common accounts for el_GR

[Reinke Bonte]

It's a bit off topic, but wouldn't it be a good idea to store the data
at least in the correct encoding in the XML file. It should be UTF-8 or
an encoding which is named in the head. At the moment all non Latin-1
data is forced to Latin-1 and then translated into HTML-entities. This
ensures that the file can only be opened under the same encoding,
because then this strange procedure is reversed. 

I had tried to fix this, but my C knowledge is very limited. Somehow
libxml1 does not behave as I expect it to and I couldn't get gnucash
work with libxml2, which fixes the internationalization problems of
libxml1. Once gnucash uses libxml2, it is a very easy fix, one just need
to convert all char variables to UTF-8 before casting them XMLchar.

Repairing this would be great help for everyone who works in more than
one encoding. I have to filter the XML file through a recode shell
script which imitates the GNUcash conversion procedure each time I want
to work in another encoding.

Reinke




On Thu, 2003-05-22 at 12:44, Linas Vepstas wrote:
> On Wed, May 21, 2003 at 10:52:44PM -0400, Derek Atkins was heard to remark:
> > Well, it DOES live under el_GR/...  so it's quite clear (once it's
> > in the hierarchy) that it's greek.
> 
> yes, I know.  But since I was recently burned by the fact that
> some vital info about a chunk of data failed to travel with the data
> itself, it got me to thinking.   Its now like a design pattern
> in my head, or more like a conspiracy theory: everywehere I turn,
> I see this little problem...
> 
> --linas
> 
> p.s. yes, I have a list of about 10 items I need to do to modernize the
> gnucash website. I'm getting to it.  I just spent the last two evenings
> fighting off gremlins again, this time on a different server.