You have really fat server-side, like something complicated on server, then migration will be problem, so then it depends on how important it is to have stability, bug-free etc,... if stability is critical, then sapper or wait (I think SvelteKit 1.0 is not that far in future, maybe month), if it's not that critical software, you can use SvelteKit already.

Btw. Sapper was already for some long time (year+) maintained the way "just fixing critical bugs", but it was already half-way "unsupported", so it's not like something changed now in sapper development,...

p.s. I think SvelteKit is already usable for some types of projects, and there are already some web apps that run on SvelteKit (even critical ones)... I myself am developing 2 web apps in SvelteKit, one non-critical and one critical, that will go to production. But it also depend on how much you know about JS and Svelte,... if You are not advanced developer, then use Sapper for now, because it will problem to fix some problems for You, that SvelteKit can have.

[benmccann]

I would generally recommend SvelteKit for new projects at this point. It's fairly far along towards getting to a 1.0 release. You can see what's been done vs what's left here: https://github.com/sveltejs/kit/milestone/2. If you need to launch quickly and need one of those things that's left to be done or need IE support which also hasn't been added yet then you may consider Sapper. SvelteKit may still have some API changes, but at this point that should be less work to adapt to than migrating from Sapper to SvelteKit

[MTyson]

"except when You have really fat server-side, like something complicated on server"

This. Is what I am concerned about.

When you assume serverless, that's great for deploying to serverless platforms, but you really ramp up the difficulty in doing server-oriented tasks. Like file upload.

Don't get me wrong, I root for sveltekit's success and look forward to using it, but I think the runtime adapter idea looks clean on paper but is gnarly in practice.

[benmccann]

SvelteKit supports but does not assume serverless deployment. Deploying to servers or generating static files for deployment are also first-class deployment targets. You can use adapter-node to deploy to a traditional server environment and that is possibly the most common way to deploy SvelteKit apps.

This thread is getting off-topic and probably isn't the right place to discuss questions about how to use SvelteKit given the myriad of different questions people might have