cosbgn Posts - Answer Overflow

•Created by cosbgn on 3/22/2025 in #help

joins overwrite select?

Am I crazy or this is not supposed to happen? My schema is:

export const pets = sqliteTable('pets', {
    id: text({ length:12 }).primaryKey().$default(() => nanoid()),
    owner_id: text().notNull().references(() => owners.id, { onDelete: 'cascade' }),
    name: text().notNull(),
})

export const owners = sqliteTable('owners', {
    id: text({ length:12 }).primaryKey().$default(() => nanoid()),
    name: text().notNull(),
})

export const pets = sqliteTable('pets', {
    id: text({ length:12 }).primaryKey().$default(() => nanoid()),
    owner_id: text().notNull().references(() => owners.id, { onDelete: 'cascade' }),
    name: text().notNull(),
})

export const owners = sqliteTable('owners', {
    id: text({ length:12 }).primaryKey().$default(() => nanoid()),
    name: text().notNull(),
})

When I do:

const pet_id = 122
const { ...pet_columns } = getTableColumns(pets) 
const [ pet ] = await db
    .select({ ...pet_columns, owner_name: owners.name })
    .from(pets)
    .innerJoin(owners, eq(pets.owner_id, owners.id))
    .where( eq(pets.id, pet_id) )
    .limit(1)

const pet_id = 122
const { ...pet_columns } = getTableColumns(pets) 
const [ pet ] = await db
    .select({ ...pet_columns, owner_name: owners.name })
    .from(pets)
    .innerJoin(owners, eq(pets.owner_id, owners.id))
    .where( eq(pets.id, pet_id) )
    .limit(1)

Data get's mixup, owner_id is assigned to pet id and also the names

1 replies

•Created by cosbgn on 3/5/2025 in #❓・help

routeRules help

Can someone help me with this:

routeRules: { 
        '/chat/:chat_id': { proxy: '/api/v1/:chat_id/public' },
    },

routeRules: { 
        '/chat/:chat_id': { proxy: '/api/v1/:chat_id/public' },
    },

it doesn't work and I can't find any documentation anywhere

7 replies

•Created by cosbgn on 3/1/2025 in #❓・help

Does anyone knows how to set CORS for static files (/public)?

I know with nitro, but I'm talking about files in the public folder. I tried with routeRules but they don't seem to apply to static files.

4 replies

•Created by cosbgn on 2/5/2025 in #workers-help

Why CF doesn't cache this worker?

Can someone help me figure out why this worker never gets cached: https://oparse.com/api/cache The response correctly sets cache-control max-age=60, stale-while-revalidate=60 (you can see it in the browser inspector) Yet from both logs ad DR flare extension I can see the response is never cached. What am I doing wrong? The worker is:

export default defineEventHandler(async (event) => {
  const now = Date.now()
  const cacheHeaders = "max-age=60, stale-while-revalidate=60"
  setResponseHeader(event, "Cache-Control", cacheHeaders)
  return `Timestamp is: ${now}`
})

export default defineEventHandler(async (event) => {
  const now = Date.now()
  const cacheHeaders = "max-age=60, stale-while-revalidate=60"
  setResponseHeader(event, "Cache-Control", cacheHeaders)
  return `Timestamp is: ${now}`
})

9 replies

•Created by cosbgn on 1/31/2025 in #❓・help

Can't start nuxt: Transform failed with 1 error: <stdin>:11:3: ERROR: Unexpected ","

I tried everything and I can't get nuxt to start after upgrading to 3.15.4 I tried also with:

  "overrides": {
    "vite": "5.4.11"
  },

  "overrides": {
    "vite": "5.4.11"
  },

Does anyone has any suggestion?

8 replies

•Created by cosbgn on 12/29/2024 in #workers-help

email worker `send to a worker` - can I select a specific path?

I have a worker which handles a lot of things, it's a hono api backend. Can I send email to a specific endpoint?

3 replies

•Created by cosbgn on 12/19/2024 in #workers-help

How to cache workers with static assets (beta)?

10 replies

•Created by cosbgn on 12/9/2024 in #help

D1 Remote in localhost?

Has anyone figured out how to connect to the production d1 DB in localhost? in wrangler is enough to add --remote

2 replies

•Created by cosbgn on 10/9/2024 in #❓・help

Please help me understand why cloudflare can't connect to postgres on nuxt

So postgres works on cloudflare workers, but with nuxt it can't connect. My guess is that it has to do with sockets. Is there a way to debug better this error? This is a quick reproduction: Nuxt: https://github.com/cosbgn/pg-cloudflare/blob/main/server/api/test.js or: https://github.com/cosbgn/pg-cloudflare/blob/main/server/api/psql.js DB_PASSWORD: Please DM (it's a demo PSQL) Try it locally with:

NITRO_PRESET=cloudflare_module npm run build

NITRO_PRESET=cloudflare_module npm run build

and:

npx wrangler dev --remote .output/server/index.mjs --site .output/public --log-level debug

npx wrangler dev --remote .output/server/index.mjs --site .output/public --log-level debug

With just nuxt dev it works perfectly, it's only on workers that it hungs and timesout. Any help is very appreciated!

