Re: [XaraXtreme-dev] XML notes

[Phil Martin <phil@xxxxxxxx>]

Alex Bligh wrote:

> --On 10 May 2006 15:22 +0100 Phil Martin <phil@xxxxxxxx> wrote:
>
> > * Remember that libxml2 uses UTF-8 encoding in memory no matter what the
> > program's native encoding or the encoding of the original doc. So you
> > have to be careful to do correct conversion to and from UTF-8 when
> > passing strring parameters to it. I used wxString for this because it 
> > has
> > useful encoding conversion functions.
>
>
> So I am guessing it is impossible to reliably encode non-UTF-8 in XML -
> using this library anyway? Can you confirm that's the case? If so,
> things like saving a text story as XML would be inherently lossy. Fine
> if so, but we should know.
>
> Alex

No, it's only the in-memory storage that uses UTF-8 and remember that 
UTF-8 can encode all unicode characters.

llibxml2 can load and save to any encoding you like so long as you have 
the appropriate encoding handlers installed and the default handlers 
include UTF-16 (big and little endian), ISO-Latin-1, ASCII, HTML and 
more depeing on platform and support libraries.

So there's no reason to lose data if you load one encoding and save to 
another.

BTW: libxml2 records the encoding that a document had when it was loaded 
and will save back in the same encoding unless you specify otherwise.

Phil