That's correct. Pages is full-domain.

IIsaac McFadyen That's correct. Pages is full-domain.

S

Spunkie•10/31/24, 9:36 PM

:weicry:

IIsaac McFadyen In general this is a good way to cache-bust. It seems "hacky" but it's not reall...

M

Mozzy•10/31/24, 9:37 PM

I know it's helpful, but I would like the option to have stricter cache rules

M

Mozzy•10/31/24, 9:37 PM

Maybe a rewrite middleware before the caching?

SSpunkie :weicry:

I

Isaac McFadyenOP•10/31/24, 9:37 PM

You could write a Pages Function to proxy some routes to another backend but that would increase latency and always bill you for Function invocations.

MMozzy Maybe a rewrite middleware before the caching?

I

Isaac McFadyenOP•10/31/24, 9:37 PM

Oh I might have misunderstood. You want to disable that so it still hits the cache?

M

Mozzy•10/31/24, 9:37 PM

Exactly

I

Isaac McFadyenOP•10/31/24, 9:37 PM

If so there's a setting you can change in the Cloudflare dashboard.

I

Isaac McFadyenOP•10/31/24, 9:37 PM

Sec.

I

Isaac McFadyenOP•10/31/24, 9:38 PM

https://dash.cloudflare.com/?to=/:account/zone/caching/configuration

I

Isaac McFadyenOP•10/31/24, 9:38 PM

Try that

I

Isaac McFadyenOP•10/31/24, 9:38 PM

Then change Caching Level to "Ignore Query String"

I

Isaac McFadyenOP•10/31/24, 9:38 PM

Keep in mind that will ignore all query string not just the incremental cache bust there.

IIsaac McFadyen You could write a Pages Function to proxy some routes to another backend but tha...

S

Spunkie•10/31/24, 9:39 PM

Ya I thought about that but it kind of defeats why I wanted to use pages. Which is mostly to make it easy to understand and update for non technical users.

S

Spunkie•10/31/24, 9:39 PM

But that is prob what I'll do :thisisfine:

I

Isaac McFadyenOP•10/31/24, 9:39 PM

You could also create a Cache Rule that ignores a specific query string value but I believe it requires a Business plan

M

Mozzy•10/31/24, 9:39 PM

Something I failed to show is that I do need to bust cache if other search params change

I

Isaac McFadyenOP•10/31/24, 9:39 PM

Yeah one sec then

M

Mozzy•10/31/24, 9:39 PM

ahh

I

Isaac McFadyenOP•10/31/24, 9:40 PM

Unfortunately it requires Enterprise to exclude specific query params actually.

M

Mozzy•10/31/24, 9:41 PM

would a rewrite middleware that only keeps certain params be enough then?

M

Mozzy•10/31/24, 9:41 PM

Never done rewrites, so not sure it works

I

Isaac McFadyenOP•10/31/24, 9:42 PM

I believe that would work? Always hard to know because different kinds of Rules run in different orders and there's no real reference that I've seen

IIsaac McFadyen I believe that would work? Always hard to know because different kinds of Rules ...

C

Chaika•10/31/24, 9:43 PM

https://developers.cloudflare.com/ruleset-engine/reference/phases-list/#request-phases

Cloudflare Docs

Phases list | Cloudflare Ruleset Engine docs

The following tables list the phases of Cloudflare products powered by the Ruleset Engine, in the order those phases are executed. Some products such as the Cloudflare Web Application Firewall have more than one associated phase.

CChaika https://developers.cloudflare.com/ruleset-engine/reference/phases-list/#request-...

I

Isaac McFadyenOP•10/31/24, 9:44 PM

Ah TY

C

Chaika•10/31/24, 9:44 PM

actual cache runs at the very end though, if the question is "can I use rules to modify what normal cache uses" pretty sure all of them work, transform headers at least def does

F

Flux•10/31/24, 10:17 PM

Wondering if anyone knows any ways to turn a PDF to a PNG/Image with workers. Is this even possible considering the v8 isolate constraints? (worst case setup lambda but dont want to)

CChaika Then you have to use CF For SaaS, which doesn't support wildcards full stop for ...

E

elpupper•10/31/24, 10:24 PM

ok so I'm doing some research and I dont think i get it. is CF for SaaS basically nginx? I cant use workers/worker routes here? Do I have to get my own server that cloudflare can point to?

MMozzy would a rewrite middleware that only keeps certain params be enough then?