1 replies

•Created by cosbgn on 10/8/2024 in #❓・help

Has anyone manage to connect nuxt & postgres on cloudflare?

My database always timesout on wrangler dev / or when deployed. I believe it's an issue on how nuxt deals with sockets. Did anyone manage to connect postgres?

1 replies

•Created by cosbgn on 10/5/2024 in #pages-help

Postgres timesout on cloudflare:

I have a PG db at digital ocean, it works perfectly in my local machine, however when deployed on cf I get: write CONNECT_TIMEOUT localhost:5432 which is weird as the port is not even 5432. This is a reprodution: https://github.com/cosbgn/pg-cloudflare/blob/main/server/api/test.js This is where it times out: https://pg-cloudflare.pages.dev/api/test Please DB me for the DB connection string and I'll share that. If someone has any idea on why it doesn't work on CF I would love it. P.s. IP are unlocked on all location.

14 replies

•Created by cosbgn on 10/1/2024 in #❓・help

Best nitro hook to kill a DB connection?

I would like to use hyperdrive which from their docs requires to clean up the client using waitUntil:

    const sql = postgres(env.HYPERDRIVE.connectionString);
    const result = await sql`select * from pg_tables LIMIT 10`;
    // Clean up the client, ensuring we don't kill the worker before that is
    // completed.
    ctx.waitUntil(sql.end());

    const sql = postgres(env.HYPERDRIVE.connectionString);
    const result = await sql`select * from pg_tables LIMIT 10`;
    // Clean up the client, ensuring we don't kill the worker before that is
    // completed.
    ctx.waitUntil(sql.end());

I would like to avoid doing this in each route, so I'm thinking a plugin is the right wait, should I use afterResponse? Something like;

  nitroApp.hooks.hook("afterResponse", (event, { body }) => {
    // SQL is auto-exported from server/utils
    event.waitUntil(sql.end());
  });

  nitroApp.hooks.hook("afterResponse", (event, { body }) => {
    // SQL is auto-exported from server/utils
    event.waitUntil(sql.end());
  });

1 replies

•Created by cosbgn on 9/26/2024 in #pages-help

All my .env variables seem suddenly undefined. Did something change internally?

22 replies

•Created by cosbgn on 9/21/2024 in #❓・help

How would you store a ref of an external packages in a state (useState)

I have this:

import { useChat } from '@ai-sdk/vue';
const { messages, handleSubmit } = useChat()

import { useChat } from '@ai-sdk/vue';
const { messages, handleSubmit } = useChat()

handleSubmit is a method and messages is a ref(). How can I wrap this in useState to be able to call handleSubmit and access messages from any component and keep the state alive on route change?

7 replies

•Created by cosbgn on 9/10/2024 in #help

Any way to use D1 remote DB (production one) while locally?

On wrangler is enough to add --remote but not sure how it works on drizzle

1 replies

•Created by cosbgn on 9/5/2024 in #❓・help

Do I need to rename all template refs from `ref` to `templateRef` now with vue 3.5?

Is this needed?

2 replies

•Created by cosbgn on 9/4/2024 in #help

generated columns reference another table

I have a table called users and another called teams. I would like to have users.plan be generated from teams.id === user.team_id and from that team get teams.plan. So pretty much user.plan is generated from the right team plan. Is this possible?

2 replies

•Created by cosbgn on 8/30/2024 in #❓・help

fetch full screen error on production

I use a custom fetch:

export const $api = $fetch.create({
    async onResponseError({ response }) {
        const possibilities = [response?._data?.statusMessage, response?.error?.data?.statusMessage, response?._data?.message, response.message, response?.statusText, "Unknown Server Error"]
        const withData = possibilities.filter(f => f?.length)
        const err_msg = `${response.status}: ${withData[0]}`
        if (import.meta.browser){
            toast(err_msg, "error")
        } else {
            console.error(err_msg)
        }
    }
})

export const $api = $fetch.create({
    async onResponseError({ response }) {
        const possibilities = [response?._data?.statusMessage, response?.error?.data?.statusMessage, response?._data?.message, response.message, response?.statusText, "Unknown Server Error"]
        const withData = possibilities.filter(f => f?.length)
        const err_msg = `${response.status}: ${withData[0]}`
        if (import.meta.browser){
            toast(err_msg, "error")
        } else {
            console.error(err_msg)
        }
    }
})

This works super well on dev, when an error shows it simply shows the toast. On production instead it shows the toast but also the full screen nuxt-error. How can I get that fixed?

1 replies