cosbgn Posts - Answer Overflow

•Created by cosbgn on 7/4/2024 in #❓・help

Is there an easy way to get `event`?

I use drizzle and cloudflare so I have an util like this one:

export const useDb = (event) => drizzle(event.context.cloudflare.env.DB, { schema })

export const useDb = (event) => drizzle(event.context.cloudflare.env.DB, { schema })

My issue is that I need always event and it's annoying to pass it around betwen functions. I know there is useEvent() but it requires an experimental asyncContext which is not available on cloudflare. How is everyone solving this?

3 replies

•Created by cosbgn on 6/1/2024 in #❓・help

Props based on route generate hydration missmatch

// /pages/tools/[tool_slug].vue
<template>
    <div>
        <tool :config="tools[tool_slug]" />
    </div>
</template>
<script setup>
import tools from '~/assets/templates/en.json'
const route = useRoute()
const { tool_slug } = route.params
if (!tools?.[tool_slug]?.title) {
    throw createError({ statusCode:404, statusMessage:"Tool not found" })
}
</script>

// /pages/tools/[tool_slug].vue
<template>
    <div>
        <tool :config="tools[tool_slug]" />
    </div>
</template>
<script setup>
import tools from '~/assets/templates/en.json'
const route = useRoute()
const { tool_slug } = route.params
if (!tools?.[tool_slug]?.title) {
    throw createError({ statusCode:404, statusMessage:"Tool not found" })
}
</script>

I want to do something like the above. My issue is that I get a missmatch because confi is empty on server and filled in client. Is there a way to have it filled also on server?

8 replies

•Created by cosbgn on 5/24/2024 in #❓・help

get `event` in `/utils/db.js` - Can I be done?

I would like to do something like this:

// utils/db.js
import { drizzle } from 'drizzle-orm/d1';
import * as schema_file from "../db/schema.js";

let cached_db = null

export const useDb = () => {
    if (!cached_db) { 
        const event = useRequestEvent()
        cached_db = drizzle(event.context.cloudflare.env.D1_PROD_BINDING, { schema: schema_file }) 
    }
    return cached_db
}

export const db = useDb()
export const schema = schema_file

// utils/db.js
import { drizzle } from 'drizzle-orm/d1';
import * as schema_file from "../db/schema.js";

let cached_db = null

export const useDb = () => {
    if (!cached_db) { 
        const event = useRequestEvent()
        cached_db = drizzle(event.context.cloudflare.env.D1_PROD_BINDING, { schema: schema_file }) 
    }
    return cached_db
}

export const db = useDb()
export const schema = schema_file

So then in any route I can do db.select() - However useRequestEvent() is not defined somehow? I know I can simply pass const db = useDb(event) on top of each route, I just would like to avoid passing event on every route and hopefully get better types. Should this instead be a plugin? What's the best practice here?

2 replies

•Created by cosbgn on 5/3/2024 in #❓・help

Did anyone manage to handle CORS with h3/nuxt/nitro?

I'm trying everything like this:

export default defineEventHandler( async (event) => {
    const cors_headers = { 
        origin: '*', 
        preflight: { statusCode: 204 }, 
        methods: '*' 
    }
    await handleCors(event, cors_headers)
    await appendCorsHeaders(event, cors_headers)
    await appendResponseHeaders(event, {
        "Access-Control-Allow-Origin": "*",
        "Access-Control-Allow-Methods": "GET, HEAD, POST, OPTIONS",
        "Access-Control-Allow-Headers": "*"
    })
return "hello"
 })

export default defineEventHandler( async (event) => {
    const cors_headers = { 
        origin: '*', 
        preflight: { statusCode: 204 }, 
        methods: '*' 
    }
    await handleCors(event, cors_headers)
    await appendCorsHeaders(event, cors_headers)
    await appendResponseHeaders(event, {
        "Access-Control-Allow-Origin": "*",
        "Access-Control-Allow-Methods": "GET, HEAD, POST, OPTIONS",
        "Access-Control-Allow-Headers": "*"
    })
return "hello"
 })

I always get: request blocked: (Reason: CORS header ‘Access-Control-Allow-Origin’ missing).

15 replies

•Created by cosbgn on 4/18/2024 in #help

Cannot read properties of null (reading 'dialect')

await migrate(db, { migrationsFolder: "server/db/migrations" }) is failing. I'm using d1. What could be the issue?

1 replies

•Created by cosbgn on 4/10/2024 in #pages-help

Mailchannels: sender is not authorized

I'm getting:

0    "Failed to send email: 550 5.7.1 This sender is not authorized to send from my-site.com. See https://bit.ly/domain-lockdown. cfid=my-site-v3.pages.dev"

0    "Failed to send email: 550 5.7.1 This sender is not authorized to send from my-site.com. See https://bit.ly/domain-lockdown. cfid=my-site-v3.pages.dev"

I have a TXT record with: name: _mailchannels
value: =mc1 cfid=my-site-v3.pages.dev (copied as is from error message) --- Do I need to just wait for DNS to catch up, or am I missing something?

