Charset decoding problem

Hi,

I've got a very strange problem with UTF-8 encoded data outside ASCII range.

While on localhost all went smoothly, the same pages on the server show � (Latin-1 chars (ä, ö, ü, ß, ...)) and ? (above Latin-1 range (typographics)). Even the support does not really have a clue (that could help me).

reference: http://test.kulturbeutel-leipzig.net/main.php?f=presse

Javascript on – all works fine (data are fetched directly from a MySQL DB via AJAX)
Javascript off – (a bit more complicated) data are fetched from DB (stored there as WDDX serialized data) and deserialized into an object, which in turn is responsible for output.

maybe there's some problem with the deserialization.....

Does anyone have an idea, how I can find out the source of the problem?

thanks

PS: the DB should contain the same data, because I used a SQL dump of one to build the other.

PPS: if you need class definitions, just ask (it would be too much to list all incorporated classes at once)

local system: Darwin Melchior 9.6.0 Darwin Kernel Version 9.6.0: Mon Nov 24 17:37:00 PST 2008; root:xnu-1228.9.59~1/RELEASE_I386 i386 / PHP 5.2.8.
(= Mac OS 10.5)

public system: Linux Custom Build 64 Bit prohost.de XEON SMP x86_64 (Red Hat Enterprise Linux) / PHP 5.2.6.

re: Charset decoding problem

Hi.

I don't really know much about WDDX, but as I understand it, it is basically XML?
I had similar problems when passing XML files around a while ago, where the server was sending stuff as Unicode, the browser was rendering using Unicode, but the output was all mangled.

Turned out all I had to do to fix this was add:
And everybody suddenly started understanding each other.

My mistake was to assume that the XML file would adopt the charset passed with a Content-Type header like HTML pages do.

Perhaps you left this out as well?

re: Charset decoding problem

yepp, WDDX is XML (useful if you have your configuration stored as XML)

but the XML header was there from the start.... and obviously Javascript has no problems at all with it.

sample WDDX:
note * = soft hyphen (escaped by bytes' editor)

re: Charset decoding problem

there seems to be something wrong with the deserializer, after some testing I can say the problems occur right after deserialization.

does anyone know, how I can determine the encoding/charset of a variable content? (that would be interesting to know)

thanks

re: Charset decoding problem

PHP strings (until version 6) don't have any native support for Unicode, or any other charset for that matter.
A string character is essentially the same as a byte.

Try running the variable content through utf8_encode. See if that helps any.

re: Charset decoding problem

Though it converts the Latin-1 characters, it's no help with the characters initially showing up as '?' („ “ – ’ … and the like)

re: Charset decoding problem

finally got the problem somehow sorted by converting all non-ascii characters using unicode entities and this little function: http://de2.php.net/manual/de/functio...code.php#75941

re: Charset decoding problem

Very interesting indeed. Interesting enough that I scoured the net and I believe this resource: http://www.mozilla.org/projects/intl...Detection.html to be the most detailed and closest any person/group has yet come to solving this extremely challenging problem.