First we install a database. I decided to use MariaDB, but MySQL works just as fine. So open a Terminal and enter

brew install mariadb

Once it is installed, follow the instructions on screen to finalize the installation:

- Setting up the database:

unset TMPDIR
mysql_install_db

Creating the startup item:

cp /usr/local/Cellar/mariadb/5.3.5/homebrew.mxcl.mariadb.plist ~/Library/LaunchAgents/
launchctl load -w ~/Library/LaunchAgents/homebrew.mxcl.mariadb.plist

And finally starting the server

mysql.server start

Next we install PHP with fcgi support. There is a great PHP repository for Homebrew, so we will use it.

Tap into the repository with the following command:

brew tap josegonzalez/homebrew-php

Then install PHP with the following command:

brew install -dv php --enable-phar --with-mariadb --with-cgi --enable-fcgi --with-suhosin

or alternatively for MySQL:

brew install -dv php --enable-phar --with-mysql --with-cgi --enable-fcgi --with-suhosin

Now comes the tricky part – Cherokee itself. With your favourite text editor, edit the cherokee formula located at

/usr/local/Library/Formula/cherokee.rb

and modify the install function so that it looks like this (the hilighted lines are the added lines)

def install
system "./configure", "--disable-dependency-tracking",
                      "--prefix=#{prefix}",
                      "--sysconfdir=#{etc}",
                      "--localstatedir=#{var}/cherokee",
                      "--with-wwwuser=#{ENV['USER']}",
                      "--with-wwwgroup=www",
                      "--with-wwwroot=#{var}/cherokee/wwwroot",
                      "--without-mysql",
                      "--enable-internal-pcre"
system "make install"

prefix.install "org.cherokee.webserver.plist"
(prefix+'org.cherokee.webserver.plist').chmod 0644
(share+'cherokee/admin/server.py').chmod 0755
end

save your modifications and then install cherokee:

brew install cherokee

After a while all should be ready to run, now it’s time to configure PHP in Cherokee. Cherokee comes with a nice wrapper function to start the admin interface. in your terminal window enter

cherokee-admin-launcher

and press Enter. A dialog pops up asking for your user password – enter it and press Enter. A moment later your default web browser opens and displays the admin interface. Initially I had Safari as my default browser, but that was a little pain in the behind because Safari thought the adin interface is a phishing site. After I made Firefox the default browser all was fine again.
On the main screen you see an overview of your Cherokee server. From here you can also stat and stop the server.

Click the VServers icon.
The page that opens displays the settings for the default server:

Click on “Rule Management” in the “Behavior” Tab

Click the + symbol to add a new behavior. A popup window opens with all available options:

Click on “Languages” then on “PHP” and finally on “Add” to add PHP:

Click “Create” to setup PHP:

Click “Save” to save the settings. If Cherokee is already running, you will see a dialog asking you to restart the server, Chose “Graceful Restart” in that case.

That’s all, you now have a web development environment with Cherokee, PHP 5.3 and MariaDB.
To start Cherokee, either use the start/stop button on the “Home” screen of the admin interface, or enter “sudo cherokee” in a terminal window.

On their website they even have a little web interface to build your own custom style, so I thought that is just perfect for a lazy dude like myself.But I was wrong. It does allow a bit of customization and works all in all ok, but the one thing that mattered to me, the navigation bar, just would not want to play along nicely. I changed the size to 60 pixel height, generated my stylesheet, attached it to the HTML document I prepared and….TADAAAAA – It looked horrible!

I wasted a bit of time trying to adjust the CSS with no results, so I finally decided to involve one of my friends, G. Oogle. I kept asking him questions about the navbar and he gave me plenty of food for my brain, but not quite what i was looking for. It was frustrating and I even considered using a different template, but somehow I really wanted the Twitter thingy to work. So I asked Dr G. some more questions and finally I got what I was looking for. Unfortunately it involved getting into LESS, but turned out that it was not such a big deal after all.

Preparations

The steps here are for Mac OS, but my friend G. Oogle has a great homepage that you can use to get in touch with him. Other options include:

• You can bing for results on Google
• You can google for results on Bing

All that matter is that you get the Yahoo effect in the end. Having just started to get back into Ruby on Rails, I quickly entered “gem install less” to ghet the less processor installed – turned out to be a stupid idea.

