Create and maintain your own bower package




  1. a public git(github/gitlab/bitbucket) repo (for private repos, you need to run a private registry, which won’t be discussed in this article)
  2. have bower installed


A step by step example

1.Have a public git repo ready

I have this really simple sample gitlab repo:

This pic shows its file structure:

I want to create a bower package with everything inside “css” and “js” folders. So my bower.json is like this:

  "name": "christy-bower-package",
  "description": "Christy's bower package example",
  "keywords": [
  "license": "MIT",
  "ignore": [
  "dependencies": {
    "jquery": "1.9.1 - 3",
    "bootstrap": "3.3.7"

If you have questions on what each field means, please read bower specs:


2. register the package

2.1 Run command to register package

The command to register is:

bower register <my-package-name> <git-endpoint>

So I run:

bower register christy-bower-package

The following is what happens:

2.2 Check successful registration

Bower decides package version based on tags. Since I haven’t pushed any tag to my repo yet. As you can see in the following picture, when I run “bower info christy-bower-package”, it shows at the end, “No versions available”:


3. Add  version to the package

3.1 create a git tag and push to remote

git tag 'v0.0.1'
git push origin master --tag

3.2 Check version successfully added

bower info christy-bower-package


4. Install the package

Now If I run:

bundle install christy-bower-package

As you can see in the above picture. It downloads my christy-bower-package, also the two denpendency I listed out in bower.json: bootstrap and jquery. The downloaded file structure is like this:


5. Maintain the package

Any time I want to release a new version of my package, I can just add a tag with increasing version number. But it’s a bit of manual work, luckily, there’s a grunt plugin(grunt-bump) that helps automate this process:

5.1 Create a package.json file with content: {}


5.2 Install grunt and then install grunt-bump

npm install grunt --save-dev
npm install grunt-bump --save-dev

After successful installation, a node_modules folder will be added, and Package.json looks like this:


5.3 Add a .gitignore file and ignore node_modules folder

Since I don’t want node_modules folder appear in git repo. Add the following .gitignore file:


5.4 Set custom grunt bump configuration and load the plugin:

The default config for grunt bump is:

Since I only want to change some of the config properties, the following is my Grutfile.js:

module.exports = function(grunt) {
    bump: {
      options: {
        files: ['bower.json'],
        commitFiles: ['-a'],
        pushTo: 'origin'


5.4 Add version field to bower.json

Add version field to bower.json so that grunt bump plugin knows which version to start with:


5.5 Use grunt bump to commit changes and bump the version

grunt bump provides a set of commands to control version. The following is a screenshot of some of the commands available. Please visit its git repo to learn more.

The “grunt bump” command will do the following:
1.Look into “bower.json” file for “version” field. And the value will be the starting version.
2.It will check all files in the repo, commit current file changes. And add a commit message ‘Release v%VERSION%’.
3.It will create a tag with specified tag name and tag message.
4.It will push to remote ‘origin’.


grunt bump


5.7 Install the package again and confirm that version has been bumped


bower install christy-bower-package

Rails add a search function controller and route best practice

1.Task description

I worked on a Ruby on Rails app, and one day I got a task to add this new feature:

Create a search page(So user can search items by category).

After hitting the “search” button, it takes the user to item list page.


For simplicity and to avoid dragging  myself away from the main topic I want to talk about:

The search page could simply look like this:


The item list page could just look like this:



2.My initial approach

2.1 Strategy:

Q: How to render the search page?

A: Create a search action.

Q: How to perform the search function?

A: Create a index action, which handles the search function. The search form should submit a post request to this action.

Q: How to render the item list page?

A: Let index action render item list page.

2.2 Implementation:



get ‘items/search’, to: ‘items#search’
post ‘items/index’, to: ‘items#index’




class ItemsController < ApplicationController
def index
# perform search code goes here
render ‘index’

def search
render ‘search’



<%= form_tag items_index_path, method: ‘post’ do %>
<input type=”text” name=”category”>
<% end %>

2.3 Implementation in action


after hitting the search button:


2.4 Problem

This works perfectly, until I manually refreshed the index page.


Since there’s no get route, it errors out! Of course I could add a get route, but in that way, I will have to store the form parameters in session. Is there another approach?



3. Second approach

3.1 Strategy:

Q: How to render the search page?

A: Create a search action.

Q: How to perform the search function?

A: Create a index action, which handles the search function. The search form should submit a get request to this action, so that the search parameters are in the URL, so that I don’t need to worry about storing them in session.

Q: How to render the item list page?

A: Let index action render item list page.

A note: 

In the future, there might be other tasks to add view item detail, edit item functions and on. So it’s better  use rails default resource:

3.2 Implementation:



get ‘items/search’, to: ‘items#search’
resources :items, only: [:index] do



The same as first approach.




<%= form_tag items_path, method: ‘get’ do %>
<input type=”text” name=”category”>
<% end %>

3.3 Implementation in action


after hitting the search button:



3.4 Concern

We can see the parameters are passed in the URL, and I’m good even though I refresh.