M

Mozzy•10/31/24, 10:44 PM

This seems to do the trick, however cursed it is

export const stripUnusedSearchParams = createMiddleware<{ Bindings: CloudflareBindings }>(async (c, next) => {
  const url = new URL(c.req.url);
  let foundInvalidSearchParams = false;

  for (const key of url.searchParams.keys()) {
    if (!validSearchParams.includes(key)) {
      url.searchParams.delete(key);
      foundInvalidSearchParams = true;
    }
  }

  if (foundInvalidSearchParams) {
    return c.redirect(url.searchParams.size ? `${url.pathname}?${url.searchParams}` : url.pathname, 302);
  }

  await next();
});

export const stripUnusedSearchParams = createMiddleware<{ Bindings: CloudflareBindings }>(async (c, next) => {
  const url = new URL(c.req.url);
  let foundInvalidSearchParams = false;

  for (const key of url.searchParams.keys()) {
    if (!validSearchParams.includes(key)) {
      url.searchParams.delete(key);
      foundInvalidSearchParams = true;
    }
  }

  if (foundInvalidSearchParams) {
    return c.redirect(url.searchParams.size ? `${url.pathname}?${url.searchParams}` : url.pathname, 302);
  }

  await next();
});

export const stripUnusedSearchParams = createMiddleware<{ Bindings: CloudflareBindings }>(async (c, next) => {
  const url = new URL(c.req.url);
  let foundInvalidSearchParams = false;

  for (const key of url.searchParams.keys()) {
    if (!validSearchParams.includes(key)) {
      url.searchParams.delete(key);
      foundInvalidSearchParams = true;
    }
  }

  if (foundInvalidSearchParams) {
    return c.redirect(url.searchParams.size ? `${url.pathname}?${url.searchParams}` : url.pathname, 302);
  }

  await next();
});

export const stripUnusedSearchParams = createMiddleware<{ Bindings: CloudflareBindings }>(async (c, next) => {
  const url = new URL(c.req.url);
  let foundInvalidSearchParams = false;

  for (const key of url.searchParams.keys()) {
    if (!validSearchParams.includes(key)) {
      url.searchParams.delete(key);
      foundInvalidSearchParams = true;
    }
  }

  if (foundInvalidSearchParams) {
    return c.redirect(url.searchParams.size ? `${url.pathname}?${url.searchParams}` : url.pathname, 302);
  }

  await next();
});

M

Mozzy•10/31/24, 11:12 PM

sadly, req.url is readonly

E

Elliot Hesp•11/1/24, 4:48 PM

Hey, if I have a RPC Target like so

class Foo extends RpcTarget {
  async fetch(request: Request): Promise<Response> {
    return new Response('Hello!');
  }
}

class Foo extends RpcTarget {
  async fetch(request: Request): Promise<Response> {
    return new Response('Hello!');
  }
}

class Foo extends RpcTarget {
  async fetch(request: Request): Promise<Response> {
    return new Response('Hello!');
  }
}

class Foo extends RpcTarget {
  async fetch(request: Request): Promise<Response> {
    return new Response('Hello!');
  }
}

Is there anyway I can just return that response from my worker?

At the moment, I get a type error:

Type '{ clone: Stub<() => Response>; status: number; statusText: string; headers: { get: Stub<(name: string) => string | null>; getAll: Stub<(name: string) => string[]>; ... 9 more ...; [Symbol.iterator]: Stub<...>; }; ... 12 more ...; blob: Stub<...>; } & Disposable' is not assignable to type 'Response'.
  The types of 'clone(...).status' are incompatible between these types.
    Type 'Promise<number>' is not assignable to type 'number'.ts(2322)

Type '{ clone: Stub<() => Response>; status: number; statusText: string; headers: { get: Stub<(name: string) => string | null>; getAll: Stub<(name: string) => string[]>; ... 9 more ...; [Symbol.iterator]: Stub<...>; }; ... 12 more ...; blob: Stub<...>; } & Disposable' is not assignable to type 'Response'.
  The types of 'clone(...).status' are incompatible between these types.
    Type 'Promise<number>' is not assignable to type 'number'.ts(2322)

