This page is a reference for all of the Figwheel configuration options.
You can enter these options in a figwheel-main.edn
file that is in the root
of your project directory.
Example figwheel-main.edn
file:
{:watch-dirs ["src" "admin-src"]
:css-dirs ["resources/public/css"]}
The options can also be entered as metadata in a Figwheel build file
in your project's root directory. The name of a build file has the form
[build-id].cljs.edn
where [build-id]
is an identifier of your
choice.
An example dev.cljs.edn
build file that supplies figwheel config options.
^{:watch-dirs ["src" "admin-src"]
:css-dirs ["resources/public/css"]}
{:main example.core}
Any options provided in the metadata of the build file will override the
options in the figwheel-main.edn
file.
The options below are listed in order of importance
A list of ClojureScript source directories to be watched and compiled on change.
The directories in :watch-dirs
are passed to the compiler as source
directories. For this reason, any entry in the watch directories must
be on the classpath and must point to the root directory of a
ClojureScript namespace source tree.
I.E. If your example.core
namespace is located at
src/cljs/example/core.cljs
you cannot use src
as an element of
:watch-dirs
, you must use the path to the root directory of the
namespace tree src/cljs
.
:watch-dirs ["cljs-src"]
A list of CSS source directories to be watched and reloaded into the browser.
:css-dirs ["resources/public/css"]
A symbol or string indicating a ring-handler to embed in the figwheel.repl server. This aids in quickly getting a dev server up and running. If the figwheel server doesn't meet your needs you can simply start your own server. The figwheel.client will still be able to connect to its websocket endpoint. Default: none
:ring-handler my-project.server/handler
All the options to forward to the ring-jetty-adapter/run-jetty
function
which figwheel.main uses to run its ring server.
All the available options are documented here: https://github.com/ring-clojure/ring/blob/master/ring-jetty-adapter/src/ring/adapter/jetty.clj#L127
This will normally be used to set the :port
and :host
of the server.
Most uses of these options are considered advanced. If you find yourself using many of these options you problably need to run your own server outside of figwheel.main.
By default Figwheel engauges a Rebel readline editor when it starts the ClojureScript REPL in the terminal that it is launched in.
This will only work if you have com.bhauman/rebel-readline-cljs
in
your dependencies.
More about Rebel readline: https://github.com/bhauman/rebel-readline
Default: true
:rebel-readline false
When :pprint-config
is set to true. The figwheel.main
will print the
computed config information and will terminate the process. Useful for
understanding what figwheel.main adds to your configuration before it
compiles your build.
Default: false
:pprint-config true
A path to an executable shell script that will be passed a file and line information for a particular compilation error or warning.
A script like this would work ie. in ~/bin/myfile-opener
#! /bin/sh
emacsclient -n +$2:$3 $1
Then add this script in your config:
:open-file-command "myfile-opener"
But that's not the best example because Figwheel handles emacsclient
as a special case. So as long as emacsclient
is on the shell path you can
simply do:
:open-file-command "emacsclient"
and Figwheel will call emacsclient
with the correct args.
Whether to include the figwheel.core library in the build. This enables hot reloading and client notification of compile time errors. Default: true
:figwheel-core false
Whether or not figwheel.core should hot reload compiled ClojureScript. Only has meaning when :figwheel is true. Default: true
:hot-reload-cljs false
Whether or not figwheel.core should reload the namespaces that
depend
on the changed namespaces in addition to the changed
namespaces themselves. Only has meaning when :figwheel is true.
Default:true
:reload-dependents false
The url that the figwheel REPL client will use to connect back to the server.
This url is actually a template that will be filled in. For example
the default :connect-url
is:
"ws://[[config-hostname]]:[[server-port]]/figwheel-connect"
The available template variables are:
For the server side:
[[config-hostname]] the host supplied in :ring-server-options > :host or "localhost"
[[server-hostname]] the java.InetAddress localhost name - "Bruces-MacBook-Pro.local" on my machine
[[server-ip]] the java.InetAddress localhost ip interface - normally 192.168.x.x
[[server-port]] the port supplied in :ring-server-options > :port or the default port 9500
On the client side:
[[client-hostname]] the js/location.hostname on the client
[[client-port]] the js/location.port on the client
If the url starts with a Websocket scheme "ws://" a websocket connection will be established. If the url starts with an http scheme "http" an http long polling connection will be established.
Either a boolean value false
or a string that indicates the url
that the figwheel REPL will open in the browser after the source code
has been compiled. A false
value will disable this behavior.
