I kept trying to force the request with the following failing code:

#don't use this, it's bad
# I want to reroute /reviews to /admin
get "/admin"

That was clearly wrong, but nothing immediately jumped out at me as to how to fix it. After much googling around, I finally found the quick and easy solution:
ruby code on github here

describe "/admin" do
    before(:each){
      get :index
    }
 
    it { should route(:get, "/admin").to(:controller => :reviews, :action => :index) }
    it { should respond_with(:success) }
end

Hope this saves someone else some time.

Well, this month I’m fortunate enough to have a small cameo in “The Goods” section on page 54 (see the crm review online). In it, I talk about the merits of Fat Free CRM over a few other customer relationship managers tools I’ve used in the past.

Thanks to Inc for the interview. Given this exposure, I’ve expedited a plan that was in the works to go live in the 1st quarter of 2010. BME now provides hosting services for Fat Free CRM. To learn more, see the hosting fat free crm website or contact me for more information.

I really hope this adds a ton of value to new or established businesses looking for more efficiencies in their web-contact-to-deal-close processes.

The plugin installs just like any other Wordpress plugin. It should be pretty intuitive to anyone that has even limited experience administrating a Wordpress site. The plugin can be downloaded from Wordpress or at my repo on github.

Fat Free CRM is a great customer relationship manager (CRM) I’ve been using for a little while now instead of Highrise. I used Highrise for a little while before I grew frustrated by its hosting costs and lack of features. The guys at 37signals don’t seem to update it very often either.

For the rails project, I forked the repository and made a few minor changes like adding access via an API key. I also added a few usability options in the view codes to make integration with the plugin seamless.

For new businesses that are just implementing a CRM solution, I’d be happy to host the CRM application on my servers for a nominal cost. Feel free to contact me for more details. (hint, the contact form is an example of the plugin in action )

To use the plugin and CRM together, just follow these simple instructions:

1. Download or clone the CRM from my github repo. Install on my servers (contact me for pricing) or on your own
2. Download and install the fat-free-crm-lead plugin either from Wordpress or my github repo
3. Configure the plugin with the URL of your CRM and an API token (details are provided in the admin interface of the plugin)
4. Watch the leads come rolling in and enjoy not having to copy and paste from emails any longer

Hopefully, this is really easy. If you run into any problems or have feature requests, please feel free to contact me.

====UPDATE====

The plugin has been accepted into the WP plugin directory. Take a look at the fat-free-crm-lead-form at Wordpress.

Given the budget of next to nothing and the relatively small size of the office, I decided it was best to proceed with a Linux based Samba server. Although their office is 100% Microsoft (a mix of XP and Vista machines), Linux would be a great light weight solution. The added benefit was we re-purposed an ancient workstation (circa 1996) to run the Samba server so no new hardware was purchased and we prevented another machine from polluting a land fill.

We created three shared drives for the business.

1. A share everyone in the office could access
2. A share only the executive could access
3. A share for the exclusive purpose of backing up everyone’s workstation

I simply installed Centos 5.2 distribution on the old workstation, created three distinct user accounts and group permissions, enabled the samba service, and locked down access through iptables firewall.

Creating a secure, readable, smb.conf file was really important to me. Not only for readability but also because I won’t be supporting this server after installation. Any other administrator should be able to walk into the office and quickly & clearly see how to make edits for upgrades or changes.

The smb.conf file:

#smb.conf
 
[global]
	#set up what the end user would see on their desktop
	workgroup = H4LSharedFiles
	server string = common files for H4L
 
	#general user perms
	security = user
	passdb backend = tdbsam
 
	#printing support
	load printers = no
        printing = bsd
        printcap name = /dev/null
        show add printer wizard = no
        disable spools = yes
 
[AllSharedFiles]
	comment = Files for everyone to share
	path = /sambashares/allAccess
	writeable = yes
	browseable = yes
	guest ok = yes
	write list = +allemployees
 
[Backup]
	comment = Files to back up everyone's computer
	path = /sambashares/backups
	writeable = yes
	browseable = yes
	guest ok = yes
	write list = +allemployees
 
[AdminOnly]
	comment = Access for sensitive files
	path = /sambashares/admins
	writeable = yes
	browseable = yes
	guest ok = no
	write list = +administrators

After hours of desperate searching, I finally found the solution to appropriate exception handling with php on IIS on the iis blog. The solution was more of a footnote about halfway down the page in the php.ini section. The trick is to set fastcgi logging to false. Not the most ideal solution, but I was finally able to handle exceptions the way php code should.

