problem with plain edit-field and UTF-8 characters

Hello,

we have a MODx installation, and we use no WYSIWYG-editor, just a plain text-field for text-entry. When i enter words with UTF8-karakters, like f.e.:

één

the word seems to be encoded in some kind of way before it is entered into the mySQL database. The result is that the output is not correct. When i paste the text directly into the database-content using PHPmyAdmin everything works correctly.

Our config:

mySQL5 with latin1-swedish-ci collation on all modx-tables
mySQL default character set is : UTF-8 Unicode (utf8)

MODx character encoding is set default to UTF8

any ideas what’s happening?

Olaf

Do you have an example URL?

Make sure there’s an explicit header in the PHP like:
header("Content-type: text/html; charset=utf-8");

because the default charset in PHP is still ISO-8859-1, iirc.

Quote from: ganeshXL at May 01, 2007, 06:17 AM

Do you have an example URL?

Make sure there’s an explicit header in the PHP like:
header("Content-type: text/html; charset=utf-8");

because the default charset in PHP is still ISO-8859-1, iirc.

we’re using AMFPHP for this site, so there is no HTML, we’re using the modx db-API to get the content directly from the database. For the AMFPHP gateway we’re using $gateway->setCharsetHandler( "utf8_decode", "ISO-8859-1", "ISO-8859-1" );

Olaf

we’re using AMFPHP for this site, so there is no HTML

Well, HTML or not: doesn’t matter. Encoding is encoding, whether you output plain text or HTML.

Did you see this tutorial? http://www.sephiroth.it/tutorials/flashPHP/amfphp_iconv/
Might help you out with this...

will take a look into it, tnx!

Quote from: olafmol at May 01, 2007, 05:13 AM

mySQL5 with latin1-swedish-ci collation on all modx-tables
mySQL default character set is : UTF-8 Unicode (utf8)

MODx character encoding is set default to UTF8

The problem is, if your database content is using latin1_swedish_ci, you have to use latin1 encoding to connect and see the proper character set. Also, check the database charset and collation vs. the table charset and collation to make sure they match. Even if the database charset is UTF8, if you have tables with a collation corresponding to a different charset, you need to set the MODx variable to match the charset for the value in your tables.

If you are using a release candidate of MODx 0.9.6, change the $modx_connection_charset variable in your manager/includes/config.inc.php to equal ’latin1’. Starting in 0.9.6, MODx now executes a SET CHARACTER SET immediately after connecting to the database; this was to correct issues with certain PHP/MySQL configurations on Windows where PHP was not respecting the character set of the database.

tnx...the info in the link Ganeshxl provided solved our problem. I installed IconV.dll in PHP, and changed the AMFPHP gatewaysettings to iconv and UTF8 for reading and writing....works like a charm now!

Olaf