Re: [xml] Charset trouble

["Abylai Ospan" <aospan netup ru>]

Hello!

I have problem with charsets again ;-(
I'm add in c++
sub_subtree = xmlNewChild(subtree, NULL,(xmlChar*)
"param_value",(xmlChar*)mstring.c_str());

but following error ocure when mstring in KOI8-R charset:

output conversion failed due to conv error
Bytes: 0xC0 0xE1 0xFB 0xEB
xmlOutputBufferWrite: encoder error
output conversion failed due to conv error
Bytes: 0xE1 0xFB 0xEB 0xE0
xmlOutputBufferWrite: encoder error
output conversion failed due to conv error
Bytes: 0xFB 0xEB 0xE0 0xE9
xmlOutputBufferWrite: encoder error
output conversion failed due to conv error
Bytes: 0xEB 0xE0 0xE9 0x20
xmlOutputBufferWrite: encoder error
output conversion failed due to conv error
Bytes: 0xE0 0xE9 0x20 0xCE

How define source charset (like <?xml version="1.0" encoding="koi8-r"?> in
xml file) on xmlNewChild ?
Thanks!

wbr, Abylai

----- Original Message ----- 
From: "Peter Jacobi" <pj walter-graphtek com>
To: "Abylai Ospan" <aospan netup ru>
Sent: Tuesday, October 07, 2003 11:57 PM
Subject: Re: [xml] Charset trouble


> Hi Abylai,
>
> I fear you are somewhat on the wrong list for your questions?
>
> But I hope some of your fellow Russians on the list help you nevertheless.
>
> Until their responses arrive:
>
> 1. Have looked at http://koi8.pp.ru/framed-koi8.html and generally
> speaking googled around to find your answers?
>
> > I have some trouble with html-forms contains russian letters. For
example, I
> > see (using tcpdump) that all available browsers (Opera 5x, Mozilla, IE
6x)
> > send %D1%85 for russian letter "x" and for all russian letter they send
hex
> > digits. How to convert this to native KOI8-R Charset ?
>
> 2. "%D1%85" is HTML-escaped for "\xD1\x85" which
> is UTF-8 encoded U+0445 which is CYRILLIC SMALL LETTER HA,
> which seems to be OK.
> So you must HTML-decode and then transcode into your preferred
> charset, using iconv on Linux or MultiByteToWideChar+
> WideCharToMultiByte on Win32.
>
> > Or how to tell
> > browser send form data in KOI8-R ?
>
> 3. This seems to be widely broken, according to the
> browser comparison table at
> http://koi8.pp.ru/framed-koi8.html
>
> > Or how to generate html-page using
> > libxml/libxslt in KOI8-R charset (I've read that only utf-8/16 output
> > available) ?
>
> 4. If compiling with iconv, you can specify every charset you think
> of as an output character set.
>
> Regards,
> Peter Jacobi
> Hamburg, Germany