From iis blog: “FastCGI logging should be disabled on IIS. If it is left enabled, then any messages of any class are treated by FastCGI as error conditions which will cause IIS to generate an HTTP 500 exception.”

php.ini

fastcgi.logging = 0 ;

So, here it is, a fairly beta version of a horizontal content slider written with the Mootools 1.2 framework. Simply download (or copy and paste) the ContentSlider class into your page and then instantiate with the required arguments.

Usage Example:

the ContentSlider class

/**
 * Content Slider - A Useable Horizontal Scroller
 *
 *  Requires a parent element with hidden overflow and child elements floated inside the parent.
 *  Take any two elements as previous and next buttons and scolls the child elements through the parent element
 *
 * Dependencies: MooTools 1.2 (and probably 1.1.x but not tested)
 *
 * adapted from Antonio Lupetti's original blog post at
 *    http://woork.blogspot.com/2009/01/elegant-animated-weekly-timeline-for.html
 *
 * @version		0.9
 *
 * @license		MIT-style license
 * @author		Bryce McDonnell bryce {at} bridgetownint {dot} com   www.brycemcdonnell.com
 * @copyright	Bryce McDonnell
 */
var ContentSlider = new Class({
    config: {
		totIncrement: 0,
		increment: 0,
		maxRightIncrement: 0,
		container: $empty,
		slidingEl: "",
		slideAtATime: 1,
		previous_btn: $empty,
		next_btn: $empty,
		duration: 1000,
		transition: Fx.Transitions.Back.easeInOut
	},
 
	initialize: function(config) {
 
		this.config = $merge(this.config, config);
 
		this.config.maxRightIncrement = this.config.increment * -( ( $$(this.config.slidingEl).length / this.config.slideAtATime) - 1 );
		if (this.config.previous_btn && this.config.next_btn) {
			this.activateMoveBtns();
		};
 
		return this.config
	},
 
	move: function() {
		this.config.fx = new Fx.Tween(this.config.container, {
		property: 'margin-left',
		duration: this.config.duration,
		transition: this.config.transition,
 
		onStart: function() {
			this.config.previous_btn.removeEvents('click');
			this.config.next_btn.removeEvents('click');
		}.bind(this),
 
		onComplete: function() {
			this.activateMoveBtns();
		}.bind(this)
 
		});
	},
 
	activateMoveBtns: function() {
		//Next Button
		this.config.next_btn.addEvents({
			'click' : function(event){
				this.config.totIncrement = (this.config.totIncrement>this.config.maxRightIncrement) ? this.config.totIncrement - this.config.increment : 0;
				event.stop();
				this.move();
				this.config.fx.start(this.config.totIncrement);
	 		}.bind(this)
 
		 });
 
		//Previous Button
		this.config.previous_btn.addEvents({
			'click' : function(event){
				this.move();
				event.stop();
				this.config.totIncrement = (this.config.totIncrement < 0) ? this.config.totIncrement + this.config.increment : this.config.maxRightIncrement;
				this.config.fx.start(this.config.totIncrement);
			}.bind(this)
 
		});
	}
});
ContentSlider.extend(new Fx.Tween);

html markup

<script src="/lib/js/homeslider.js" type="text/javascript"><!--mce:0--></script>
<div id="worksamples" class="slide-box">
	<a id="slide-left" class="slider-arrows"> previous </a>
	<a id="slide-right" class="slider-arrows">next</a>
<div id="workimgs">
<ul id="slider-list">
	<li><img class="aligncenter size-full wp-image-168" title="BME logo small" src="http://blog.thedevranch.net/wp-content/uploads/2009/05/BME-logo-small.jpg" alt="BME logo small" width="300" height="405" /></li>
	<li><img src="/lib/img/slide/img.jpg" alt="meth" width="292" height="184" /></li>
	<li><img src="/lib/img/slide/img.jpg" alt="nuskin" width="292" height="184" /></li>
	<li><img src="/lib/img/slide/img.jpg" alt="choice" width="292" height="184" /></li>
	<li><img src="/lib/img/slide/img.jpg" alt="meth" width="292" height="184" /></li>
	<li><img src="/lib/img/slide/img.jpg" alt="nuskin" width="292" height="184" /></li>
	<li><img src="/lib/img/slide/img.jpg" alt="choice" width="292" height="184" /></li>
</ul>
</div>
</div>

the javascript

new ContentSlider({
		increment: 904,
		container: $("slider-list"),
		slidingEl: "#slider-list li",
		slideAtATime: 3,
		next_btn: $("slide-right"),
		previous_btn: $("slide-left")
	});

the css markup:

