The beauty of this method is that you can still run all your validation checks and then hash the data between validation and saving. Why is that important? Consider you have a rule that says a password should be no more than 15 characters. When you hash it, it would break the rule and your record would never save.

All you have to do is hash your fields in the beforeSave function of your model like this example from a User Model:


public function beforeSave() {
App::uses('Utitlity','Security');
if(!empty($this->data['User']['password'])) {
$this->data['User']['password'] = Security::hash($this->data['User']['password']);
}
return true;
}


One import aspect of this is to always return true. If you don’t your record will NEVER save.

I’ve been building with PHP for 10 years, but over the past 2 years all of my projects have been built using CakePHP. I’ve never looked back as far as stability and speed of development is concerned.

So if you are looking for a CakePHP developer, please visit and maybe hire me.

However, when no database is present CakePHP throws a warning that it can’t connect to the database. The solution is to create a dummy database connection to convince CakePHP that all is ok. And, it’s as simple as this:

Create a new folder app/models/datasources/dbo

Create a new file at app/models/datasources/dbo/dbo_dummy_db.php and add the following:

1
2
3
4
5
6
7

<?php   
    class DboDummyDb extends DboSource {         
                 function connect()  {                
                          $this->connected = true;
                          return $this->connected;
                  }
            }

In your app/config/database.php find the default connection array – a line that starts $default = array
and replace the line for the driver from
'driver' => 'mysql',
to
'driver' => 'dummy_db',

And that’s all there is to using CakePHP without a database.

When I develop CakePHP apps, my local database configiration file is very different from the live one. Therefore, when I am using Git to push my changes to a live environment I want to be sure this file is not pushed. Fortunately Git  has the ability to let me do this with the .gitignore file. I can specify the following:

app/config/database.php

One path to file on each line will ignore those files.

However, I recently found myself in the situation where I had already added files I wanted to ignore to my repository. After reading the docs and several forums, the answer seemed to be using the following command:

git rm --cached filename

I did this, commited my changes, then pushed them live. Horror. The database.php had been removed on my local repo and this was subsequently pushed to the live site which duly collapsed. Fortunately I used Git to quickly revert my changes. I’ve no idea why the command above didn’t work for me, but I did find a failsafe way of removing a tracked file without deleting it.

1. make sure the path to file is listed in your .gitignore
2. copy the file outside of your repo.
3. do a git rm -rf path/to/file.ext
4. do a git commit to remove the file from the repo
5. copy the file back from where you placed it in step 2

You should be able to change the file now and when running git status it should not show up.

It may seem long winded and a more-than-likely unorthodox to do it this way, but it makes me feel safer.

Returns a string with all spaces converted to underscores (by default), accented characters converted to non-accented characters, and non word characters removed.

Slug is the perfect method for generating a search engine friendly URL from any string. For instance in an application I am currently developing there is a categories model with a field called title. This makes a great field to automagically convert to a url part for SEO purposes. Now instead of a URL like /category/browse/1 ; I can have something like /category/browse/1/restauraunts-cafes-and-bars.

For my taste of URL structure slug doesn’t work perfectly out of the box: it uses underscores to seperate words, rather than hypens; and it doesn’t convert all characters to lower-case. Having said that, a workable solutions is never far away in CakePHP.

The easiest solution for me is to build a reusable method in my /app/controllers/app_controller.php that I can access from any other controller or view. This way I can create a slug on the fly in my view or commit them to my database on create or update. So, let me introduce my hyper-simple private method, toSlug() :

function _toSlug($string) {
return low(Inflector::slug($string, '-''));
}

Call this method like this:

$this->_toSlug("Any String or Data You like")

Which returns any-string-or-data-you-like

Full-text indexes can be used only with MyISAM tables, and can be created only for CHAR, VARCHAR, or TEXTcolumns.

It did take a little bit of experimentation to get the syntax for the query correct in CakePHP as the query is a little different from the regular “SELECT field FROM table WHERE somevalue = somevalue“.

