I've discovered my issue is related to the /u switch. Interestingly enough,
the moment I add /u it no longer matches anything, and returns an empty
string. I've even added A-z, a-zA-Z, and [:alpha:] to the pattern at
different times. If I remove the 'u' (unicode switch) from the pattern, I
get the ascii chars back. But if I put it in, it strips out everything:
Returning: 'Bertr'; from: 'Bertrán '; pattern; /[^\p{L}]/
Returning: '''; from: 'Bertrán ''; pattern; /[^\p{L}]/u
Maybe this will narrow down more exactly what's going on. It's really hard
Googleiing this issue.
Philip
On Wed, Mar 26, 2008 at 1:44 AM, Karl Katzke <[EMAIL PROTECTED]> wrote:
> I'll run some tests on it tomorrow, but I think that Zend_Filter_Input
> pulls from the post variables in the request object, which are processed
> before the $_POST variable that you set above would be set.
>
> Unicode issues are *tricky* to get pinned down. You also need to have the
> right charset being passed in the headers from the server, the right
> override in your browser, etc. so on so forth. We've had problems with a
> bunch of our servers when they were going through a misconfigured squid
> cache that stripped the UTF-8 headers off ... but only when you hit it with
> a request for the site in the Turkish charset. All kinds of things could
> cause the behaviour you're seeing. And knowing how much translation and
> internationalization that is happening, I think we'd have seen this fairly
> simple case show up before now.
>
> -K
>
> On Tue, Mar 25, 2008 at 11:53 PM, Philip G <[EMAIL PROTECTED]> wrote:
>
> >
> > Well, the script I wrote was a hack job to show what I was doing.
> > Technically, the data is being based through _POST. Are you saying the only
> > way to get this to work is to Unicode the scripts on the file system? I
> > never herd of that requirement.
> >
> > I'm trying to make a form submit function more unicode friendly. I have
> > the required UTF-8 accet charset in place. I have verified that 'Bertrán'
> > is being passed across properly. I even verified that PHP doesn't mess with
> > it (by printing out _POST). However, once it goes through the filter
> > process, it gets stripped.
> >
> > Philip
> >
> >
> > On Tue, Mar 25, 2008 at 7:20 PM, Guillaume Rossolini <
> > [EMAIL PROTECTED]> wrote:
> >
> > > Hi,
> > >
> > > I suppose the script is encoded as Unicode in your filesystem? Your
> > > sample code fails if encoded as ANSI, but succeeds if encoded as UTF-8.
> > >
> > > <?php
> > > require_once 'Zend/Filter/Input.php';
> > >
> > > $_POST['first_nm'] = "Bertrán";
> > >
> > > $filters = array ( 'first_nm' => 'Alpha' );
> > > $validators = array( 'first_nm' => array ('Alpha', 'presence' =>
> > > 'required') );
> > > $input = new Zend_Filter_Input($filters, $validators, $_POST);
> > >
> > > echo $input->isValid() ? 'ok' : 'dead';
> > >
> > >
> > > Regards,
> > >
> > >
> > >
> > >
> > > On 3/25/08, Philip G <[EMAIL PROTECTED]> wrote:
> > > >
> > > >
> > > > I'm trying to get a form to validate with unicode characters.
> > > >
> > > > For a very simple example, I have:
> > > >
> > > > $_POST['first_nm'] = "Bertrán";
> > > >
> > > > $filters = array ( 'first_nm' => 'Alpha' );
> > > > $validators = array( 'first_nm' => array ('Alpha', 'presence' =>
> > > > 'required) );
> > > > $input = Zend_Filter_Input($filter, $validators, $_POST);
> > > >
> > > >
> > > > Now, a simple $input->isValid() call and it fails. Returns:
> > > > [first_nm] => Array ( [stringEmpty] => ' ' is an empty string )
> > > >
> > > > Is there something special I need to do in order to get
> > > > Zend_Filter_Input to understand UTF-8? I've checked the data being
> > > > passed to
> > > > PHP; it is valid. I checked the code and it seems to do some utf8
> > > > check, but
> > > > apparently it's not working for some reason.
> > > >
> > > > Thanks
> > > >
> > > > --
> > > > Philip
> > > > [EMAIL PROTECTED]
> > > > http://www.gpcentre.net/
> > >
> > >
> > >
> > >
> > > --
> > > Guillaume Rossolini
> >
> >
> >
> >
> > --
> > Philip
> > [EMAIL PROTECTED]
> > http://www.gpcentre.net/
> >
>
>
--
Philip
[EMAIL PROTECTED]
http://www.gpcentre.net/
