Do functions work with NextJS deployments? Wrangler doesn't seem to be picking up my functions.
Do functions work with NextJS deployments? Wrangler doesn't seem to be picking up my functions.
JIf you are using
@cloudflare/next-on-pages_worker.js
Sgot it. ty!
Fhttps://developers.cloudflare.com/pages/tutorials/build-an-api-with-pages-functions/
not a repo but detailed instructions for building one
not a repo but detailed instructions for building one
In this tutorial, you will build a full stack Pages application. Your application will contain:
Sdo functions run in front of the cache just as workers?
KYup
KWell, Pages itself is a Worker so there isn’t really any origin cache in the first place - just the Cache API
EStatic assets have their own internal cache
Edynamic responses from functions do not, so those need to use cache api
Ahas anyone gotten langchain to work on cloudflare pages? it says
crypto.getHashes has not been implemented yet when i try to deploy
VIt should work using fetch()
AThanks for the reply. Can you specify where I should use fetch()? the error occurs in deployment when i have this inside /server/api/query.ts (using nuxt 3):
UI would assume both are exactly the same results just slightly different response 
Vsorry, I've never used langchain, so can't really tell you specifically how to make it work
Aokay thanks anyway!
Uare you getting errors?
EI would recommend contacting MailChannels support with as much detail as possible https://support.mailchannels.com/hc/en-us/requests/new
TJust jumped on Pages and I am liking it so far, added a trivial function, setup route filtering etc. All is good, but I can't for the life of me figure out when I should use functions and when I should use workers. Functions are workers but with a different API and some file based routing sprinkled on top? When would you use a worker in a Pages/function deploy?
DIf you want better analytics per worker then use workers, if you want to be able to specifically choose which billing method (bundled or unbundled) per api route then use workers
DPages is a little behind workers and currently works like a blanket system that I assume is mainly there to help streamline SSR, but you would get more fine grained control using a worker
TAll right, that makes sense, too bad workers doesn't have the nice GitHub-integration that pages has
DIt kind of does though, if you work on it locally you can deploy the worker with wrangler and I believe it will use the commit description that you last used on it. You can also then go to the dashboard and see that history and rollback to whatever deployment you had
TSure and that works well for a single dev, but if I had multiple people working on it and perhaps mulitple workers (I'm currently looking at having just one for now and do routing) I'd prefer everyone not having to deploy from their machines 
TBut it's a minor thing really, for now I just want to get more clarity on when to use what
DApparently from what I've learnt it's ok to have 1 worker containing everything you need since it only really affects cold start times, and as long as it's warm it should be instant I think
DJust make sure any new npm packages don't bloat your worker, I had some random packages I didn't even use much load entire 7MB json files into the worker
TThe dangers of npm, I'm a frequent user of bundlephobia.com for that reason
Acould be a compression issue
Aim having a similar sort of problem with a specific deployment of my site, but every other deployment is fine
AI have not managed to fix it
Ait seems to only happen on my own network which is weird
JI'm going to escalate this for you - we've seen this happening quite a bit recently.
JThanks Matt
