php.activerecord

nevermind, figured it out myself.

What I was looking for was validates_confirmation_of, which didn't exist in phpar. I wrote it, you can get my fork with the changes from http://github.com/sdellysse/php-activerecord. I submitted a pull request also.

examples for those having similar issues:
class User extends ActiveRecord\Model {
static $before_validation = array(
'encrypt_password_if_changed',
);
static $validates_confirmation_of = array(
array('password'),
);
static $validates_presence_of = array(
array('ecrypted_password'),
);
public function encrypt_password_if_changed() {
if(isset($this->password)) {
$this->encrypted_password = User::pw_encryption_scheme($this->password);
}
}

public function set_password($p) {
    $this->password = $p;
  }

public function set_password_confirmation($p) {
    $this->password_confirmation = $p;
  }

Nice, but I prefer custom validation to do this "matches" rule
http://www.phpactiverecord.org/projects/main/wiki/Validations#validate

take a look at this commit
https://github.com/devi/kohana-activerecord/commit/53388c9662e1cd3a4bd1c39a4fb8a9cc0dcdfd33

While using validate() works, why would you use that instead of the patch I submitted? Rails's ActiveRecord implementation has this feature available, and since this is a port of that to PHP, why not make this functionality available out of the box? Also, I can tell that it was meant to be there originally but got put on the backburner as the error message was already in place before I started working on it.

why would you use that instead of the patch I submitted?

Because my goal is different with yours.
I'm focusing on "wrapper" for Kohana Framework instead patching php-activerecord.

That's the beauty of open source right ?

Devi Mandiri wrote:

[...]

Because my goal is different with yours. I'm focusing on "wrapper" for Kohana Framework instead patching php-activerecord.

That's the beauty of open source right ?

True, that it is. I now understand; here's to hoping my pull request gets accepted, as I'm not a fan of patching a library either... I'm using that code for a project, and I'd like to keep library modifications to minimum for the next person to work on what I've built.

Shawn Dellysse wrote:

here's to hoping my pull request gets accepted

Don't worry, if I'd be kla, I would accept it immediately (it behaves as in Rails, it may be useful for the general audience, it's documented, and heck, it works!). The fact that he didn't respond only means he's busy with other stuff right now.

@Devi: can you tell me more why attribute_is_dirty() doesn't work for you? It was committed not long ago, if I remember well I even wrote some unit tests for it.

Still checking my box, attribute_is_dirty() always return false, don't know why.

Shawn,

This looks pretty cool! One question though. It looks like you would have to set a "password_confirmation" property on your model for this to work. If that is the case, how do you keep password_confirmation from being stored in your database?

Thanks,
Taylor

Taylor,

What I do is in the actual database, I have a property called `encrypted_password`. In the PHP Model, I place the code I wrote above. Notice the functions prefixed with `set_`. These create attributes in the Model, but not in the database. When the `before_validation` callback is ran, it checks to see if the virtual attribute `password` is set, and if so, make the encrypted password from that.

Did I explain this well enough for you? My writing skills are not the greatest.