Figwheel builds your ClojureScript code and hot loads it into the browser as you are coding!
Get a quick idea of what figwheel does by watching the 6 minute flappy bird demo of figwheel.
Learn even more by watching a 45 minute talk on Figwheel given at ClojureWest 2015.
Read the introductory blog post.
If you write reloadable code, figwheel can facilitate automated live interactive programming. Every time you save your ClojureScript source file, the changes are sent to the browser so that you can see the effects of modifying your code in real time.
You can use figwheel to live code ClojureScript in Node.js!
The inclusion of a static file server allows you to get a decent
ClojureScript development environment up and running quickly. For
convenience there is a :ring-handler
option so you can load a ring
handler into the figwheel server.
Figwheel will reload your CSS live as well.
Figwheel can live reload your JavaScript source files.
Figwheel has a non-intrusive heads up display that gives you feedback on how well your project is compiling. By writing a shell script you can click on files in the heads up display and they will open in your editor!
When you launch figwheel it not only starts a live building/reloading process but it also optionally launches a CLJS REPL into your running application. This REPL shares compilation information with the figwheel builder, so as you change your code the REPL is also aware of the code changes. The REPL also has some special built-in control functions that allow you to control the auto-building process and execute various build tasks without having to stop and rerun lein-figwheel.
Figwheel's connection is fairly robust. I have experienced figwheel
sessions that have lasted for days through multiple OS sleeps. You can
also use figwheel like a REPL if you are OK with using print
to output
the evaluation results to the browser console.
Figwheel broadcasts changes to all connected clients. This means you can see code and CSS changes take place in real time on your phone and in your laptop browser simultaneously.
Figwheel will not load a file that has not been required. It will also respond well to new requirements and dependency tree changes.
Figwheel does its best to only reload what needs to be reloaded. This minimizes the surface area of dynamically reloaded code, which in turn should increase the stability of the client environment.
If your ClojureScript code is generating compiler warnings Figwheel won't load it. This, again, is very helpful in keeping the client environment stable. This behavior is optional and can be turned off.
My cofounder, Danny, and I can't imagine working without Figwheel. We have been using it since December to build Precursor, a collaborative drawing tool. It has saved us a ton of time and contributed a lot towards Danny's appreciation of ClojureScript for design.
Danny King and Daniel Woelfel
My major issue is that Figwheel is too awesome.
Julian Leviston
Make sure you have the latest version of leiningen installed.
You can try figwheel out quickly with the flappy bird demo:
git clone https://github.com/bhauman/flappy-bird-demo.git
then cd into flappy-bird-demo
and type
lein figwheel
You can now goto localhost:3449/index.html
and open up
src/flappy_bird_demo/core.cljs
with your favorite editor and start
coding. Make sure you open your browser's development console so you
can get feedback about code reloads.
If you would prefer to greenfield a new project you can use the figwheel leiningen template.
lein new figwheel hello-world
Or optionally:
lein new figwheel hello-world -- --om ;; for an om based project
lein new figwheel hello-world -- --reagent ;; for a reagent based project
If you are brand new to ClojureScript it is highly recommended that you do the ClojureScript Quick Start first. If you skip this you will probably suffer.
There is a lot to learn when you are first learning ClojureScript, I recommend that you bite off very small pieces at first. Smaller bites than you would take when learning other languages like JavaScript and Ruby.
Please don't invest too much time trying to set up a sweet development environment, there is a diverse set of tools that is constantly in flux and it's very difficult to suss out which ones will actually help you. If you spend a lot of time evaluating all these options it can become very frustrating. If you wait a while, and use simple tools you will have much more fun actually using the language itself.
If you are new to Figwheel here is a Quick Start tutorial. Working through this Quick Start will probably save you a tremendous amount of time.
First make sure you include the following :dependencies
in your project.clj
file.
[org.clojure/clojure "1.7.0"]
[org.clojure/clojurescript "1.7.170"]
Then include lein-figwheel
in the :plugins
section of your project.clj.
[lein-figwheel "0.5.0-2"]
You also need to have your :cljsbuild
configuration set up in your
project.clj
.
Here is an example:
:cljsbuild {
:builds [ { :id "example"
:source-paths ["src/"]
:figwheel true
:compiler { :main "example.core"
:asset-path "js/out"
:output-to "resources/public/js/example.js"
:output-dir "resources/public/js/out" } } ]
}
The important part here is that you have to have at least one build
that has :optimizations
set to :none
or nil
.