  /***  SLIDER   ***/
#worksamples {
	width:960px;
	height:180px;
}
#workimgs{
	width:900px;
	overflow:auto;
	overflow:hidden;
	height:200px;
	margin:0 auto;
}
#slider-list{
	width:20000px;
	margin:0;
	padding:0;
	left:400px;
}
#slider-list li{
	list-style:none;
	margin:0;
	margin-right:5px;
	padding:2px;
	float:left;
	width:292px;
	height:200px;
}
.slider-arrows {
	display:block;
	width:18px;
	height:42px;
	text-indent:-9999px;
	overflow:hidden;
	margin-top:74px;
	cursor:pointer;
}
#slide-left {
	float:left;
	background:url(../img/left-arrow.png) no-repeat !important;
	background:url(../img/left-arrow.gif) no-repeat;
}
#slide-right {
	float:right;
	background:url(../img/right-arrow.png) no-repeat !important;
	background:url(../img/right-arrow.gif) no-repeat;
 
}

Explanation:

This serves as a pretty rudimentary example. The markup is simply a list with containing images that need scrolling. For the next and previous buttons, I got creative and got rid of the text and set an image for the background. That provides a good looking UI to the user.

Usage requires two basic elements: a container and child elements to scroll through. The parent should hide all overflow and the child elements should float so as to be one long horizontal array of scrolling elements. Simply include the class on any page and then instantiate the class passing in the required arguments.

Arguments:

increment : the number of pixels the container should scoll. You could call $(”inner-el”).width here if you want to just scroll one element. I’m working on eliminating this requirement and instead just calculating it using the other arguments passed in.

container: A reference to the parent element with hidden overflow

slidingEl: collection of elements to scroll horizontally. Can be a css selector object

slideAtATime: an integer passed in to indicate how many child elements to scroll at when either the previous or next buttons are pushed. defaults to 1

next_btn: a reference to the element to scroll to the right. A click event is added to this element when the class is instantiated

previous_btn: a reference to the element to scroll to the right. A click event is added to this element when the class is instantiated

Closing:

Happy MooTool’ing ! Be sure to post a comment if you see something wrong or would like to add to this.

This is an old version of an improved Mootools Content Slider

I recently implemented a javascript based slider using MooTools. I’m increasingly becoming a huge fan of Mootools. Their documentation is great and the code I write just seems to be much cleaner. I don’t feel like it’s as fragile as js I’ve written in prototype/scriptaculous.

The content slider, which I pretty much implemented verbatim using Antonio Lupetti’s great tutorial at http://woork.blogspot.com/2009/01/elegant-animated-weekly-timeline-for.html.

Antonio’s tutorial is really great. It is missing two elements which I think improve the code base:

1) you have to set the number of elements to scroll explicitly

2) it does not handle some crazy user clicking like mad on either the left or right buttons

To solve both those pains, I’ve updated the code a bit to work with MooTools 1.2 and made a couple of enhancements seen below.

var totIncrement = 0;
var increment = 387;
var totalIncrements = $$("#sliderList li").length - 1
var maxRightIncrement = increment*(- totalIncrements);
 
var fx = new Fx.Tween('sliderList', {
   property: 'margin-left',
   duration: 1000,
   transition: Fx.Transitions.Back.easeInOut,
   onStart: function() {
      $("slide-left").removeEvents('click');
      $("slide-right").removeEvents('click');
   },
   onComplete: function() {
      appEvent();
   }
});
 
function appEvent() {
   // Previous Button
   $('slide-left').addEvents({
      'click' : function(event){
         if(totIncrement<0){
            event.stop();
            totIncrement = totIncrement + increment;
            fx.start(totIncrement);
        }
      }
   });
 
   // Next Button
   $('slide-right').addEvents({
      'click' : function(event){
         if(totIncrement>maxRightIncrement){
            event.stop();
            totIncrement = totIncrement - increment;
            fx.start(totIncrement);
        }
      }
   });
}
appEvent();

By encapsulating the right and left click events into its own function, I can easily turn on and off the events to prevent the end user from invoking the slide event in the middle of an event that may already be going on. A shorter way to say that is users can’t double click.

