Deprecations
This document describes the difference between wrangler
versions 1 and 2, specifically deprecations and breaking changes.
Configuration
wrangler
version 2.0.0 introduces some new fields for configuration, while also deprecating some redundant fields.
New fields
main
: string, optionalThe
main
field is used to specify an entry point to the Worker. It may be in the established service worker format[], or the newer, preferred modules format[
]. An entry point is now explicitly required, and can be configured either via themain
field, or passed directly as a command line argument; for example,wrangler dev index.js
. This field replaces the legacybuild.upload.main
field (which only applied to modules format Workers). Read more at (ref:)[]rules
: array, optionalThe
rules
field is an array of mappings between module types and file patterns. It instructswrangler
to interpret specific files differently than JavaScript. For example, this is useful for reading text-like content as text files, or compiled WASM as ready to instantiate and execute. These rules can apply to Workers of both the established service worker format, and the newer modules format. This field replaces the legacybuild.upload.rules
field (which only applied to modules format Workers). Read more at (ref:)[]
Non-mandatory fields
A few configuration fields which were previously required, are now optional in particular situations; they can either be inferred, or added as an optimization. No fields are required anymore when starting with wrangler
version 2, and you can gradually add configuration as the need arises.
name
: stringThe
name
configuration field is now not required forwrangler dev
, or any of thewrangler kv:*
commands. Further, it can also be passed as a command line argument as--name <name>
. It is still required forwrangler publish
.account_id
: stringThe
account_id
field is not required for any of the commands. Any relevant commands will check if you are logged in, and if not, will prompt you to log in. Once logged in, it will use your account ID and will not prompt you again until your login session expires. If you have multiple account IDs, you will be presented with a list of accounts to choose from.You can still configure
account_id
in yourwrangler.toml
file, or as an environment variableCLOUDFLARE_ACCOUNT_ID
; this will make startup faster, and bypass the list of choices if you have multiple IDs. TheCLOUDFLARE_API_TOKEN
environment variable is also useful for situations where it is not possible to login interactively - refer to Running in CI/CD .workers_dev
boolean, default:true
when no routes are presentThe
workers_dev
field is used to indicate that the Worker should be published to a*.workers.dev
subdomain. For example, for a Worker namedmy-worker
and a previously configured*.workers.dev
subdomainusername
, the Worker will get published tomy-worker.username.workers.dev.com
. This field is not mandatory, and defaults totrue
whenroute
orroutes
are not configured. When routes are present, it defaults tofalse
. If you want to neither publish it to a*.workers.dev
subdomain nor any routes, setworkers_dev
tofalse
. This useful when you are publishing a Worker as a standalone service that can only be accessed from another Worker with (services
)[TODO:link/to/services/docs] bindings.
Deprecated fields (non-breaking)
A few configuration fields are deprecated, but their presence is not a breaking change yet. It is recommended to read the warning messages and follow the instructions to migrate to the new configuration. They will be removed and stop working in a future version.
zone_id
: string, deprecatedThe
zone_id
field is deprecated and will be removed in a future release. It is now inferred fromroute
/routes
, and optionally fromdev.host
when usingwrangler dev
. This also makes it simpler to deploy a single Worker to multiple domains.build.upload
: object, deprecatedThe
build.upload
field is deprecated and will be removed in a future release. Its usage results in a warning with suggestions on rewriting the configuration file to remove the warnings.build.upload.main
/build.upload.dir
are replaced by themain
fields and are applicable to both service worker format and modules format Workers.build.upload.rules
is replaced by therules
field and is applicable to both service worker format and modules format Workers.build.upload.format
is no longer specified and is automatically inferred bywrangler
.
Deprecated fields (breaking)
A few configuration fields are deprecated and will not work as expected anymore. It is recommended to read the error messages and follow the instructions to migrate to the new configuration.
site.entry-point
:string
, deprecatedThe
site.entry-point
configuration was used to specify an entry point for Workers with a[site]
configuration. This has been replaced by the top-levelmain
field.type
:rust
|javascript
|webpack
, deprecatedThe
type
configuration was used to specify the type of Worker. It has since been made redundant and is now inferred from usage. If you were usingtype = "webpack"
(and the optionalwebpack_config
field), you should read the TODO: webpack migration guide to modify your project and use a custom build instead.
Deprecated commands
The following commands are deprecated in Wrangler as of Wrangler 2.
build
:
The wrangler build
command is no longer available for building the Worker.
The equivalent functionality can be achieved by wrangler publish --dry-run --outdir=path/to/build
.
config
:
The wrangler config
command is no longer available for authenticating via an API token.
Use wrangler login
/ wrangler logout
to manage OAuth authentication, or provide an API token via the CLOUDFLARE_API_TOKEN
environment variable.
preview
:
The wrangler preview
command is no longer available for creating a temporary preview instance of the Worker.
Try using wrangler dev
to try out a worker during development.
subdomain:
The wrangler subdomain
command is no longer available for creating a workers.dev
subdomain.
Create the workers.dev
subdomain on your Workers dashboard.
generate:
The wrangler generate
command is no longer available to initialize a Worker from a template repository.
Initialize a basic Worker project via wrangler init
, which provides a good starting point for most projects.
Alternatively, try cloning the template repository and initializing it manually.
route:
The wrangler route
command is no longer available to configure a route for a Worker.
Routes are specified in the wrangler.toml
configuration file.
Other deprecated behaviour
wrangler
will no longer useindex.js
in the directory wherewrangler dev
is called as the entry point to a Worker. Use themain
configuration field, or explicitly pass it as a command line argument, for example:wrangler dev index.js
.wrangler
will no longer assume that bare specifiers are file names if they are not represented as a path. For example, in a folder like so:project ├── index.js └── some-dependency.js
where the content of
index.js
is:import SomeDependency from "some-dependency.js";addEventListener("fetch", (event) => {// ...});wrangler
1 would resolveimport SomeDependency from "some-dependency.js";
to the filesome-dependency.js
. This will also work inwrangler
2, but will also log a deprecation warning; in the future, this will break with an error. Instead, you should rewrite the import to specifiy that it is a relative path, like so:- import SomeDependency from "some-dependency.js";+ import SomeDependency from "./some-dependency.js";