so wrangler cannot tear everything down automatically with one command?
so wrangler cannot tear everything down automatically with one command?
IThe resource names in the TOML just links the resource with your Worker.
IWrangler can tear down the worker, and Wrangler can tear down the resource.
IIf both are down, then there's no link between them because neither exist. Similarely, if you delete the Worker the resource will continue to exist, and you can't delete the resource without deleting the Worker (I believe it will block you).
Lso wrangler does both?
IYes, but in seperate steps. You need to create the resource (with Wrangler or the dashboard), and then add it to your TOML to tell Wrangler to link it to that specific Worker when you deploy the Worker (because a resource can be linked to many Workers at once, they're not specific to a single Worker).
IIt's similar to how you might spin up an RDS instance on AWS, or a DB instance on Azure, and then add the credentials to whatever app needs to access it (Wrangler does this for you)
Lwhy it cannot do it like terraform
IYou can delete the app without deleting the DB.
IBecause it's not designed to do that. Wrangler is a deployment tool, not an IaC tool like Terraform.
Lbut sst does both
IWith Wrangler you say "deploy this Worker for me, oh also the Worker needs to access this (possibly shared) resource", but with Terraform/SST you say "I need this resource, and this Worker, and they need to talk, create it all for me"
Lanyway, i tried to update all deps to latest available version, and it finds conflicting peer deps
IYou can fix that by changing the version of vitest to 1.3.0, but if you're not actually seeing any issues then I wouldn't.
Lno, alrady tried that
L
IIt doesn't really matter unless you're actually seeing errors in Wrangler or Vitest.
Lby default C3 uses vitest-pool-runner 0.1.19 which is crazy old
Ls
wrangler.tomlwrangler.toml
HFor the "multiple workers in a project" part, you can just put your
wrangler.tomlwrangler.tomlso need a monorepo to handle multiple workers
HYou can also split it into multiple repos too if you want
NHi all! New contributor here, so apologies if this is not the appropriate place to ask. I'm trying to run the unittests in the workers-sdk repository, but so far can't get it to work. I have plenty of developer experience, but it's been a long time since I last worked with Typescript and I'm not too familiar with vitest or how to set things up locally...
This is what I was trying to do, but apparently it's not the right way to run the tests.
➜ workers-sdk git:(main) ✗ npx vitest run packages/create-cloudflare
▲ [WARNING] Cannot find base config file "@cloudflare/workers-tsconfig/tsconfig.json" [tsconfig.json]
If anyone could point me in the right direction it would be very much appreciated!
This is what I was trying to do, but apparently it's not the right way to run the tests.
➜ workers-sdk git:(main) ✗ npx vitest run packages/create-cloudflare
▲ [WARNING] Cannot find base config file "@cloudflare/workers-tsconfig/tsconfig.json" [tsconfig.json]
If anyone could point me in the right direction it would be very much appreciated!
Nis wrangler or c3 in more active development?
JJlast wrangler release was just on friday. And last c3 release was just a couple weeks ago
ZI love all the work that's gone into running Workers locally... I started using Cloudflare several years ago and the situation is so much nicer now for local development. However, as more and more awesome new products get added, I get frustrated every time I want to start using something that doesn't have local development support. It would be amazing if we had the option of still binding to some remote services while running everything locally that can be local. For example, I really want to start using Vectorize, but if I can't use it in my normal local dev stack, then it's a non-starter. Am I just missing something? Is there a way to bind things like Vectorize, AI gateway, etc to local dev?
AIs anyone having failures with
npx wrangler login today? It was working fine yesterday, but dtoday I'm getting a non-json response back from the API SyntaxError: Unexpected token < in JSON at position 0
Jsadly not. Some bindings simply don't work locally today unfortunately. https://docs.google.com/spreadsheets/d/1T7j7FZZO7zmkrEegzhMeMtlFPIh-efYGMc7BuenMPWY/edit#gid=1961597022 is a good breakdown.
Aat least with vectorize, you could hit the API from local https://developers.cloudflare.com/api/operations/vectorize-upsert-vector
Interact with Cloudflare's products and services via the Cloudflare API
SHi! Will be supported calling DO using rpc from two different wrangler sessions? For service binding it is already supported.
ZThanks, that's a good resource
ZTrue. I guess I could make my own local binding proxy to remote - and using RPC, the binding could look just like a real Vectorize binding
II’m trying to understand the relationship between all the cloudflare sdk tools. Could you correct me if I’m wrong here, in that create cloudflare (c3) just sets up the project (files, scripts, npm stuff), wrangler is a cli that wraps mini flare to make it easy to use basic features, mini flare is a full featured tool that manages the bindings and stuff, and uses workerd to actually run the workers and stuff?
If that’s correct does that mean I can use mini flare without wrangler to customize the dev process more? Like for features wrangler doesn’t have, I can do it through mini flare?
If that’s correct does that mean I can use mini flare without wrangler to customize the dev process more? Like for features wrangler doesn’t have, I can do it through mini flare?
IOr do I still need wrangler
RI used to create a miniflare server directly for testing. You should be able to start it from a node script without wrangler. https://www.npmjs.com/package/miniflare
ZIs there a way to bind the same local KV store to 2 different apps when developing locally?
HUse
--persist-to to point them both to the same directory. As long as they have the same ID, they should pull from the same place
ZAwesome thanks.
JHas anybody run into issues (or is it a common issue) with R2 bindings running locally via wrangler, where binary data (eg. a png) is corrupted when downloaded (
env.R2.get(...)). And appears the encoding is borked somewhere along the way. I can see that the file is not corrupted on the disk (within the .wrangler folder) but when I download the file, via the worker, it "corrupts". I the screenshot attached, the left is the original file, and the right is after downloading from R2. It's worth mentioning that if I upload a text file (eg. xml) - it works fine.
JRunning the same code on an actual Cloudflare Worker works fine.
BI started seeing authz errors today when trying to import data to a newly created db.
is now giving me this:
I just updated wrangler to 3.57.0. I'll report if someone thinks it might be a bug. Thanks
wrangler d1 execute <DBNAME> --remote --file=./creation_insertion_stmts.sqlis now giving me this:
I just updated wrangler to 3.57.0. I'll report if someone thinks it might be a bug. Thanks
Knot authorized means you tried using a pragma/function that isn't allowlisted
BThank you, I do have that
PRAGMA foreign_keys=OFF;KD1’s foreign key enforcement is equivalent to SQLite’s PRAGMA foreign_keys = on directive. Because D1 runs every query inside an implicit transaction, user queries cannot change this during a query or migration.
KKYou can use
defer_foreign_keys thoughBthank you, looks like I need to start leaving off the
https://developers.cloudflare.com/d1/build-with-d1/import-data/#convert-sqlite-database-files
I missed the change.
BEGIN TRANSACTION; and COMMIT; as well. I'll review that allowlist. Thanks again, I've been able to import everything remotely now.https://developers.cloudflare.com/d1/build-with-d1/import-data/#convert-sqlite-database-files
I missed the change.
Cloudflare Docs
D1 allows you to import existing SQLite tables and their data directly, enabling you to migrate existing data into D1 quickly and easily. This can be …