The string value is actually a template that can provide optional
template variables. For example the default :open-url
is:
"http://[[server-hostname]]:[[server-port]]"
The available template variables are:
For the server side:
[[server-hostname]] the host supplied in :ring-server-options > :host or "localhost"
[[server-port]] the port supplied in :ring-server-options > :port or the default port 9500
Figwheel naively reloads clj
and cljc
files on the :source-paths
.
It doesn't reload clj dependent files like tools.namspace.
Figwheel does note if there is a macro in the changed clj
or cljc
file
and then marks any cljs namespaces that depend on the clj
file for
recompilation and then notifies the figwheel client that these
namespaces have changed.
If you want to disable this behavior:
:reload-clj-files false
Or you can specify which suffixes will cause the reloading
:reload-clj-files #{:clj :cljc}
The name of a file to redirect the figwheel.main logging to. This will only take effect when a REPL has been started.
:log-file "figwheel-main.log"
The level to set figwheel.main java.util.logger to.
Can be one of: :error
:info
:debug
:trace
:all
:off
:log-level :error
The log level to set the client side goog.log.Logger to for
figwheel.repl and figwheel.core. Can be one of:
:severe
:warning
:info
:config
:fine
:finer
:finest
:client-log-level :warning
figwheel.main logging prints out compile time syntax errors which
includes displaying the erroneous code.
Setting :log-syntax-error-style
to :concise
will cause the logging to
not display the erroneous code.
Available options: :verbose
, :concise
Default: :verbose
:log-syntax-error-style :concise
If there are warnings in your code emitted from the compiler, figwheel does not refresh. If you would like Figwheel to load code even if there are warnings generated set this to true. Default: false
:load-warninged-code true
Figwheel makes an effort to provide colorful text output. If you need
to prevent ANSI color codes in figwheel output set :ansi-color-output
to false. Default: true
:ansi-color-output false
Whether to validate the figwheel-main.edn and build config (i.e.".cljs.edn") files. Default: true
:validate-config false
Whether to validate the figwheel-main command line options Default: true
:validate-cli false
A String that specifies the target directory component of the path where figwheel.main outputs compiled ClojureScript
The default :output-dir
is composed of:
[[:target-dir]]/public/cljs-out/[[build-id]]
The default :output-to
is composed of:
[[:target-dir]]/public/cljs-out/[[build-id]]-main.js
If you are using the default figwheel.repl server to serve compiled assets, it is very important that the :target-dir be on the classpath.
The default value of :target-dir
is "target"
:target-dir "cljs-target"
A boolean that indicates whether you want figwheel to automatically launch Node. Defaults to true.
A boolean that indicates whether you want figwheel to enable remote inspection by adding "--inspect" when it launches Node. Defaults to true.
A String indicating the Node.js executable to launch Node with. Defaults to "node"
Figwheel optionally launches a JavaScript host environment when it
starts a REPL or runs a script. You see this behavior when it opens a
browser or starts Nodejs. This behavior can be overridden with the
:launch-js
option.
Can take the name of an executable script on your system and will pass it either the path to the compiled JavaScript (when the target is Nodejs) or the URL to the JavaScript (when the target is the browser).
Script example:
#! /bin/sh
chrome --headless --disable-gpu --repl --remote-debugging-port=9222 $1
If the above script is named headless-chrome-launcher
and is on your
path, then you would add this to your config:
:launch-js `headless-chrome-launcher`
Can also take a vector that represents a shell command to invoke. The
vector can contain the keywords :output-to
and :open-url
which
will be replaced with the the path or the URL to the compiled
JavaScript.
Shell command vector example:
:launch-js ["chrome" "--headless" "--repl" "--disable-gpu" :open-url]
The :launch-js
option can also take a namespaced symbol
representing a function to invoke. The function will be passed a map
containing the keys :open-url
and :output-to
.
Symbol example:
:launch-js user/start-js-environment
and in your user.clj file:
(defn start-js-environment [{:keys [output-to open-url]}]
(clojure.java.shell/sh "headless-chome" open-url))
The :launch-js
option will take precedence over any node
configurations like :node-command
or :launch-node
.
A boolean that indicates whether to include binaryage/devtools into the clojurescript build. Defaults to true when the target is a browser and the :optimizations level is :none, otherwise it is false.
:cljs-devtools false
A boolean that indicates whether figwheel should try and be helpful by adding classpaths to help you get started, or whether you want to have complete control over classpaths. Advanced users will want to disable this option.
:helpful-classpaths false
Support for importing webpack bundles.
