This image contains everything you need to have a working development environment for ember-cli. The container's working dir is /myapp so that you can setup a volume mapping your project dir to /myapp in the container.
Supported tags and respective
This image was originally based on: geoffreyd/ember-cli (hat tip)
ember-cli 3.1.2 + node 8.11.1/9.11.1 + npm 5.6.0/5.6.0 + bower 1.8.4 + yarn 1.5.1 + chrome 65.0.3325.181 + watchman 4.9.0
Important Change in ember-cli:3.1.1
The test server port exposed in the Dockerfile has changed from 5779 to 7357. I know, they should make up their mind. The default exposed ports for this image will follow along with the default ports used by ember-cli. So, make sure that you now expose that port if you want to hit the test server on its new port:
docker run --rm -ti -v $(pwd):/myapp -p 4200:4200 -p 7020:7020 -p 7357:7357 danlynn/ember-cli:3.1.1 bash root@9ad4805d2b50:/myapp# ember test --server
Important Change in ember-cli:3.0.0
PhantomJS support is not longer supported in Ember 3.0.0. Thus, it has been removed from the docker image. Ember 3.0.0 has moved away from script tag driven development. Thus, it no longer supports Bower in that context. However, your own projects may still be using Bower to pull in dependencies. Therefore, we have retained Bower 1.8.2 in the docker image.
It is notable that Ember 3.0.0 also drops support for IE9 and IE10. Furthermore, IE11 is only supported as a target if ember env (
EMBER_ENV) is 'production' -OR- if the new
CI env var is truthy. Therefore, if you want to support IE11 in 'development' or 'test' modes then include this env var in your docker commands with
--env CI=true like:
docker run --env CI=true --rm -ti -v $(pwd):/myapp -p 4200:4200 -p 7020:7020 -p 5779:5779 danlynn/ember-cli:3.0.0 bash root@9ad4805d2b50:/myapp# ember test
Similarly, you may update your
docker-compose.yml file to include this env var like:
server: image: danlynn/ember-cli:3.0.0 environment: CI: true volumes: - .:/myapp ports: - "4200:4200" - "7020:7020" - "5779:5779"
Important Change in ember-cli:2.17.0
Node has moved the LTS tag from 6.11 to 8.9. Since ember always supports the LTS version of node, we have bumped the version of node on the 'latest' docker hub tag to 8.9. As always, the latest current version of node is also available tagged as '2.17.0-node_9.2'.
Important Change in ember-cli:2.16.2
The test server port 5779 is now exposed in the Dockerfile. In order to be able to hit the test server on your host OS, you will still need to expose the port in either your docker-compose.yml file or in your docker command line - just like you do with the ember server and livereload.
docker run --rm -ti -v $(pwd):/myapp -p 4200:4200 -p 7020:7020 -p 5779:5779 danlynn/ember-cli:2.16.2 bash root@9ad4805d2b50:/myapp# ember test --server
Yarn is now part of the docker image. You can still rely on npm and bower to install your dependencies. But, now you have the choice of using the shiny new yarn instead.
docker run --rm -ti -v $(pwd):/myapp -p 4200:4200 -p 7020:7020 -p 5779:5779 danlynn/ember-cli:2.16.2 bash root@9ad4805d2b50:/myapp# ember init --yarn root@9ad4805d2b50:/myapp# ember install ember-cli-mirage --yarn
Important Change in ember-cli:2.15.0
Since ember-cli 2.15.0 now uses headless chrome by default on new projects when running
ember test, google-chrome is now installed in addition to phantomjs. Existing projects will continue to run tests just fine with their testem.js configured for phantomjs. Also note that we have switched to installing the official phantomjs binary releases.
As of ember-cli:2.15.0, the default 'latest' docker tag will now always use the LTS (long term support) version of node (as recommended by the ember-cli project). From this point forward, there will be the following tagged releases:
latest- using the latest version of ember-cli and the LTS version of node. This version will also be tagged with the LTS node version like:
2.15.0-node_6.11. Additionally, this version will be tagged with the ember-cli version like:
- A second tagged release will be provided with each new ember-cli release that uses the current node release with the latest features. It will be tagged like:
Important Change in ember-cli:2.14.2
As of ember-cli:2.14.2, this image has changed the default 'latest' docker tag to always use the most up-to-date version of node. Previously, 'latest' was stuck at an old "stable" version of node which has proven to be less than stable with newer versions of ember. Thus, we are dropping the old node 4.8.4 support. From now on, the 'latest' image will be using the latest version of node that works well with ember-cli. Older versions of node will be supported in alternate build tags (like 'ember-cli-2.14.2-node_6.11'). Note that this release has also updated watchman from version 3.5.0 to 4.7.0. However, phantomjs has dropped back to 2.1.1 for node 8.2.1 since the latest version has build issues with the latest node.
Important Change in ember-cli:2.11.1
As of ember-cli:2.11.1, ember-cli livereload listens on port 7020 by default insteadof 49152. Dockerfile now exposes 49153.
Important Change in ember-cli:2.10.0
As of ember-cli:2.10.0, this image has changed its
ENTRYPOINT to the docker convention of
/bin/sh -c. This means that if you have any shell scripts which assume the previous
/usr/local/bin/ember then you will need to update them. The default CMD is now
ember server which means that the syntax for most critical action of launching the server is unchanged. However, other actions may need to be updated.