If you leave out the :optimizations
key the ClojureScript compiler
will default to :none
.
Setting :figwheel true
or :figwheel { :on-jsload "example.core/reload-hook" }
will
automagically insert the figwheel client code into your application.
If you supply :on-jsload
the name of a function, that function will
be called after new code gets reloaded.
If you want to serve the HTML file that will host your application
from figwheel's built in server, then the output directory has to be
in a directory that can be served by the static webserver. The default
for the webserver root is "resources/public" so your output files need
to be in a subdirectory "resources/public" unless you change the
webserver root. For now the webserver root has to be in a subdirectory
of resources
.
If you are serving your application HTML from your own server you can
configure :output-to
and :output-dir
as you like.
Start the figwheel server. (This will get the first :optimizations
:none
build)
$ lein figwheel
or optionally give the name of the build
$ lein figwheel example
This will start a server at http://localhost:3449
with your
resources being served via the compojure resources
ring handler.
So you can load the HTML file thats hosting your ClojureScript app
by going to http://localhost:3449/<yourfilename>.html
If you are using your own server please load your app from that server.
This is not neccessary but you can configure the figwheel system. At
the root level of your project.clj
you can add the following server
side configuration parameters:
:figwheel {
:http-server-root "public" ;; this will be in resources/
:server-port 3449 ;; default
:server-ip "0.0.0.0" ;; default
;; CSS reloading (optional)
;; :css-dirs has no default value
;; if :css-dirs is set figwheel will detect css file changes and
;; send them to the browser
:css-dirs ["resources/public/css"]
;; Server Ring Handler (optional)
;; if you want to embed a ring handler into the figwheel http-kit
;; server
:ring-handler example.server/handler
;; Clojure Macro reloading
;; disable clj file reloading
; :reload-clj-files false
;; or specify which suffixes will cause the reloading
; :reload-clj-files {:clj true :cljc false}
;; To be able to open files in your editor from the heads up display
;; you will need to put a script on your path.
;; that script will have to take a file path and a line number
;; ie. in ~/bin/myfile-opener
;; #! /bin/sh
;; emacsclient -n +$2 $1
;;
:open-file-command "myfile-opener"
;; if you want to disable the REPL
;; :repl false
;; to configure a different figwheel logfile path
;; :server-logfile "tmp/logs/figwheel-logfile.log"
;; Start an nREPL server into the running figwheel process
;; :nrepl-port 7888
;; Load CIDER, refactor-nrepl and piggieback middleware
;; :nrepl-middleware ["cider.nrepl/cider-middleware"
;; "refactor-nrepl.middleware/wrap-refactor"
;; "cemerick.piggieback/wrap-cljs-repl"]
}
Make sure you have setup an html file to host your cljs. For example
you can create this resources/public/index.html
file:
<!DOCTYPE html>
<html>
<head>
</head>
<body>
<div id="main-area">
</div>
<script src="js/example.js" type="text/javascript"></script>
</body>
</html>
Instead of setting :figwheel true
in your cljsbuild configuration
you can pass a map of options as below:
:cljsbuild {
:builds [ { :id "example"
:source-paths ["src/"]
;; put client config options in :figwheel
:figwheel { :websocket-host "localhost"
:on-jsload "example.core/fig-reload"}
:compiler { :main "example.core"
:asset-path "js/out"
:output-to "resources/public/js/example.js"
:output-dir "resources/public/js/out"
:optimizations :none } } ]
}
The following configuration options are available:
;; configure a websocket host, figwheel already knows the port
;; this is helpful if you want to broadcast to devices
:websocket-host "localhost" ;; or "www.myhost.com", "192.168.0.112"
;; An important configuration option for :websocket-host
;; if you set it to :js-client-host it will set the host based on the
;; js/location.host param of the browser
;; This is helpful for multiple device settings where you are using
;; fighweel to serve your app.
;; :websocket-host :js-client-host
;; optional callback
:on-jsload "example.core/fig-reload"
;; if you want to do REPL based development and not have
;; have compiled files autoloaded into the client env
:autoload false
;; The heads up display is enabled by default to disable it:
:heads-up-display false
;; when the compiler emits warnings figwheel blocks the loading of files.
;; To disable this behavior:
:load-warninged-code true
Whole files will be reloaded on change so we have to make sure that we write reloadable code.
Please check out the example project in the example
directory.
To see all the client side config options look here.
