I know not an ideal solution but when using functions I let wrangler do its thing to run the functio

I know not an ideal solution but when using functions I let wrangler do its thing to run the function with the ./dist but I also run a the vite server for my frontend on its own. And access the :5173 instead of the :87xx by wrangler.
This way when using the functions it uses the wrangler's endpoint to get the data but otherwise I take advantage of Vite's HMR for the frontend.
I am not sure how this will work for other CF services run through wrangler.

Oom Idk how to setup DO in pages. How to export

Hello, I’m Allie!•4/15/24, 2:07 PM

You need a Worker to export the DO

Hello, I’m Allie!•4/15/24, 2:08 PM

As for RPC, it should work

Vvimo I know not an ideal solution but when using functions I let wrangler do its thin...

IEatBeans•4/15/24, 3:33 PM

I thought about that but won’t the frontend try to use the 5173 port when fetching data, or do just just manually specify the 8788 port for all fetch’s?

IIEatBeans I thought about that but won’t the frontend try to use the 5173 port when fetchi...

vimoOP•4/15/24, 3:45 PM

yes manually specify it. using a .env.local file as a proxy of sorts in development. its cumbersome but works.

Vvimo Yes, when theres a 500 internal error it can't send a response. I am fetching f...

IEatBeans•4/15/24, 4:29 PM

So when the api you are calling from within the function is offline, it fails to fetch, and the worker is just returning 500 because the response never exists.

If this is the case then I think you can use a .catch() after the fetch and put a callback that returns whatever error you want your function to return when that happens.

IEatBeans•4/15/24, 4:30 PM

In a .then() (or if using await) then you can do basically the same with the status codes.

IIEatBeans So when the api you are calling from within the function is offline, it fails to...

vimoOP•4/15/24, 4:41 PM

Thanks for still thinking about it. Unfortunately, that doesn't work or not at least locally.
It does not return anything as the fetch never completes, so the function just shows 500 internal error on the server but doesn't execute any further. I tried shortcircuit with datares.status to check for 500 but it doesn't work that way as by that time the function errors out and not the origin server (for which datares checks for).
The 500 error is the function error but nothing is checking for the function error itself.
I am not sure if I am explaining it well.

I found out this, apparently the 500 error will only be returned if its timeout by due to DNS error https://community.cloudflare.com/t/community-tip-fixing-error-500-internal-server-error/44453 but other reasons it will not return anything like in my case where the origin server does not respond. The linked docs page in that article seems helpful for all the possible CF error codes

Vvimo Thanks for still thinking about it. Unfortunately, that doesn't work or not at l...

IEatBeans•4/15/24, 4:43 PM

Sorry but im still not sure I’m understanding.

IEatBeans•4/15/24, 4:44 PM

Maybe you could share some code around this spot?

kajuwa•4/16/24, 6:03 AM

hello everyone where can i ask about cloudflare calls beta about inactivity calls to disconnect as i cannot find any snippet code to do this

Erisa•4/16/24, 10:48 AM

#stream

Paolo•4/18/24, 12:39 PM

Is there some authentication support for Functions? what I've seen for Okta seems to require express

PPaolo Is there some authentication support for Functions? what I've seen for Okta seem...

IEatBeans•4/18/24, 4:37 PM

Lucia Auth library supports functions I think, though it definitely requires more work on your end than a paid solution

sudo__su--•4/18/24, 7:11 PM

does cloudflare offer a way to run a nest.js backend serverless to conenct to like a gitlab (i was thinking of doing vercel but i like whats in CF capabilitys

Akash•4/20/24, 2:24 PM

I have created a pages function with following code in a file called index.js and put it directly under the functions folder -

export async function onRequestGet(context) {
    console.log('Inside index.html proxy function ..')
    return context.next()
}

export async function onRequestGet(context) {
    console.log('Inside index.html proxy function ..')
    return context.next()
}

The idea behind this function is to execute it before serving index.html file ONLY and to avoid it for other static assets from pages application. This is something which I could not achieve with worker or functions middleware. While this works in local setup, this is not getting executed when deployed on cloudflare. So my question is which behavior (local or cloudflare) is correct?