This approach definitely works. The only concern is that it brought in a search action, which is not part of rails default resource; and the search page and item list page are on different URL.

I want to utilize rails default resource, and keep search and item list page on the same URL.

So here comes the third approach.



4. Third approach

4.1 Strategy:

Q: How to render the search page?

A: Use index action. Check if user is not performing a search, render search page. (Trick is add a hidden input in the form)

Q: How to perform the search function?

A: Perform it in index action.

Q: How to render the item list page?

A: Use index action. Check if user is performing a search, render index page.



4.2 Implementation:



resources :items, only: [:index] do


class ItemsController < ApplicationController
def index
if params[:search].blank?
render ‘search’
# perform search code goes here
render ‘index’



<%= form_tag items_path, method: ‘get’ do %>

<input type=”hidden” name=”search” value=”true”>
<input type=”text” name=”category”>
<% end %>

3.3 Implementation in action


after hitting the search button:




Deploy rails app with GoCD and Capistrano simplified log file

Clean work directory   “/var/lib/go-agent/pipelines/ROMUI_DEV_DEPLOY”

Create git artifact    “pipelines/ROMUI_Dev_Deploy/Configuration”

Set GoCD ENV variable

Start to build task    ROMUI_Dev_Deploy/649/dev_deployment/1/ha-deploy-runOnAll-1 on  [/var/lib/go-agent]

    Fetch Artifact from CI    

  • execute task: <fetchartifact pipeline=”ROMUI_CI_Deploy” stage=”ci_deployment” job=”run_tests” srcdir=”romui”/>
  • fetch artifact [romui] from [ROMUI_CI_Deploy/675/ci_deployment/run_tests
  • save artifact to [pipelines/ROMUI_Dev_Deploy]

Fetch Artifact from config repo

     bundle install 

bundle exec cap development_integration deploy target=localhost stack=04 branch=develop

  • Run passenger: running /usr/bin/env which passenger as rails@localhost
  • Run RVM:
  •       running /usr/bin/env [-d~/.rvm] as rails@localhost
  •       command: /usr/local/rvm/bin/rvm   default do ruby –version
  • Check git access:  command: (export GIT_ASKPASS=”bin/echo” GIT_SSH=”/tmp/romui/”; /usr/bin/env git ls-remote –heads
  • INFO The repository mirror is at /rails/apps/deploy/repo
  • command: /usr/bin/env  ln -s /rails/apps/deploy/shared/vendor/bunlde   /rails/apps/deploy/releases/20160728145958/vendor/bundle
  • command: ln -nfs /rails/apps/config/current/.env   /rails/apps/deploy/releases/20160728145958/.env
  • Run ‘rake db:migrate’
  •      command: cd /rails/apps/deploy/releases/20160728145958 && (export RAILS_ENV=”development_integration”; /usr/local/rvm/bin/rvm    default to bundle exec rake db:migrate)
  • Restart app
  •      running /usr/bin/env passenger_config restart_app /rails/apps/deploy –ignore -app-not-running as rails@localhost
  •      command: cd /rails/apps/deploy/releases/20160728145958 && /usr/bin/env passenger-config restart-app /rails/apps/deploy –ignore-app-not-running
  •       restarting /rails/apps/deploy/current
  • Clean up
  •     INFO: keeping 5 of 6 deployed releases on localhost
  •     running /usr/local/rvm/bin/rvm    default to bundle exec rake tmp:clear as rails@localhost


Capistrano deploy flow outline

Official flow reference:

An example flow:

Deploy: starting

Rbenv ensures that the version we configured on config/deploy.rb is installed and that it can write on disc.

Checks git repository

Create needed folder tree








Deploy: updating

Clone the whole repo in /var/www/app_name/repo

Creates a new folder date-time named on /var/www/app_name/releases/20140717180754

Copies the project there


Deploy: publishing

Runs bundle install in last release folder

Runs assets:precompile copying our minimized assets to shared/public/assets

Creates a soft link from last release to /var/www/app_name/current

Restar server.



Logs deployment to revisions.log

Capistrano commands

cap deploy:setup

This will SSH to your server and create some directories in the folder you specified with deploy_to, where Capistrano will store your releases, and your shared files (e.g., logs, configs, static assets like updates, etc.). If something goes wrong with permission or SSH access, you’ll see some error messages. Fix these before proceeding so you know you can actually make a connection to your server.

This should setup the following directories(if you’ve specified to install your app in /app_name/):





The “releases” directory is where copies of all your actual code are stored.

“shared” is a place where you can put common, shared files like logs, static assets, and maybe config files like database.yml.