Experimental feature! This feature may change or be removed entirely. Only available in
0.1.8-SNAPSHOT
or higher
This also works best with ClojureScript >= 1.10.339
.
Currently takes a map with only one valid key :bundles
. The value of
the :bundles
key must be a map of bundled JavaScript files to the
index JavaScript files that they are compiled from.
:npm {:bundles {"dist/index.bundle.js" "src/webpack/index.js"}}
This feature will read an index.js file like:
import React from 'react';
import ReactDom from 'react-dom';
window.React = React;
window.ReactDom = ReactDom;
and will then generate a :foreign-libs
entry for it. For example the
above index.js would cause the following to be added to your compiler
options:
:foreign-libs [{:file "dist/index.bundle.js"
:provides ["react" "react-dom"]
:global-exports {react React
react-dom ReactDom}}]
This will set :npm-deps
to false
if it hasn't been previously set.
This will set :infer-externs
to true
if it hasn't been
previously set.
You can learn more about ClojureScript and Webpack here: https://clojurescript.org/guides/webpack
A collection of symbol or strings that represent Clojure functions to call just before your ClojureScript sources get built.
These functions will be called before every build.
These functions will be passed the current configuration of the
system. This is a fairly complex data-structure and contains the
:options
for the current build among other things.
:pre-build-hooks [user/gen-testfile]
A collection of symbol or strings that represent clojure functions to call just after your ClojureScript sources have been built.
These functions will be called after every build.
These functions will be passed the current configuration of the
system. This is a fairly complex data-structure and contains the
:options
for the current build among other things.
:post-build-hooks [user/gen-alternate-main-js]
A map of keyword ids to Clojurescript option maps.
:extra-main-files
will output extra main files besides the one that was
configured in your ClojureScript options.
For example this will output a dev-main-test.js file for your tests:
::extra-main-files {:tests {:main example.tests.test-runner}}
This file will be created in addition to the dev-main.js
file.
The options will be merged with the ClojureScript options for the
current build. Keep in mind that this merge supports keywords prefixed
with extra-
when you want the values of these keys to be merged. If
you supply :extra-preloads
in the options map they will be
concatenated with the existing :preloads
. Since
Figwheel works by injecting itself into your config with :preloads
and :closure-defines
it is recommended that you always use
:extra-preloads
and :extra-closure-defines
if you want to change
these values.
This extra main will have all the same configured Figwheel options as the main build. In other words, the extra main will connect to the Figwheel REPL and get reloads just like the main build.
This feature will only output the ClojureScript bootstrap file that
you will require on your host page, it will not cause any files to be
compiled. So you will need to make sure that you have added all the
needed source directories to your :watch-dirs
and your classpath.
This will only work under :optimizations
level :none
.
Figwheel provides a default host page for extra mains so that you do
not have to configure one. The default host page can be found at
/figwheel-extra-main/[id]
where id is the id you supplied as a key
in the config you passed to the :extra-main-files
. For example the
config above you would be able to find the :tests
main at
/figwheel-extra-main/tests
. Keep in mind that the id
of the app
div on the default host page will be app-[id]
.
If you don't want to use the default host page you will need to create a your own host page for it. See https://figwheel.org/docs/your_own_page for help.
This feature is perfect for adding cljs-test-display and devcards to your workflow.
Also keep in mind that you can insert extra behavior with :preloads
and you can even change the :target
to :nodejs
if you want to work
on a Nodejs app in parallel with your main build.
::extra-main-files {:devcards {:main example.devcards}}
Build inputs are passed as the first argument to the CLJS compiler.
Build inputs are normally a list of sources (files and directories) for the compiler to compile.
Figwheel attempts to provide build inputs to the ClojureScript
compiler based on your current configuration. The logic is roughly: if
you are using :optimizations
level :none
and not only building
once, use the :watch-dirs
as the build inputs, otherwise use the
:main
namespace as the build input.
Using the :watch-dirs
as a build input has the advantage that
Figwheel will watch and compile all the source files in the
:watch-dirs
even if they are not required in your application
yet. This allows Figwheel to provide compiler feedback while you are
working on files that are not in your require tree.
When you provide a :build-inputs
in your config you will be
overriding the default Figwheel behavior and be specifing which
specific inputs you want to send to the compiler.