Example Changes To Common Actions
Launch ember server (unchanged):
OLD: $ docker run -ti --rm -v $(pwd):/myapp -p 4200:4200 -p 7020:7020 danlynn/ember-cli:2.9.1 NEW: $ docker run -ti --rm -v $(pwd):/myapp -p 4200:4200 -p 7020:7020 danlynn/ember-cli:3.1.2
Run ember tests:
OLD: $ docker run -ti --rm -v $(pwd):/myapp danlynn/ember-cli:2.9.1 test NEW: $ docker run -ti --rm -v $(pwd):/myapp danlynn/ember-cli:3.1.2 ember test
Launch bash shell:
OLD: $ docker run -ti --rm -v $(pwd):/myapp --entrypoint=/bin/bash danlynn/ember-cli:2.9.1 NEW: $ docker run -ti --rm -v $(pwd):/myapp danlynn/ember-cli:3.1.2 bash
How to use
Create new project dir and add a docker-compose.yml file similar to the following:
ember: &defaults image: danlynn/ember-cli:3.1.2 volumes: - .:/myapp npm: <<: *defaults command: npm bower: <<: *defaults command: bower --allow-root server: <<: *defaults command: server --watcher polling ports: - "4200:4200" - "7020:7020" - "7357:7357"
Make sure that your docker-machine is already running:
$ docker-machine start default
Or, if you haven't created one yet:
$ docker-machine create --driver virtualbox default
Create an ember app in the current dir:
$ docker-compose run --rm ember init
Start the ember server:
$ docker-compose up
This launches the ember-cli server on port 4200 in the docker container. As you make changes to the ember webapp files, they will automagically be detected and the associated files will be recompiled and the browser will auto-reload showing the changes.
Note that if you get an error something like
server_1 | Error: A non-recoverable condition has triggered. Watchman needs your help! server_1 | The triggering condition was at timestamp=1450119416: inotify-add-watch(/myapp/node_modules/ember-cli/node_modules/bower/node_modules/update-notifier/node_modules/latest-version/node_modules/package-json/node_modules/got/node_modules/duplexify/node_modules/readable-stream/doc) -> The user limit on the total number of inotify watches was reached; increase the fs.inotify.max_user_watches sysctl server_1 | All requests will continue to fail with this message until you resolve server_1 | the underlying problem. You will find more information on fixing this at server_1 | https://facebook.github.io/watchman/docs/troubleshooting.html#poison-inotify-add-watch
Then watchman is running out of resources trying to track all the files in a large ember app. To increase the
fs.inotify.max_user_watchescount to something that is more appropriate for an ember app, stop your docker-compose server by hitting ctrl-c (or
docker-compose stop serverif necessary) then execute the following command:
$ docker run --rm --privileged danlynn/ember-cli:3.1.2 sysctl -w fs.inotify.max_user_watches=524288
Note that this will affect all containers that run on the current docker-machine from this point forward because
fs.inotify.max_user_watchesis a system-wide setting. This shouldn't be a big deal however, so go ahead and give it a try. Then start the docker-compose service again with
$ docker-compose up
Launch the ember webapp:
You will need to first determine the IP of the docker container:
$ docker-machine ip default -or- $ boot2docker ip 192.168.59.103
Next open that ip address in your browser on port 4200:
Command Usage for
The ember, bower, and npm commands can be executed in the container to effect changes to your local project dir as follows. You basically put a "docker-compose run --rm" in front of any of the 3 commands and pass the normal command options as usual.
$ docker-compose run --rm npm install $ docker-compose run --rm bower install bootstrap $ docker-compose run --rm ember generate model user
Note that the
--rm prevents any changes outside of your project dir (/myapp in the container) from being persisted by docker-compose. Usually don't care about anything outside of your project dir.
Command Usage for
docker run --rm -ti -v $(pwd):/myapp danlynn/ember-cli:3.1.2 before any command you run.
$ docker run --rm -ti -v $(pwd):/myapp danlynn/ember-cli:3.1.2 npm install $ docker run --rm -ti -v $(pwd):/myapp danlynn/ember-cli:3.1.2 bower --allow-root install bootstrap $ docker run --rm -ti -v $(pwd):/myapp danlynn/ember-cli:3.1.2 ember generate model user $ docker run --rm -ti -v $(pwd):/myapp -p 4200:4200 -p 7020:7020 -p 7357:7357 danlynn/ember-cli:3.1.2
Note that the
--rm prevents a bunch of stopped containers from accumulating from these one-off commands. They take up space and since pretty much any change made by these commands will only affect what is in your project dir (/myapp in the container), there is no need to keep them around.
Alternatively, you could simply launch into a bash shell and execute the commands in the normal fashion:
$ mkdir new_ember_app $ cd new_ember_app $ docker run --rm -it -v $(pwd):/myapp -p 4200:4200 -p 7020:7020 -p 7357:7357 danlynn/ember-cli:3.1.2 bash root@9ad4805d2b50:/myapp# ember init root@9ad4805d2b50:/myapp# ember init --yarn root@9ad4805d2b50:/myapp# npm install root@9ad4805d2b50:/myapp# bower --allow-root install root@9ad4805d2b50:/myapp# ember server root@9ad4805d2b50:/myapp# ember test root@9ad4805d2b50:/myapp# ember test --server
Note that bash had to be launched with
-p 4200:4200 -p 7020:7020 in order to be able to access the
ember server on port 4200 and enable Livereload on port 7020. The
-p 7357:7357 is needed if you intend to run
ember test --server.
Also note that the
npm install is done automagically by the
ember init command on newer versions of ember. Also,
bower --allow-root install is not used as much anymore. There are no bower packages or dependencies in the default project created by
ember init. Using the
--yarn option on
ember init --yarn will use yarn instead of npm to install dependencies.
Also note that both npm and bower are pretty much being replaced by yarn. Newer versions of ember-cli have built-in support for yarn on many commands. Yarn works smoothly in place of npm. Yarn can also replace your use of bower. However, even though yarn used to support bower file formats directly, it no longer does. You should instead use yarn's support for installing front-end web components.