Like, one server replying for all requests?

yep.

a wee baby server.

Is there any trick to debugging in production? miniflare is awesome for getting something running, but now that I publish i get totally useless "internal error" in the logs with no extra information

wrangler tail

has some support for Durable Objects, so depending on exactly what error you're getting, it may yield some insight.

Got it. The trick is to wrap everything in try/catch and manually dump it to the log. Grump.

When you find the error could you let me know what it was? Miniflare should probably throw a similar one... 👀

Yeah I have it. It is:

CompileError: WebAssembly.compile(): Wasm code generation disallowed by embedder

workers support wasm, but you have to upload it statically.

which is bummers for me.

I'm not sure how your impl works, but if you are embedding v8, the relevant api is: https://bugs.chromium.org/p/v8/issues/detail?id=6756

ah, got it, I probably just need to set

wasm: false

here: https://github.com/mrbbot/miniflare/blob/master/src/scripts.ts#L17 (https://nodejs.org/api/vm.html#vm_vm_createcontext_contextobject_options)

is there a way to know what pop i'm in? both for the outer worker and the DO ?

The latency is higher than I expect. I feel like when I tested this before I was getting absurdly low latency like 10ms, now its 100.

nevermind, found request.cf

For me, it's a great use case for creating an instance of a room where state and real-time magic needs to happen. For instance my online pictionary prototype skissan.com uses DO instance for each room, and a single DO instance for the lobby

@User https://latencies.codesurge.workers.dev/ping

^ this should tell you 'from' is the worker, 'to' is the durable object.

Anyone any tips to have a global setInterval in the durable object that I can actually stop? I get the following error:

Cannot clear a timeout created in a different request context.

Yeah thanks was just looking at this 😄

Sorry on my phone, couldn't link to it

No worries, thanks 😄

It's too bad we can't clear these though 🤔

I guess with an setInterval implementation using setTimeout, I can just stop the next setTimeout. I'll try that.

Yeah it sounds like it's something that will be fixed

That's what I do, yes.

Maybe I should add setInterval to my workaround gist? It appears to be quite popular 😅

Definitely 🙂

I attempted to look for it for fun, but no luck!