Update worker-deploy.yml

[Keyrxng]

relates to ubiquity-os/plugin-sdk#34

[gentlementlegen]

@Keyrxng I just realized something in these PRs: it won't work to generate the link. Look at an example I tried that would work:
https://github.com/ubiquity-os-marketplace/command-query/pull/18/files#diff-7733aa903e17a9c146f50a2c0397966c6767a5af4a27fdd6c2105baab37c9429R36

[Keyrxng]

My implementation works without the need to add the worker name as seen in the QA for the PR that was closed

your-plugin-name is the default name that comes with plugin-template

gentlementlegen/ubiquity-os-kernel#1 (comment)

---

I pull the plugin name from the url of the worker itself within the fetch

  const workerUrl = new URL(inputs.ref).origin;
    let workerName;

    if (workerUrl.includes("localhost")) {
      workerName = "localhost";
    } else {
      workerName = `${workerUrl.split("//")[1].split(".")[0]}`;
    }

Is this what you meant, sorry, I think now maybe I misunderstand what you were pointing out.

[Keyrxng]

@gentlementlegen the task (ubiquity-os/plugin-sdk#34) was dropped from P3 to P1 which is why I focused on higher priority tasks but I think they have been pretty much more or less cleared or are in the process of.

Perhaps the task priority can be re-raised and I'll focus on it again?

Also regarding the callback builder PR that was closed, likewise if you could label the task I'd be happy to jump back into SDK work

[gentlementlegen]

If that works to build the URL pointing to the Cloudflare Dashboard then I am fine with it. I just remembered that I had issues building the URL and I thought about it afterwards.

[Keyrxng]

If that works to build the URL pointing to the Cloudflare Dashboard then I am fine with it. I just remembered that I had issues building the URL and I thought about it afterwards.

Yeah that works just fine mate inputs.ref from the kernel is always going to be the worker URL and our deploy workflow uses the plugin name, so name extraction is simple so long as our deploy flow doesn't change.

Deploy injects ID, so it's available in the worker env when deployed but not locally but that's impossible anyway so I just name it localhost instead.