6 replies

•Created by cosbgn on 4/5/2024 in #❓・help

error loading dynamically imported module

1 replies

•Created by cosbgn on 4/3/2024 in #help

How do you deal with migrations? on postbuild?

So, usually while developing locally, whenever I change my schema I run:

drizzle-kit generate:sqlite
// and then 
node ./server/db/migrate.js

drizzle-kit generate:sqlite
// and then 
node ./server/db/migrate.js

This first generates the migrations and then applies them to my local db. Once I deploy in my package.json postbuild script I have:

postbuild:"node ./server/db/migrate.js"

postbuild:"node ./server/db/migrate.js"

So if my build runs without issues, it will apply the migrations to my production DB. Is this correct?

1 replies

•Created by cosbgn on 3/13/2024 in #help

How do I mark "migration as applied" after using "push"?

I've been using push for a while, now my migrations don't work anymore because:

[ResponseError: error executing a request on the primary: table `actions` already exists in CREATE TABLE `actions`

[ResponseError: error executing a request on the primary: table `actions` already exists in CREATE TABLE `actions`

10 replies

•Created by cosbgn on 3/13/2024 in #help

returning with onConflictDoUpdate - How do I know if it was updated?

If I do:

await db.insert(users)
  .values({ id: 1, name: 'Dan' })
  .onConflictDoUpdate({ target: users.id, set: { name: 'John' } })

await db.insert(users)
  .values({ id: 1, name: 'Dan' })
  .onConflictDoUpdate({ target: users.id, set: { name: 'John' } })

Without returning I know if the row was added or edited, but I don't get the ID - With returning, I get the id but I don't know if it was created or just updated. Any way to know both?

1 replies

•Created by cosbgn on 2/27/2024 in #help

Do I need to cache `createClient`?

I use turso on the edge (CF Workers). I see a lot of implementations do something like this:

let cached_db = null
export const useDb = () => {
    if (!cached_db){
        const client = createClient({ url, authToken })
        cached_db = drizzle(client, { schema:schema_file })
    }
    return cached_db
}
// And then use it like:
const db = useDb()

let cached_db = null
export const useDb = () => {
    if (!cached_db){
        const client = createClient({ url, authToken })
        cached_db = drizzle(client, { schema:schema_file })
    }
    return cached_db
}
// And then use it like:
const db = useDb()

Is that the right approach? Or can I simply do:

  const client = createClient({ url, authToken })
  export const db = drizzle(client, { schema:schema_file })

  const client = createClient({ url, authToken })
  export const db = drizzle(client, { schema:schema_file })

Like can I call createClient() hundreds of times without issues or I should actually cache it?

1 replies

•Created by cosbgn on 1/25/2024 in #help

Why do I need to duplicate names in my schema?

For example:

name: text('name').notNull(),

name: text('name').notNull(),

Why do I need name twice? I can easily forget to update one or another creating a mess, like name: text('old_name') Any way to avoid it?

2 replies

DDeno

•Created by cosbgn on 1/22/2024 in #help

Subhosting API protection

Is there a way to setup custom auth via some header for example for all subhosting routes? I.e. Deno will throw an error is Authorization!=="foo"?

2 replies

•Created by cosbgn on 1/19/2024 in #❓・help

Do I need to use "lazy" with `useFetch/useAsyncData` when deploying to vercel?

So before I used to have a fully static fronted (nuxt-generate) and a backend API. Now with nuxt3 this got blurred. When I deploy to vercel (without changing any config), nuxt still builds the frontend fully static and backend is lambdas right? Meaning it's all generated at build time, which means I always have to use lazy right?

13 replies

•Created by cosbgn on 11/12/2023 in #workers-help

Rewrite traffic based on hostname

I'm building a site with custom domains. I would like to have my base domain (base-example.com) render normally, but IF hostname !== base-example.com I want to render the content of base-example.com/api/custom-content?host=some-hostname.com - How can I do that? I use workers, but could switch to pages if needed.

2 replies

•Created by cosbgn on 10/25/2023 in #pages-help

does cloudflare pages cache workers?

Can I be sure that my backend functions are not cached? They return different data depending on the logged in user.

1 replies

•Created by cosbgn on 10/21/2023 in #workers-help

How to set `logpush=true` without `wrangler.toml`

I use nuxt.js which automatically builds the app for cloudflare pages. This is great, however I do not have access to a wrangler.toml file. Is there any way to set logpush=true without a wrangler file?

2 replies

•Created by cosbgn on 8/30/2023 in #general-help

Best tool for long lasting background work? (e.g. 15 mins)

I have an app which runs on cloudflare pages. This means that my backend runs on workers on the edge, and have a 30 second timeout. Is there a tool in cloudflare, which allows me to run a specific worker/function for longer? like 15 minutes? It doesn't need to return anything, it needs just to compute something and save it in the database.

4 replies