Figwheel normally reloads any file that has changed. If you want to prevent certain files from being reloaded by figwheel, you can add meta-data to the namespace declaration like so:
(ns ^:figwheel-no-load example.core)
Figwheel will not load or reload files that haven't been required by your application. If you want to force a file to be loaded when it changes add the follwoing meta-data the namespace declaration of the file:
(ns ^:figwheel-load example.core)
It can be very helpful to have a file reload every time a file changes in your ClojureScript source tree. This can facilitate reloading your main app and running tests on change.
To force a file to reload on every change:
(ns ^:figwheel-always example.test-runner)
When you run lein figwheel
a REPL will be launched into your application.
You will need to open your application in a browser in order for the REPL to connect and show its prompt.
This REPL is a little different than other REPLs in that it has live
compile information from the build process. This effectively means
that you will not have to call (require
or (load-namesapce
unless
it is a namespace that isn't in your loaded application's required
dependencies. In many cases you can just (in-ns 'my.namespace)
and
everything you need to access will be there already.
The REPL doesn't currently have built-in readline support. To have a
better experience please install rlwrap. You can do this on OSX
using brew: brew install rlwrap
.
When rlwrap
is installed you can now execute lein figwheel as so:
$ rlwrap lein figwheel
This will give you a much nicer REPL experience with history and line editing.
The Figwheel REPL has the following control functions:
Figwheel Controls:
(stop-autobuild) ;; stops Figwheel autobuilder
(start-autobuild [id ...]) ;; starts autobuilder focused on optional ids
(switch-to-build id ...) ;; switches autobuilder to different build
(reset-autobuild) ;; stops, cleans, and starts autobuilder
(build-once [id ...]) ;; builds source one time
(clean-builds [id ..]) ;; deletes compiled cljs target files
(fig-status) ;; displays current state of system
These functions are special functions that poke through the ClojureScript env into the underlying Clojure process. As such you can't compose them.
You can think of these functions having an implicit set of build ids that they operate on.
If you call (reset-autobuild)
it will stop the figwheel autobuilder,
clean the builds, reload the build configuration from your
project.clj
and then restart the autobuild process.
If you call (stop-autobuild)
it will stop the figwheel autobuilder.
If you call (start-autobuild)
it will start the figwheel autobuilder
with the current implicit build ids.
If you call (start-autobuild example)
it will start the figwheel
autobuilder on the provided build id example
. It will also make
[example]
the implicit set of build ids.
start-autobuild
and switch-to-build
are the only functions that
update the build-id set.
clean-builds
and build-once
both allow you to do one off builds and
cleans. They do not alter the implicit build ids.
fig-status
displays information on the current Figwheel system state,
including whether the autobuilder is running, which build ids are in
focus, and the number of client connections.
You may want a REPL in your editor. This makes it much easier to ship code from your buffer to be evaluated.
If you use
lein repl
or something that invokes it like CIDER, you are using nREPL. A ClojureScript REPL will not just run over an nREPL connection without Piggieback.
If you are just starting out I would use the Figwheel console REPL because it's aready set up and ready to go, complexity conquered!
If you want to integrate a REPL into your editor, here are my top recommendations:
Emacs:
- use
inf-clojure
as described on the wiki page - alternatively use Cider and nREPL. Using the ClojureScript REPL over an nREPL connection is considered advanced
Cursive: use the instructions on the wiki page
Vi: use tmux
mode to interact with the figwheel REPL, still trying to get a wiki page for this if you can help that would be great
If you are going to use nREPL with Figwheel please see:
As your development workflow grows in complexity, the declarative
approach of lein
can be limiting when you want to launch and control
different services (ie. SASS compilation). It is really helpful to use
Clojure itself to script whatever workflow services you want.
Figwheel has a Clojure API that makes it easy to start, stop and control Figwheel from Clojure.
In order for the following examples to work, you will need to have
[figwheel-sidecar "0.5.0-2"]
in your dependencies.
To start Figwheel from a script, you will need to require the
figwheel-sidecar.repl-api
and provide your build configuration to
figwheel-sidecar.repl-api/start-figwheel!
like so:
(require '[figwheel-sidecar.repl-api :as ra])
;; this will start figwheel and will start autocompiling the builds specified in `:builds-ids`
(ra/start-figwheel!