Type '{ clone: Stub<() => Response>; status: number; statusText: string; headers: { get: Stub<(name: string) => string | null>; getAll: Stub<(name: string) => string[]>; ... 9 more ...; [Symbol.iterator]: Stub<...>; }; ... 12 more ...; blob: Stub<...>; } & Disposable' is not assignable to type 'Response'.
  The types of 'clone(...).status' are incompatible between these types.
    Type 'Promise<number>' is not assignable to type 'number'.ts(2322)

Type '{ clone: Stub<() => Response>; status: number; statusText: string; headers: { get: Stub<(name: string) => string | null>; getAll: Stub<(name: string) => string[]>; ... 9 more ...; [Symbol.iterator]: Stub<...>; }; ... 12 more ...; blob: Stub<...>; } & Disposable' is not assignable to type 'Response'.
  The types of 'clone(...).status' are incompatible between these types.
    Type 'Promise<number>' is not assignable to type 'number'.ts(2322)

B

bkyerv•11/1/24, 6:23 PM

is firebase sdk still not supported in workers?

C

Cross•11/1/24, 9:40 PM

Hi guys, i have a quick question, if i want to implement worker for my domain abc.com, on setting worker whats you think the best practice, set a route or a subdomain?

G

Geert Jan Sloos•11/1/24, 11:04 PM

Is it difficult to create an Email Worker, and check if the incoming email has an PDF?

O

Omar Abid•11/2/24, 2:21 PM

worker deployment is now failing 50% of the time for me. this has been on going for the last couple days.

OOmar Abid worker deployment is now failing 50% of the time for me. this has been on going ...

E

elpupper•11/2/24, 5:05 PM

using rust?

D

DoggeSlapper•11/2/24, 5:22 PM

I wanna create an image with workers, using a base image + some text using specific fonts, can anyone recomend a library that works well on workers?

1

1984 Ford Laser•11/3/24, 1:22 AM

Does the

url

url

field of the draw object need to be a URL, or can it be an image object/blob? I have a second Worker that generates images, to be used as layers in an overlay, but I need to fetch the images to get dimensions before they're drawn as the overlay
https://developers.cloudflare.com/images/transform-images/draw-overlays/

Cloudflare Docs

Draw overlays and watermarks | Cloudflare Images docs

You can draw additional images on top of a resized image, with transparency and blending effects. This enables adding of watermarks, logos, signatures, vignettes, and other effects to resized images.

Eelpupper using rust?

O

Omar Abid•11/3/24, 3:23 AM

Both rust and javascript. it happens randomly.

Original message was deleted

F

fry69•11/4/24, 6:52 AM

See -> https://developers.cloudflare.com/workers/platform/limits/#daily-request

W

Walshy•11/4/24, 12:05 PM

Don't think this exists yet but I'm sure it's something the builds team is tracking

E

elpupper•11/4/24, 3:50 PM

hey so whats the file limits for the static assets?

Eelpupper hey so whats the file limits for the static assets?

F

fry69•11/4/24, 3:53 PM

Apparently those for Pages -> https://developers.cloudflare.com/pages/platform/limits/#file-size

Aabkux ~Hello, can I use postgresql on workers with hono?

T

TravisTravis•11/4/24, 6:30 PM

I was planning to test that out also.

They have a tutorial that shows using the

postgres

postgres

postgres

postgres package:

https://developers.cloudflare.com/workers/tutorials/postgres/

I assume it should work fine with hono.

Cloudflare Docs

Connect to a PostgreSQL database with Cloudflare Workers | Cloudfla...

In this tutorial, you will learn how to create a Cloudflare Workers application and connect it to a PostgreSQL database using TCP Sockets and Hyperdrive. The Workers application you create in this tutorial will interact with a product database inside of PostgreSQL.

T

TravisTravis•11/4/24, 6:35 PM

On that same topic, does anyone know how the TCP connection is reused in that example?

They don't seem to be creating a connection explicitly. Is that part handled by hyperdrive?

Aabkux ~yes, but still anyone could help me with prisma one would be great

T

TravisTravis•11/4/24, 6:58 PM

I was actually looking into that as well.

As far as I can tell it's still a preview feature:

https://www.prisma.io/docs/orm/prisma-client/deployment/edge/overview

Looks like

pg

pg

pg

pg actually is supported though.

Overview: Deploy Prisma ORM at the Edge | Prisma Documentation

Learn how to deploy your Prisma-backed apps to edge functions like Cloudflare Workers or Vercel Edge Functions

That's correct. Pages is full-domain.

Similar Threads

Similar Threads

Similar Threads