To get started, the instructions on George Ornbos blog were very helpful which are (assuming you have homebrew installed) as simple as this:

• Install node.js
brew install node
• Install  NPM
curl http://npmjs.org/install.sh | sh
• open your .bash_profile
  pico ~/.bash_profile (could also be called .bashrc)
• Declare the npm home variable
export NODE_PATH="/usr/local/lib/node"
• modify your PATH
/usr/local/share/npm/bin:$PATH (but make sure you keep exsisting modifications)
• reload your envidonment
  source ~/.bash_profile
• Install uglify
  npm install uglify-js -g

Modifying Twitter Bootstrap

If you don’t already have a terminal window open, open one and create a folder for your project and check out the git repository for Twitter Bootstrap:

git clone git://github.com/twitter/bootstrap.git

This will download the current version of the Twitter Bootstrap project to your disk. My search on Google brought me to a discussion on github, and this post pretty much sums up the changes we have to make:

The files we need to change are in the ‘less’ folder which is located in the ‘bootstrap’ folder of the project we cloned from github. First open the file variables.less. Around line 140 you will see the section that configures the navbar variables:

 // Navbar
// -------------------------
@navbarHeight:                    60px;
@navbarBackground:                @black;
@navbarBackgroundHighlight:       @grayDark;

@navbarText:                      @grayLight;
@navbarLinkColor:                 @grayLight;
@navbarLinkColorHover:            @white;
@navbarLinkColorActive:           @navbarLinkColorHover;
@navbarLinkBackgroundHover:       @blue;
@navbarLinkBackgroundActive:      @blue;

@navbarSearchBackground:          lighten(@navbarBackground, 25%);
@navbarSearchBackgroundFocus:     @white;
@navbarSearchBorder:              darken(@navbarSearchBackground, 30%);
@navbarSearchPlaceholderColor:    #ccc;

Customize it to your liking. That was the easy part. But the next part is easy, too. The definitions for the navbar style are in the file navbar.less (Who would have guessed ) and the definitions that fixes the navbar are:

.navbar .nav
{
    height: @navbarHeight;
}

.navbar .nav > li > a
{
    padding: (@navbarHeight-20)/2 10px ((@navbarHeight - 20)/2 + 1);
}

.navbar .brand
{
    padding: (((@navbarHeight-20)/2) * 0.8) 20px (((@navbarHeight - 20) / 2) * 1.2);
}

So, we find the definition for .navbar .nav (line 203) and add the height to it, then the definition for .navbar .nav li a and replace thepadding rule with the formular (@navbarHeight – 20)/2 10px ((@navbarHeight – 20)/2 + 1) and finally the definition for .navbar .brand (line 69) and replace the definition for padding with padding: (((@navbarHeight – 20) / 2) * 0.8) 20px (((@navbarHeight – 20) / 2) * 1.2);

Do not forget to save the changes, then go back to the terminal window. Now it’s time to generate the new stylesheets, When you cloned the git repositry, the folder bootstrap was greated to contain all the files for the project. change into that folder and you will see that it contains a Makefile. That file takes care of everything. Simply enter make bootstrap, press Enter and a few moments later you will find a bootstrap folder with the new CSS, the Javascripts and images. Those are the customized files you can you in your own Project.

I am pretty confident that in one of the upcoming updates to Twitter Bootstrap this issue will be fixed. At the time of writing, the version for Twitter Bootstrap is 2.0.2

It took me a moment to understand the limitations of the free version. At first I thought I can only have 2 hosts enabled at a time, but far off.
Horst allows you to create sets of hosts. For example, you can setup a set called “Work” that contains all the hostnames you work on at the office, and maybe another set “Home” that contains all the hostnames for your private projects, you can even create more than 2 sets, however in the free version only 2 sets can be active at a time. The paid version is only €5 and although I don’t really need it, I am seriously considering to buy it just to support the developer, as this is really so convenient and the first free/affordable solution that actually works for me out of the box. (Apart from a “dumb user” mistake, but I guess I can only blame myself for that.)

To install Horst, simply download it from the website, unpack the archive and drag the Horst application into your application folder.
Start it, by double clicking on the application icon and enter your admin password if you get asked for it. Horst modifies the /etc/host file (it does not delete anything you have already in there, just adds/modifies its own entries) so it requires administrator permissions.

When Horst opens for the first time, it looks pretty empty:

Click on the [+] in the lower left corner to create your first set of hosts:

You can double click on the text “New Set” to rename it, then start entering your hosts in the right part of the window in the format IP HOSTNAME

Once you’re happy with your setup, press [Command]+S to save, then drag the little slider next to your set name from “OFF” to “ON”

That’s it! Your hosts are now active:

I admit, the screenshot of the terminal is not necessary, but i was desperately looking for an opportunity to show off the cute header of my terminal window.

Of course you can simply open a terminal window and enter sudo mate /etc/hosts (if you have Textmate) or sudo pico /etc/hosts and enter those entries manually, but for lazy people like me, Horst just adds a lot of convenience. Plus – it makes it a lot easier to manage the entries as you can enable and disable sets of hosts simply by switching the set to ON or OFF.

If you enjoy the app as much as I do, then please support the developer and buy the full version. I just found €5 that I was still carrying in my wallet from my last trip to Germany (and I kept it there for a coffee next time I go back) and I think I will use them to buy Horst.

Here are the steps I did:

- Open the System Preferences:

- Select Network from the options:

- Select your connection and then click on “Advanced:

- Select the “Proxy” tab:

make sure “Http Proxy” is checked, then enter the proxy details.