{:figwheel-options {} ;; <-- figwheel server config goes here
:build-ids ["dev"] ;; <-- a vector of build ids to start autobuilding
:all-builds ;; <-- supply your build configs here
[{:id "dev"
:figwheel true
:source-paths ["src"]
:compiler {:main "example.core"
:asset-path "out"
:output-to "resources/public/main.js"
:output-dir "resources/public/out"
:verbose true}}]})
;; you can also just call (ra/start-figwheel!)
;; and figwheel will do its best to get your config from the
;; project.clj or a figwheel.edn file
;; start a ClojureScript REPL
(ra/cljs-repl)
;; you can optionally supply a build id
;; (ra/cljs-repl "dev")
Build config notes
It's important to remember that figwheel can autobuild and reload multiple builds at the same time. It can also switch between builds and focus on autobuilding one at a time. For this reason you need to supply the initial
:build-ids
to tell figwheel which builds you want to start building. It's also really helpful to supply your:advanced
builds because while you can't autobuild them you can callbuild-once
on them
Assuming the above script is in script/figwheel.clj
you can invoke it as follows:
$ rlwrap lein run -m clojure.main script/figwheel.clj
The above command will start figwheel and it will behave just like
running lein figwheel
.
Please note that the above command is not running the script in the
same environment as lein repl
or cider-jack-in
. Both of these
start an nREPL session. I am intentionally not using nREPL in order to
remove a lot of complexity from ClojureScript REPL communication.
If you are using nREPL, launching the ClojureScript REPL requires that you have Piggieback installed. Please see the section above titled "Editor REPLs and nREPL"
Let's make a small helper library and then initialize a Clojure REPL with it:
(require
'[figwheel-sidecar.repl-api :as ra])
(defn start []
(ra/start-figwheel!
{:figwheel-options {} ;; <-- figwheel server config goes here
:build-ids ["dev"] ;; <-- a vector of build ids to start autobuilding
:all-builds ;; <-- supply your build configs here
[{:id "dev"
:figwheel true
:source-paths ["src"]
:compiler {:main "example.core"
:asset-path "out"
:output-to "resources/public/main.js"
:output-dir "resources/public/out"
:verbose true}}]}))
;; Please note that when you stop the Figwheel Server http-kit throws
;; a java.util.concurrent.RejectedExecutionException, this is expected
(defn stop []
(ra/stop-figwheel!))
(defn repl []
(ra/cljs-repl))
The next line will call clojure.main
and initialize it with our
script and then continue on to launch a REPL.
$ rlwrap lein run -m clojure.main --init script/figwheel.clj -r
After the Clojure REPL has launched, you will now have the ability to
call (start)
, (repl)
and (stop)
as you need.
You can also call all of the functions in the figwheel-sidecar.repl-api.
This is a powerful way to work, as you now have the interactivity and generality of the Clojure programming language available.
Need to start a server? Go for it.
Need to watch and compile SASS files? No problem.
I highly recommend Stuart Sierra's component library to compose all the services you will need in your development process.
Here is an example of creating a Figwheel component and composing it with a Ring server component to serve your application.
(require
'[figwheel-sidecar.repl-api :as ra]
'[com.stuartsierra.component :as component]
'[ring.component.jetty :refer [jetty-server]])
(def figwheel-config
{:figwheel-options {} ;; <-- figwheel server config goes here
:build-ids ["dev"] ;; <-- a vector of build ids to start autobuilding
:all-builds ;; <-- supply your build configs here
[{:id "dev"
:figwheel true
:source-paths ["src/main"]
:compiler {:main "example.core"
:asset-path "/out"
:output-to "resources/public/main.js"
:output-dir "resources/public/out"
:verbose true}}]})
(defrecord Figwheel []
component/Lifecycle
(start [config]
(ra/start-figwheel! config)
config)
(stop [config]
;; you may want to restart other components but not Figwheel
;; consider commenting out this next line if that is the case
(ra/stop-figwheel!)
config))
(defn handler [request]
{:status 200
:headers {"Content-Type" "text/plain"}
:body "Hello World"})
(def system
(atom
(component/system-map
:app-server (jetty-server {:app {:handler handler}, :port 3000})
:figwheel (map->Figwheel figwheel-config))))
(defn start []
(swap! system component/start))
(defn stop []
(swap! system component/stop))
(defn reload []
(stop)
(start))
(defn repl []
(ra/cljs-repl))
Again you can run this script as so:
$ rlwrap lein run -m clojure.main --init script/figwheel.clj -r
As you can see with humble beginnings you can build up arbitrary functionality.
