Problem with XSP and ISO charset

[Andrzej Talarczyk]

First of all, hi everybody here,

I have successfully used AxKit for several months on my Debian Linux box. Recently, I have decided to migrate one of the sites to a FreeBSD box and I have stumped by an unexpected problem. I use AxKit to generate HTML files from XML source encoded in iso-8859-2 in the processing chains that looks like this:

file.xml (iso-8859-2) -> XSP -> XSLT -> XSLT -> HTML (iso-8859-2)

On Debian, intermediate xml structures (as seen with AxTraceIntermediate) are UTF-8-encoded and final HTML output is properly converted to iso-8859-2 by the second XSLT stylesheet.

On FreeBSD, however, the first intermediate (after XSP processing) looks as if there was an "additional" (unnecessary) conversion step from iso-8859-2 to UTF-8: e.g iso-8859-2 code is converted to UTF-8 (two codes) and then each of them is converted once again, resulting 4 codes from a single iso code. Needless to say, that the final HTML page is mangled, too.

This is strange, because the source files haven't been modified in any way when moving from the Debian box to the FreeBSD box, but the AxKit behaviour is very different.

Both boxes have AxKit 1.6, most recent versions of LibXML and LibXSLT. The Apache config files are identical on both machines regarding AxKit. Probably I'm missing something obvious but I'd be grateful for any hints how to pinpoint the problem (especially how to debug XSP processing).

Thanks,

Andrzej


---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]