:build-inputs
is a collection of:
strings representing paths to source files and directories
namespace symbols that are on the classpath
the keyword :main
which will be replaced with the namespace in your :main
CLJS option
the keyword :watch-dirs
which will be replaced with your configured :watch-dirs
:build-inputs [:watch-dirs example.core-tests "extra-src"]
Figwheel will automatically discover all the cljs.test based tests
that you have defined and will provide an endpoint to display them
with cljs-test-display
. It will only provide this by default when
the tests are present in your watched directories and a build is using
:optimizations
level :none
.
You can find these tests at the /figwheel-extra-main/auto-testing
HTTP endpoint on the Figwheel server.
Figwheel will automatically find all the namespaces with tests in them.
You can enable this feature by specifying:
:auto-testing true
You can specify which namespaces to test:
:auto-testing {:namespaces [example.core-tests example.logic-tests]}
You can also disable cljs-test-display
with:
:auto-testing {:cljs-test-display false}
When you use libraries from http://cljsjs.github.io they sometimes come bundled with static resources (like CSS files) that you would like to be served from the Figwheel server.
If you set :cljsjs-resources
to true
the Figwheel server will
serve the resources with the cljsjs
root.
For example: if there is a CSS file in the react-vis
jar at
cljsjs/react-vis/common/react-vis.inc.css
you will be able to access
it via the Figwheel server at the path
/react-vis/common/react-vis.inc.css
.
:cljsjs-resources true
The figwheel.repl
client can direct printed (via pr) output to the
REPL and/or the console. :client-print-to
is a list of where you
want print output directed. The output choices are :console
and :repl
Default: [:console :repl]
:client-print-to [:console]
The figwheel server has a notion of a :ring-stack
. The
:ring-stack
is a composition of basic ring-middleware (think
sessions) to wrap around a supplied :ring-handler
.
The default :ring-stack
is a slightly modified
ring.middleware.defaults/wrap-defaults
The figwheel.repl server has a notion of a :ring-stack
. The
:ring-stack
is a composition of basic ring-middleware to wrap around
a supplied :ring-handler
.
The default :ring-stack
is a slightly modified
ring.middleware.defaults/wrap-defaults.
:ring-stack-options
are the options that figwheel.repl supplies to
ring.middleware.defaults/wrap-defaults
.
The default options are slightly modified from ring.middleware.defaults/site-defaults
:
{:params
{:urlencoded true, :multipart true, :nested true, :keywordize true},
:cookies true,
:session
{:flash true, :cookie-attrs {:http-only true, :same-site :strict}},
:static {:resources "public"},
:responses {:content-types true, :default-charset "utf-8"},
:figwheel.server.ring/dev
{:figwheel.server.ring/fix-index-mime-type true,
:figwheel.server.ring/resource-root-index true,
:figwheel.server.ring/wrap-no-cache true,
:figwheel.server.ring/cljsjs-resources false,
:ring.middleware.not-modified/wrap-not-modified true,
:co.deps.ring-etag-middleware/wrap-file-etag true,
:ring.middleware.cors/wrap-cors true,
:ring.middleware.stacktrace/wrap-stacktrace true}}
You can override these options by suppling your own to :ring-stack-options
If these options are changed significantly don't be suprised if the figwheel server stops behaving correctly :)
The number of milliseconds to wait before issuing reloads. Set this higher to wait longer for changes. This is the interval from when the first file change occurs until we finally issue a reload event.
Default: 50
:wait-time-ms 50
The :mode
indicates the behavior that occurs after a compile.
Options: :repl
:serve
or :build-once
:repl
indicates that a REPL will be started:serve
indicates that a server will be started:build-once
indicates that a compile will not be follwed by any actionThis is mainly intended for use when you are launching figwheel.main from a script.
Normally defaults to :repl
Figwheel broadcasts hot reloads to all clients that have connected
since the figwheel process has started. Set :broadcast-reload
to
false
if you want to only send hot-reloads to the client where the
REPL eval occurs.
Default: true
:broadcast-reload false
In the past figwheel would broadcast REPL evaluations to all
connected clients and then print the first result received in the
REPL. Setting :broadcast
to true
will give you back this legacy
behavior. Default: false
:broadcast true
The time (in milliseconds) it takes for the REPL to timeout. Evaluating any given expression in cljs can take some time. The REPL is configured to throw a timeout exception as to not hang forever.
This config option will determine how long the REPL waits for the result of an eval before throwing.
Default: 8000
:repl-eval-timeout 10000 ;;waits for 10 seconds instead of 8
If you need to watch files with polling instead of FS events. This can be useful for certain docker environments.
:hawk-options {:watcher :polling}
Can you improve this documentation?Edit on GitHub
cljdoc is a website building & hosting documentation for Clojure/Script libraries
× close