The proxy server is proxy.trueinternet.co.th and the port is 8080

Click “OK” when you are done:

Click “Apply” to save the settings:

TADAAAAA!!!!! – AppStore is working:

"restart")
$0 restart-zdd
$0 restart-monitor-node
$0 restart-apache
$0 restart-lighttpd
$0 restart-jb
$0 restart-jqd
;;
I looked at it without my glasses on, I put my glasses on and had another look, but somehow I do not see mysql mentioned there anywhere. So how does the script know that it is supposed to start MySQL?
I modify the start and restart section as follows:

"start")
$ECHO_CMD "Starting $PRODUCT_NAME $PRODUCT_VERSION ..n"
$0 start-zdd %
$0 start-monitor-node
$MYSQL_EN && $0 start-mysql %
$0 start-apache %
$0 start-lighttpd %
$JB_EN && $0 start-jb %
$0 start-jqd %
$ECHO_CMD "n$PRODUCT_NAME started..."
;;

"restart")
$0 restart-zdd
$0 restart-monitor-node
$0 restart-apache
$0 restart-mysql
$0 restart-lighttpd
$0 restart-jb
$0 restart-jqd
;;
Modified the “stop” section like so:
"stop")
$ECHO_CMD "Stopping $PRODUCT_NAME $PRODUCT_VERSION ..n"

$0 stop-apache %
sleep 2
$ZCE_PREFIX/bin/clean_semaphores.sh
$0 stop-lighttpd %
$JB_EN && $0 stop-jb %
$MYSQL_EN && $0 stop-mysql
$0 stop-jqd %
$0 stop-zdd %
$0 stop-monitor-node %
$ECHO_CMD "n$PRODUCT_NAME stopped."
;;
and for extra fun, added the re-start option:

"stop-mysql")
if $MYSQL_EN; then
$ZCE_PREFIX/mysql/bin/mysql.server stop
fi
;;


For extra credits you can add the status option, too, but I can’t be bothered.

Today I realized that I have a few posts which promise some sort of “to be continued…” but never do. Luckily, nobody actually comes here, so there’s virtually no harm done.

This is where the original post ended. I have to admit, that I wrote the Post in the same spirit as Zend seems to release the CE edition (i.e. In theory those changes should do the trick, so why bother testing?!) and of course that backfired. The above modifications to the script actually can start, stop , restart MySQL server when it is executed. Tiziano however found out that MySQL is still not starting during the boot process and I was able to happily confirm that the next morning when I re-booted the Mac in my office. He also found the solution to this problem:

I found the solution! You have to replicate changes in this file:

/Library/StartupItems/ZendServer_init/zendctl.sh

it works!

Thanks, Tiziano, for finding this, sharing it with us and completing this post.

Love
Michael

What are the options to fix it?

One option is to re-install from sources using the latest sourcecode. Personally I don’t like that because I just prefer to stick to the package system of the distribution.

What else can we do?

The admin interface for Cherokee is actually written in Python. So here is what I did to get my admin interface up and running:

1) download the latest sources from http://www.cherokee-project.com/downloads.html and unpack the tarball
2) open the file “configured.py” and modify the top entries like so:

PREFIX = "/usr"
LIBDIR = "/usr/lib"
DATADIR = "/usr/share"
DOCDIR = "/usr/share/doc/cherokee"
LOCALEDIR = "/usr/share/locale"
WWWROOT = "/var/www"
SYSCONFDIR = "/etc"
LOCALSTATE = "/var"

3) Open a terminal window and become root (sudo -i)
4) change to /usr/share/cherokee and delete the admin folder.
5) copy the admin folder from the source code (with the modified configured.py file) into /usr/share/cherokee
6) start cherokee-admin and enjoy.

I can’t really take credit for this, as I found the general idea of hiding the partition on www.radiotope.com. Reading the article and its comments, I cam to this (maybe) streamlined version.

First we need the UUID of the drive we want to hide and the easiest method seems to be to use the Disk Utility.

After starting the Disk Utility, click on the hardisk/partition you want to hide and click the info button on the toolbar.

The information you need is the “Universal Unique Identifier” – select it with your mouse and copy it.

Then open a terminal window (iTerm 2 is a brilliant alternative to the native terminal client on OS X). Now if you look at the /etc folder, you most likely won’t find the fstab file, ony fstab.hd. It confused me a little, so I did more research and found out that it does not matter. We just create the fstab file.

sudo pico /etc/fstab

Once the editor window opens, enter

UUID=[Your Universal Unique Identifier] none hfs rw,noauto 0 0


Reboot your Mac and the partition is hidden. Wohoo…
Now watch me having fun on the computer:

You can now use your web server of choice, maybe the free version of MAMP or the version that eosgarden provides.

Next time we will setup a virtual machine with Ubuntu and a few goodies to get the development server running. This is not a necessary part, but it gives you the feeling of the real deal somehow.

This post was written here: