doesn't work due to CORB now

[Tenari]

i tried to spin up a new project. To the best of my knowledge it is configured correctly.

I edited the +cors-registry as the README dictates. still get this (I am logged in to the ship locally as well)

relevant logs:

tenari@dzs-MacBook-Pro ui % npm run dev

> tasks@0.0.0 dev
> vite

http://127.0.0.1

  VITE v4.2.1  ready in 188 ms

  ➜  Local:   http://localhost:5173/apps/tasks/
  ➜  Network: use --host to expose
  ➜  press h to show help

> +cors-registry
[ requests=[n=~~http~3a.~2f.~2f.localhost l={} r={~~http~3a.~2f.~2f.127~.0~.0~.1}]
  approved=[n=~~http~3a.~2f.~2f.localhost~3a.5173 l={~~http~3a.~2f.~2f.127.0.0.1~3a.5173 ~~http~3a.~2f.~2f.localhost~3a.3000} r={}]
  rejected=~
]

when I try running my browser with security turned off, I get a slightly different error:

it seems to me that something with how vite is passing along our urbauth cookie is not working

[Tenari]

also appears to be a vite version thing, since I don't have this issue on an older project using vite version v2.9.15

[matildepark]

What happens if you go direct to localhost:5173?

[Tenari]

takes me to the grid of urbit apps, since I am logged in.

[matildepark]

Sorry, I was on a plane. Okay, so presently, I can't replicate the issue described. I will write exactly what I did.

This is on an M1 MacBook Pro. Version 111.0.5563.146 of Chrome (latest), urbit 2.1, and create-landscape-app 8.1.0.

1. I made sure my urbit was the latest, redownloading it from urbit.org. It's currently 2.1.
2. I ran ./urbit -F zod.
3. I let it compile and continued.
4. I ran npx clear-npx-cache just in case I had old copies of the create-landscape-app wizard lying around. Then I ran npx @urbit/create-landscape-app.
5. For the wizard, I set my app to tasks and my ship location to http://localhost:80.
6. I saw the ship was finished, so I went to the generated folder, cd ui and npm i && npm run dev. It pointed me to http://127.0.0.1:5173/apps/tasks/. I went there, and saw the page mounted, but no apps, since I hadn't signed in yet.
7. I went direct to http://127.0.0.1:5173/ and signed in.
8. I went back to http://127.0.0.1:5173/apps/tasks/ and the apps were there, no CORB issue.

[ajlamarc]

Sorry, I was on a plane. Okay, so presently, I can't replicate the issue described. I will write exactly what I did.

This is on an M1 MacBook Pro. Version 111.0.5563.146 of Chrome (latest), urbit 2.1, and create-landscape-app 8.1.0.

1. I made sure my urbit was the latest, redownloading it from urbit.org. It's currently 2.1.

2. I ran `./urbit -F zod`.

3. I let it compile and continued.

4. I ran `npx clear-npx-cache` just in case I had old copies of the create-landscape-app wizard lying around. Then I ran `npx @urbit/create-landscape-app`.

5. For the wizard, I set my app to `tasks` and my ship location to `http://localhost:80`.

6. I saw the ship was finished, so I went to the generated folder, `cd ui` and `npm i && npm run dev`. It pointed me to `http://127.0.0.1:5173/apps/tasks/`. I went there, and saw the page mounted, but no apps, since I hadn't signed in yet.

7. I went direct to `http://127.0.0.1:5173/` and signed in.

8. I went back to `http://127.0.0.1:5173/apps/tasks/` and the apps were there, no CORB issue.

Just ran through these instructions and got the same result (no issues).

[elizasviel]

Ran through these steps and on step 6 I get stuck

12:12:25 AM [vite] http proxy error at /~/scry/docket/charges.json:
Error: connect ECONNREFUSED ::1:80
at TCPConnectWrap.afterConnect [as oncomplete] (node:net:1300:16)

http://localhost:5173/ gives me a 500 error

[matildepark]

Are you running your ship on localhost:80? It is often :8080 on Linux.

[elizasviel]

Yes, confirmed, when I boot ship, it prints: http: web interface live on http://localhost:80

[elizasviel]

quick update: solved the issue; for some reason changing the .env to http://127.0.0.1:80 instead of http://localhost:80 did the trick. I thought these were the same thing but apparently not

[gmcz]

I can confirm that I also ran into this issue on Linux, starting from a complete blank-slate.

for some reason changing the .env to http://127.0.0.1:80 instead of http://localhost:80 did the trick

Being on linux, I had http://localhost:8080 and can confirm that switching to http://127.0.0.1:8080 resolved the issue for me as well.