AAkash I have created a pages function with following code in a file called index.js an...

Hello, I’m Allie!•4/20/24, 4:59 PM

It should just work? Just published https://functest.pages.dev/ to test, and only the index should have the header

functions: wuz here

functions: wuz here

Astro Blog

Welcome to my website!

HHello, I’m Allie!It should just work? Just published https://functest.pages.dev/ to test, and onl...

Akash•4/20/24, 7:21 PM

Can you try this with a custom domain and see if it is working? If yes then I will create a simple pages project tomorrow and test it again.

AAkash Can you try this with a custom domain and see if it is working? If yes then I wi...

Hello, I’m Allie!•4/20/24, 7:24 PM

https://functest.goalastair.com/

Astro Blog

Welcome to my website!

Akash•4/20/24, 7:29 PM

Thanks. Will test it again and see if something is wrong on my side.

HHello, I’m Allie!https://functest.goalastair.com/

Akash•4/20/24, 7:31 PM

Can you share the code of your pages function? Did you make any other change to your pages functions application to make it work?

dave•4/21/24, 2:13 PM

Does Functions support logpush?

dave•4/21/24, 3:22 PM

Oh, I already asked this last year. functions

Rok•4/21/24, 4:43 PM

Is having an "exclusively" function (worker) backed Pages project without a SPA in front considered an anti-pattern and would eventually bite me?

I guess that having a SPA in front might be cheaper to run, since not every request will trigger a worker invocation, but it could be more expensive (slower) to develop without enough experience with the selected SPA framework

HHello, I’m Allie!https://functest.goalastair.com/

Akash•4/21/24, 5:10 PM

I tested it with a simple functions application today and it is working as expected. I need to figure out why it is not working for my other application. Thank you for looking into it.

Brutal Strike•4/22/24, 9:25 AM

@HardlyWorkin' I have an issue with my Cloudflare Functions. I followed the steps to create a Cloudflare Pages project with Functions using the

wrangler pages project create

wrangler pages project create

command, but I'm encountering problems when trying to access my functions.

I created a new function named

hello.js

hello.js

inside the

functions

functions

directory with the following content:

export async function onRequest(context) {
  return new Response("Hello, World!");
}

export async function onRequest(context) {
  return new Response("Hello, World!");
}

When I tried to test my function locally using

npx wrangler pages dev ./

npx wrangler pages dev ./

, I couldn't access it at

http://localhost:8788/hello

http://localhost:8788/hello

. Additionally, after deploying my project to Cloudflare Pages, I attempted to access my function at

https://<my-project-name>.pages.dev/hello

https://<my-project-name>.pages.dev/hello

, but it didn't work either.

BBrutal Strike @HardlyWorkin' I have an issue with my Cloudflare Functions. I followed the ste...

Hello, I’m Allie!•4/22/24, 9:53 AM

Try putting your assets in a subdirectory

HHello, I’m Allie!Try putting your assets in a subdirectory

Brutal Strike•4/22/24, 10:02 AM

like this? I still have: No webpage was found for the web address: https://5538651d.test2-55l.pages.dev/hello

zegevlier•4/22/24, 10:05 AM

Put the

functions

functions

folder in the

test

test

folder, then upload the

public

public

folder.

Brutal Strike•4/22/24, 10:08 AM

thank you now its working

Paolo•4/22/24, 10:42 AM

I'm trying the Vitest testing tutorial for Functions but I get the following error "Using service bindings to the current worker requires

poolOptions.workers.main

poolOptions.workers.main

to be set to your worker's entrypoint". But I do not have an entrypoint :/

Zzegevlier Put the `functions` folder in the `test` folder, then upload the `public` folder...

Marius Gundersen•4/22/24, 5:32 PM

The way to structure the project isn't terribly clear from the documentation. It's not easy to understand from the words alone that functions should not be in public and where to put all the other special files (_redirects, _routes.json). I wish the documentation had a clear example of where these files should be

King•4/22/24, 11:57 PM

Hello everyone, Is it possible to send emails with attachments using the mailchannels plugin? Or has this feature not been implemented?

