> On Mon, 21 Jun 1999, Darren O. Benham wrote:
>
> > On Mon, Jun 21, 1999 at 05:35:08PM +0200, Ove Kaaven wrote:
> > > (And by the way, if I make it use the Norwegian ispell dictionary, why
> > > aren't the Norwegian 8-bit (iso-8859-1) characters handled right? Well,
> > > *I* can live without, since my own brain's spellchecker happens to be
> > > even better than what a machine could provide, but it's a minus in
> > > abiword's show-off value when I want to show my colleagues what you can
> > > do with linux...)
> >
> > What do you mean by "not handled right"? What's happening? The ispell
> > spelling package is a fickle beast but I don't see why it can't handle 8
> > bit characters properly.
>
> Well, just that if I write a normal word with such a character in it, like
> "år", "går", etc, then it's marked as a spelling mistake.
I looked at this with gdb, and two problems were apparent when using the
inorwegian package from debian potato:
- the upperconv[] didn't convert the 8-bit characters to uppercase
- hash.c in abiword hashed "GÅ" as 0x4700c4 % 0x14eb8 = 0x65f4, while I
found "GÅ" hashed in norsk.hash as 0x6633, which suggests that it
was compiled using 0x47 0x103 as input data?
Is this a compatibility problem (abiword implementation missing some new
conversion features), since the ispell program itself seems to work fine?