So, for this example I will use a MyISAM table called pages and search against the title, content and tags fields for the term” jquery web development”. You should substitute that term with your dynamically passed value in the code below.

Firstly I built the conditions for my query into a “params” array.

$params= array(‘conditions’ => array(
‘MATCH(Page.title,Page.description,Page.tags)
AGAINST(“jquery web development” )’ ));

Finally a call to CakePHP’s native find method will perform the search:

$this->Page->find(‘all’,$params);

One last note: MySQL full text search can be powerful when used correctly and a potential bottleneck when poorly implemented. I suggest reading the documentation for further information and more advanced usage.

The first is something I use quite alot when I am giving users the ability to administer lists of items in bulk. That including a checkbox that when checked sets the state of all the checkboxes beside a list of items to “checked”. Naturally unchecking it will uncheck all checkboxes. I’m sure you get the picture here’s the code based on two scenarios depending on your preference. But firstly you need to include the jQuery library in the head of your document (I grab mine straight from Google’s repository to save my bandwidth):

<script src="http://ajax.googleapis.com/ajax/libs/jquery/1.3.2/jquery.min.js" type="text/javascript"></script>

and then somewhere in your document a controller checkbox that toggles all the others (note the call to toggleChecked when the state of the checkbox changes):

<input type="checkbox" onclick="toggleChecked(this.checked)"> Select / Deselect All

Now here are  two versions of the toggleChecked function dependent on the semantics of your document. The only real difference is the jQuery selector for your list checkboxes:

1: All checkboxes have a class of  “checkbox” (<input type=”checkbox” class=”checkbox” />)
function toggleChecked(status) {
$(".checkbox").each( function() {
$(this).attr("checked",status);
})
}

2: All the checkboxes are contained within a div with an arbitary id:

<div id="checkboxes">
<input type="checkbox" />
<input type="checkbox" />
<input type="checkbox" />
</div>

In this case the function would look like this:

function toggleChecked(status) {
$("#checkboxes input").each( function() {
$(this).attr("checked",status);
})


Have fun!

The issue seems to be tied to PHP 5.1.2 and the way regular expressions are handled, however it seems the it may appear in other configurations too. Before I get to a workaround, note I haven’t levelled the “bug” word at either PHP nor CakePHP – for all I know it could be a difference of opinion on how something should be handled. What I do know is that I would rather harness CakePHP’s flexibility to create custom model rules than try upgrading my current live server PHP installation.

Here is an example of how you should be able to implement the alphaNumeric rule using validation in CakePHP based on a simpliefied model called “User”:
class User extends AppModel

{
var $name = 'User';
var $helpers = array('Form');
var $validate = array(
'username' => array(
'loginRule-1' => array(
'rule' => 'alphaNumeric',
'last' => true,
'message' => 'Only letters and numbers allowed'
));
}

The rule in bold italics above is the one that doesn’t always work. However this can be easily rectified with a custom rule built on a regular expression seeking out valid alphanumeric strings.

To achieve this the alphaNumeric rule just needs to be replaced with this:

"rule" => array('custom', '/^[a-z0-9]*$/i'),

Problem solved.

We built the site in co-operation with T-MAX media from Oxford. Our input was cutting the new designs into valid XHTML and CSS, building some custom jQuery functionality and programming new features into their content management system. The result is a high-impact site for a major corporate client – and all just as Zumo Internet turns 1.

They also specialize in commercial and domestic green roofs.  Green roofs / vegetated roofs or roofs with vegetated spaces – are a simple, attractive, ecologically beneficial and economically viable solution to many of today’s environmental issues. They have been adopted extensively in Europe (particularly in Germany) but have not really been taken up to the same extent in this country.

Zumo Internet designed and developed a content managed site for them as well as building intranet so they can manage their commercial roofing projects and keep a track on clients and warranties for roofing systems provided by them.