Secondly, creating the totalIncrements variable allows someone (like me) to add another list element into the markup at a later date without having to remember to update the javascript code.

These are just a couple of tweaks to an otherwise great tutorial. Thanks a lot Antonio and keep up the good work.

The goal is to dump the database to the filesystem, add it to the tar file, encrypt it and then delete the clear text version of the file.

Security considerations abound. In a nutshell however, it is important not to publish any unencrypted database dump anywhere. Dump files can be searched on easily by Google and if you house even a morsel of private user data, you’ll be in trouble very quickly if it’s not encrypted.

Getting started, the MySQL command to dump a database is:

 mysqldump -u user -psupersecret --all-databases > /some/dir/backupsql.sql

-p will prompt for a password

-all-databases will dump every database you have running on the mysql instance. If you want a specific database, replace –all-databases with your_db_name

This command pushes the output of the mysqldump command to a file on the filesystem.

Since I don’t want my cron job to be prompted for a user’s password when dumping the db, I have created a special backup user with only enough permissions to read the tables, and write their contents to a file. You can create such a user by running the following command inside a mysql prompt:

GRANT SELECT, LOCK TABLES ON *.* TO 'backup'@'localhost'
IDENTIFIED BY 'supersecret'

Make sure to flush privileges after creating users. There’s not much more frusterating than trying to figure out why that script didn’t work.

One security consideration is that if you were to run the ps command while the dump script is running, it will expose the username and password of the backup user potentially allowing other users of the system to make select statements to databases they otherwise shouldn’t be on.

By limiting the backup user access only from the localhost, this eliminates the possibility of someone else dumping your database from some other host; the user must be logged into the box the database resides on.

At the conclusion of this script, we will have all the database data and structure backed up to a file on the filesystem. In the next post, I’ll explain how to add this data to a tar archive and then encrypt the whole mess.

However, after just a morning of futzing with a few key linux commands and better utilizing a service I already back up my source code to (www.github.com) I have a found a robust and secure way to handle automated, off-site, redundant backups in a way that will let me compete with some bigger shops. I’ve posted the code below so I hope you will find it useful. Over the next few posts, I’ll unpack what I’ve written and the philosophy behind it.

A few things bothered me in the way I was doing traditional backups:

1. I knew I had to get them off-site, but actually finding time to get off-site (to a secure location) wasn’t happening.
2. The backup had to be absolutely secure. My clients’ source code is too precious and leakage too damaging to make even one mistake with security breach
3. Had to be simple and automated. I usually have 10 other things I need to do at the same time. I didn’t want backups to be number 11.
4. Small file size. Again, being a small dev shop, I didn’t want to put a lot of cost into storage of incremental backups
5. Incremental backups were key since I don’t want to go to all this trouble only to restore a copy of the bad data I was trying to replace. If I a problem isn’t made known until after the next set of backups are made, I’d be overwriting bad data with bad data; better to restore to the point before the problem happened.

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35

 
 
#!/usr/bin/sh
 
#backupdb
echo "======================================"
echo "backing up database"
mysqldump -u backup --all-databases > /var/www/html/alldatabases.sql
 
#tar and compress the directories really hacky and should either make a file for exclusions or just get rid of the crap
tar -czvvf /var/www/bkp/websqlbkp.tar.gz /var/www/html/alldatabases.sql /var/www/html/*
 
#encrypt that mug and then remove the decrypted file
echo "encrypting backup"
openssl des3 -salt -k supersecret \
        -in /var/www/bkp/websqlbkp.tar.gz \
        -out /var/www/bkp/websqlbkpencrypted.tar.gz
 
#cleanup files I don't want people to see
echo "cleaning up files"
rm -rf /var/www/bkp/websqlbkp.tar.gz
rm -rf /var/www/html/alldatabasesl.sql
 
#update the git repo
echo "committing to git `date` "
cd /var/www/bkp/
 
echo "adding to git"
/usr/local/bin/git add .
 
echo "commiting git"
/usr/local/bin/git commit -a -m "commiting backup on `date`"
 
echo "pushing to github"
/usr/local/bin/git push origin production