Paolo•4/23/24, 8:49 AM

has anyone gotten vitest working with functions like it's done with workers?

SSnorre Hello everyone. What would be the best way to co-ordinate refresh and access tok...

js•4/24/24, 12:50 AM

Hey did you ever determine the best solution for this?

tozz•4/24/24, 7:00 AM

Anyone else noticed that if you have a

onRequest

onRequest

function defined, other specific functions like

onRequestOptions

onRequestOptions

no longer gets called in the middleware? Seems like a regression since it used to work and I don't see anything about it changing in the release notes

danboyle8637•4/24/24, 1:47 PM

For some reason... it appears my service bindings are set up incorrectly but I can't figure out why.

I have a binding from a "normal" Worker to my Pages Worker. It's setup correctly in the dashboard (I am not using the wrangler.toml file in Pages yet).

In my particular case, I create a new Request object where the url is the path that matches a route in my destination Worker.

However, I always get an error stating that fetch can't call the url path I supply.

Just to be clear... I call my Request using

env.WORKER_BINDING_NAME.fetch(Request)

env.WORKER_BINDING_NAME.fetch(Request)

-- this call is made in a Pages function endpoint and I await it because I want the response.

The last wrinkle is I am testing in my staging environment. And in my CF Dashboard I bind to the correct Worker... BUT... there is a drop down when binding Workers for

environment

environment

and (for me at least) the only option I ever have it

production

production

.

So my Binding is made to my staging Worker... but the Dashboard binding drop down only ever gives

production

production

as an option.

I plan on switching to

wrangler.toml

wrangler.toml

config rather than Dashboard. It seems easier to manage for sure.

Any ideas on why it appears my Service Bindings aren't working?

Alek•4/24/24, 3:52 PM

It seems like I always get a MISS for the same URL I want to cache. Is it something I'm missing? "try {
let url = await context.request.url;

let paramString = url.split('?')[1];

if (paramString.includes('ntt=') || paramString.includes('lpurl'))
{
const today = new Date();
const tomorrow = new Date(today.getFullYear(), today.getMonth(), today.getDate() + 1);
const seconds = Math.floor((tomorrow.getTime() + 18000000 - today.getTime()) / 1000);
console.log(seconds);

let response = await fetch(fetchUrl.concat('?', paramString), {
method: "GET",
cf: {
cacheTtl: seconds,
cacheEverything: true,
cacheKey: url,
},
});
let headersObject = Object.fromEntries(response.headers);
let responseHeaders = JSON.stringify(headersObject, null, 2);
console.log(

Response headers: ${responseHeaders}

Response headers: ${responseHeaders}

);
return response;
}
} catch (err) {
return new Response('Error parsing JSON content', { status: 400 });
}"

Varna•4/24/24, 4:10 PM

I'm deploying via GitHub, got wrangler.toml:

# stuff above
compatibility_flags = ["nodejs_compat"]
[vars]
PROFILE_INFO = { baseURL = "https://example.com/", USER = "root", PASS = "toor" }
# more stuff below

# stuff above
compatibility_flags = ["nodejs_compat"]
[vars]
PROFILE_INFO = { baseURL = "https://example.com/", USER = "root", PASS = "toor" }
# more stuff below

but I fail to publish the project with error:

    Error: Failed to publish your Function. Got error: invalid or missing json property for binding PROFILE_INFO

    Error: Failed to publish your Function. Got error: invalid or missing json property for binding PROFILE_INFO

Also got warnings like

Your Worker may throw errors at runtime unless you enable the "nodejs_compat" compatibility flag.

Your Worker may throw errors at runtime unless you enable the "nodejs_compat" compatibility flag.

AAvinash Is there any way to intercept url query parameters (example.com/hello?name=joe) ...

IEatBeans•4/24/24, 4:40 PM

This is how I do it. Not intuitive but it works:

new URL(request.url).searchParams.get("days")

I know not an ideal solution but when using functions I let wrangler do its thing to run the functio

Similar Threads

Similar Threads

Similar Threads