MAnd yeah there's no limit so 
FIf the raw image size is 3,5mb this might blow up ~30% with base64 encoding, some AI APIs I know have a 4MB hard limit for image size (in base64).
FDunno if this applies to Cloudflare AI though.
MHmmm, is there a way to easily compress the image in a worker?
FMYeah, not really what's needed here I think. I want to use AI to classify some images but might not work if there's such a small limit
FYou can of course do this manually inside the worker, but you have to code this yourself (you'll kind of pay for the conversion in both ways, either via the image service or via CPU milliseconds for your worker).
MThanks for the help anyway
F
FPlease do not post your question in multiple channels/post it multiple times per the rules at #
welcome-and-rules. It creates confusion for people trying to help you and doesn't get your issue or question solved any faster.
welcome-and-rules. It creates confusion for people trying to help you and doesn't get your issue or question solved any faster.MAnyone from cloudflare who could officially answer this?
CI’d recommend asking in #workers-ai
MGood one
0is the pulumi/terraform provider don't support workers static assets yet is it?
FGit heavily relies on a working filesystem (as git effectively uses the filesystem as its database), I don't know if a VFS shim around R2/S3 will ever make git happy. It may be easier to write/modify a git implementation to use R2/S3 operations directly and skip the VFS layer (like Microsoft did with ditching VFSForGit and moving everything in their git fork)
DQuick maybe dumb question. Can you define a worker entry point in a worker and access and call the class and it’s methods in the same worker? I understand with RPC you can export the class in the same file as the module default export and it can be used in other workers. But wondering if it can be used in the source worker and other workers
II think you can bind it directly within that same worker and access it via
env , here’s a demo: https://github.com/cloudflare/js-rpc-and-entrypoints-demo/tree/main/public-admin-api-interfaces/db-service. I haven’t tried it myself though
DThanks I’ll try that out
DI’m also gonna try calling a rpc in a workflow. Wondering if that will work
GHello everyone, have any of you tried to deploy nextjs on workers?
GThanks ! Which is better cloudflare pages or workers for nextjs
FNo clue. Workers are more up-to-date and powerful though.
DWorking through this example... Is this bug described here still outstanding? https://github.com/cloudflare/js-rpc-and-entrypoints-demo/blob/12f7f2e10a4543c229554bcec4cba139e16594ce/public-admin-api-interfaces/db-service/wrangler.init.toml
GitHub
Contribute to cloudflare/js-rpc-and-entrypoints-demo development by creating an account on GitHub.
Ais there a way to have a subrequest outlive the parent request in a worker?
Ae.g. POST /foo -> worker kicks off an internal fetch() to /someapi, meanwhile responds to the POST /foo with (202)
Ain my testing the subrequest is getting killed when the 202 response is sent
Cctx.waitUntil
CCloudflare Docs
The Context API in Cloudflare Workers, including waitUntil and passThroughOnException.
Cyou only get 30s of real time though for it to finish, or it'll get cancelled anyway
Agot it!
Athat works like a charm, thanks very much
AHey 
I am trying to call a DO from Worker B when DO is exported from worker A.
Worker A looks like this:
And the index.ts in worker B is a fetch calling a DO stub (with RPC), and exporting the fetch.
On Worker A, it's a simple fetch + DO sqlite, both exported.
Worker A deploys fine and works fine with DO. I get this error on deploy for worker B (that should use the DO from worker A):
Any idea what I could be doing wrong?
I am trying to call a DO from Worker B when DO is exported from worker A.
Worker A looks like this:
And the index.ts in worker B is a fetch calling a DO stub (with RPC), and exporting the fetch.
On Worker A, it's a simple fetch + DO sqlite, both exported.
Worker A deploys fine and works fine with DO. I get this error on deploy for worker B (that should use the DO from worker A):
Cannot create binding for class in script 'worker-A' that does not exist [code: 10061]Any idea what I could be doing wrong?
AFound it! the script name needs to be worker-A-dev
Dguessing i'm out of luck? can't really use D1 if you can't bind anything on the fly.
HHave you tried using SQLite DOs instead of D1?
D1G storage is prohibitive for us. I'm just not sure why the API just returns success = true but nothing happens
KHow "on the fly" are you looking for? It wouldn't affect that Worker invocation that calls it.
Dmy plan is when someone clicks something from my CF pages site, it creates a D1 database and we need to bind that to 3 CF workers. It doesn't have to be instant but quick enough that it's not waiting 5 - 10 minutes
KHow are you using the API? Wrangler itself uses it so it should be fine
Dcant use wrangler because its done by the client in pages
Di'm using this endpoint: https://developers.cloudflare.com/api/operations/worker-script-patch-settings but it just reutrns sucess = true and bindings are never updated
Interact with Cloudflare's products and services via the Cloudflare API
Deven though it does generate a new deployment
Dafaik - as long as it's formatted right, it'll always respond success no matter the input
KWhat's your payload?
DSending body: ---------------formdata45241039827298346
Content-Disposition: form-data; name="settings"
{"settings":{"bindings":[{"database_id":"821eb65a-ebdd-49ef-aa0b-878f65b57d57","name":"inbound-5772b3b2-6bca-42ed-938b-f3e17b2781ca","type":"d1","id":"821eb65a-ebdd-49ef-aa0b-878f65b57d57"}]}}
---------------formdata45241039827298346--
Worker binding response: {
"result": {
"placement": {},
"compatibility_date": "2024-09-08",
"compatibility_flags": [],
"usage_model": "standard",
"tags": [],
"tail_consumers": [],
"logpush": false,
"bindings": [
{
"database_id": "2206f7ed-1476-40a0-9758-f1bc3d47e5ad",
"id": "2206f7ed-1476-40a0-9758-f1bc3d47e5ad",
"name": "2206f7ed-1476-40a0-9758-f1bc3d47e5ad",
"type": "d1"
},
]
},
"success": true,
"errors": [],
"messages": []
}
Worker binding process completed
Content-Disposition: form-data; name="settings"
{"settings":{"bindings":[{"database_id":"821eb65a-ebdd-49ef-aa0b-878f65b57d57","name":"inbound-5772b3b2-6bca-42ed-938b-f3e17b2781ca","type":"d1","id":"821eb65a-ebdd-49ef-aa0b-878f65b57d57"}]}}
---------------formdata45241039827298346--
Worker binding response: {
"result": {
"placement": {},
"compatibility_date": "2024-09-08",
"compatibility_flags": [],
"usage_model": "standard",
"tags": [],
"tail_consumers": [],
"logpush": false,
"bindings": [
{
"database_id": "2206f7ed-1476-40a0-9758-f1bc3d47e5ad",
"id": "2206f7ed-1476-40a0-9758-f1bc3d47e5ad",
"name": "2206f7ed-1476-40a0-9758-f1bc3d47e5ad",
"type": "d1"
},
]
},
"success": true,
"errors": [],
"messages": []
}
Worker binding process completed
Dhad to purge some of the bindings that have came back because length
Dive done a ton of different configurations, incl passing existing bindings through, matching all the flags in the result, etc with no real change
KRemove the top-level settings object