Please see Daniel Szmulewicz's excellent system which is a set of helpful components
If you are using nREPL, launching the ClojureScript REPL requires that you have Piggieback installed. Please see the section above titled "Editor REPLs and nREPL"
Please note that when you stop the Figwheel server, http-kit throws a
java.util.concurrent.RejectedExecutionException
, this is expected
Read more about the clojure.main
command line options
Read more about component
This plugin starts a ClojureScript auto builder, opens a websocket and starts static file server. When you save a ClojureScript file, Figwheel will detect that and compile it and other affected files. It will then pass a list of those changed files off to the figwheel server. The figwheel server will in turn push the paths of the relevant compiled javascript files through a websocket so that the browser can reload them.
The main motivation for lein-figwheel is to allow for the interactive development of ClojureScript. Figwheel doesn't provide this out of the box, the developer has to take care to make their code reloadable.
Figwheel relies on having files that can be reloaded.
Reloading works beautifully on referentially transparent code and code that only defines behavior without bundling state with the behavior.
If you are using React or Om it's not hard to write reloadable code, in fact you might be doing it already.
There are several coding patterns to look out for when writing reloadable code.
One problematic pattern is top level definitions that have local state.
(def state (atom {}))
The state
definition above is holding an atom that has local state.
Every time the file that holds this definition gets reloaded the state
definition will be redefined and the state it holds will be reset back
to the original state. But with figwheel we are wanting to change our
programs while maintaining the state of the running program.
The way to fix this is to use defonce
(defonce state (atom {}))
This will fix most situations where you have code that is relying on a
definition that has local state. Keep in mind though that if you
change the code that is wrapped in a defonce
you won't see the
changes, because the identifier won't be redefined.
Complicated object networks wired together with callbacks (Backbone, Ember, etc.) are also problematic. Instantiating these object callback networks and then storing them in a global var is yet another version of this problem.
Functions that maintain local state like counters and such are also definitions with local state, and as such are problematic.
You also need to look out for common setup code that hooks into the browser.
Often you will see statements like this at the bottom of a file.
(.click ($ "a.button") (fn [e] (print "clicked button")))
Every time this file gets loaded a new listener will get added to all the anchor tags with a "button" class. This is obviously not what we want to happen.
This code is very problematic and points to the why using the browser APIs directly has always been really difficult. For instance if we make it so that these hooks are only executed once, like so:
(defonce setup-stuff
(do
(.click ($ "a.button") (fn [e] (print "clicked button")))))
When you are live editing code, this doesn't work very well. If you alter your HTML template any new "a.button" elements aren't going to have the listener bound to them.
You can fix this by using an event delegation strategy as so:
(defonce setup-stuff
(do
(.on ($ "div#app") "click" "a.button" (fn [e] (print "clicked button")))))
But even with the above strategy you won't be able to edit any of the code in the setup up block and see your changes take affect.
If you are not using React and you want to build things this way and
have reloadable code we need to create setup
and teardown
functions to be invoked on code reload.
(defn setup []
(.on ($ "div#app") "click" "a.button" (fn [e] (print "clicked button"))))
(defn teardown []
(.off ($ "div#app") "click" "a.button")
;; define a reload hook in the
(defn fig-reload-hook []
(teardown)
(setup))
Now you can edit the code in the setup and teardown functions and see the resulting changes in your application.
In a way you can think of the previous definitions of setup-stuff
as
functions that have local state of sorts. They are altering and storing
callbacks in the DOM directly and this is why it is so problematic.
This is one of the reasons React is so damn brilliant. You never end
up storing things directly in the DOM. State is mediated and managed
for you. You just describe what should be there and then React takes
care of making the appropriate changes. For this reason React is a
prime candidate for writing reloadable code. React components already
have a lifecycle protocol that embeds setup
and teardown
in each
component and invokes them when neccessary.
It is worth repeating that React components don't have local state, it just looks like they do. You have to ask for the local state and React in turn looks this state up in a larger state context and returns it, very similar to a State Monad.
Reloadable code is easy to write if we are very conscious and careful about the storage of state, state transitions and side effects. Since a great deal of programming complexity stems from complex interactions (side effecting events) between things that have local state, it is my belief that reloadable code is often simply better code.
If you want to do less thinking and write more reliable front end code you should really be looking at React, Om etc.
OK enough.
Copyright © 2014 Bruce Hauman
Distributed under the Eclipse Public License either version 1.0 or any later version.