“current” is simply a symbolic link that points to the current release inside the “releases” directory (capistrano updates this on each deploy)



cap deploy:check

Check your local environment and your server and try to locate any possible issues.



cap deploy:update_code

Update application code




Set up Zend Framework Skeleton Application on Mac Step by Step

1.Install Composer

Follow Composer official guide and install it locally:

Run the following commands:

$ php -r “copy(‘;, ‘composer-setup.php’);”

$ php -r “if (hash_file(‘SHA384’, ‘composer-setup.php’) === ‘070854512ef404f16bac87071a6db9fd9721da1684cd4589b1196c3faf71b9a2682e2311b36a5079825e155ac7ce150d’) { echo ‘Installer verified’; } else { echo ‘Installer corrupt’; unlink(‘composer-setup.php’); } echo PHP_EOL;”

$ php composer-setup.php

$ php -r “unlink(‘composer-setup.php’);”

Here is the screenshot of what I did:


I got a warning about OpenSSL version, we can just ignore it for now.

Now I can see that composer.phar has been created in my file system:




2.Verify that Composer is installed successfully:

Since I installed Composer locally, so when I need to run Composer, I need to cd into the installation path, and run

$ php composer.phar



If you want to install Composer globally, you can follow this guide:



3. Install ZendSkeletonApplication

Follow the instruction on the Github repository:

Here I want to clone the application in “zend” folder, so I run this command:

$ php composer.phar create-project -n -sdev zendframework/skeleton-application zend

The screenshot:





4. Verify that ZendSkeletonApplication has been installed successfully

We can see that a “zend” folder is created and the application files are placed inside:


Start apache server(my apache root directory is parent folder of “mysites”), go to http://localhost/mysites/zend/public/

We can see the skeleton application running:





Host a website on Amazon AWS step by step guide

1.Sign in to the console using AWS credentials.

You will see a page listing all AWS web services:




2.Click on “EC2” to go to EC2 dashboard:QQ20160515-1



3.Click on “Launch Instance” to start the process of  creating a new EC2 instance:


Notice: there should be a step to create security key pairs for later access to running EC2 instance. But since I’ve already created one in my account, the system skips the step. Just remember to save the private key to a safe place, we will need it to SSH into the instance later.


3.1Choose AMI

Here I will just choose the forth one: Ubuntu Server.



3.2Choose Instance Type

Here just choose the default one:



3.3Configure Instance Details

Here we can keep everything default, no need to change any settings:




3.4Add Storage

Here we can keep everything default, no need to change any settings:




3.5Tag Instance

Here the Value we put will be the name of this instance, and it will be displayed on our AWS EC2 dashboard:




3.6Configure Security Group

This step is very important!

Firstly, I want to talk about my failure experience:

According to my experience, normally if I start something new, default is most likely to be a good choice.

So for “Assign a security group”, I selected the “Select an existing security group” radio button, and chose the first security group listed, whose name is “default”. The following is the screenshot:


And I just clicked the “Review and Launch” button.

Then I waited for a few minutes for the instance to be running. Sadly, because of the default group setting, I couldn’t SSH into the instance, which means I could do nothing about the instance. The instance was dead for me!


After checking some tutorials, I got to know that I have to create a new security group and add custom access rules. Here is the screenshot of the rules we need:


4.Then we can see the running instance in EC2 dashboard




5.SSH into the instance

In command shell, cd to the directory that the private key is located. Then run this command:

$ ssh -i [private_key_file_name] ubuntu@[public_DNS_address]

For example:



6.Run as root and access website folder

$ sudo su

$ cd /var/www/html/











Consul & Registrator & Rails Service all together in separate Docker container

1.Start a container running Consul:

Here we use this image:

$ docker run -it -h node -p 8500:8500 -p 53:53/udp progrium/consul -server -bootstrap -advertise  $DOCKER_IP


How to check that Consul is running:


In Docker-Machine, run

$ docker ps

Will see running Consul container:


We can also check the logs:

$ docker logs $CONTAINER_ID



Browse to $DOCKER_IP:8500 there is a dashboard to see the services that are registered in Consul:




2.Start a container running Registrator

Here we use this image:

Registrator automatically registers and deregisters services for any Docker container by inspecting containers as they come online.

Registrator watches for new Docker containers and inspects them to determine what services they provide. For our purposes, a service is anything listening on a port. Any services Registrator finds on a container, they will be added to a service registry, such as Consul or etcd.

$ docker run -it -v /var/run/docker.sock:/tmp/docker.sock -h $DOCKER_IP gliderlabs/registrator consul://$DOCKER_IP:8500


How to check that Registrator is running?

In Docker-Machine, run

$ docker ps

Will see running Registrator container:


We can also check the logs:

$ docker logs $CONTAINER_ID




3. Start a web service and let Registrator automatically register it on Consul

In command line, into a directory that has proper rails app, Dockerfile and docker-compose.yml file in place.

$ docker-compose build
$ docker-compose run web bundle exec rake sample_data:populate
$ docker-compose run -p 3000 -e ‘SERVICE_NAME=romui-web’ web rails s

How to check that this service is registered?


Use Consul API:

$ curl $DOCKER_IP/v1/catalog/services


We can check the service details using Consul HTTP API:

$ curl http://localhost:8500/v1/catalog/service/$SERVICE_NAME



Check in Consul UI:



4. Find the IP address and port of the service from Consul UI, and start using the service.


Now we are ready to access the service from browser: