Re: [BRLTTY] Braille coding in Gnopernicus

[remus draica <rd baum ro>]

On Thu, 2005-11-24 at 19:57, Bill Haneman wrote:

Hi,


> Hi Jan:
> 
> I am taking the liberty of CC-ing the GAP list again.
> 
> I have an idea what the problem is here, and a possible solution (below).
> 
> Jan Buchal wrote, on the BrlTTY list
> 
> >Hello,
> >
> >I'm not sure if my email is not of topic in this list but I asked in
> >GAP list already and I am not sure if the problem is solved. 
> >
> >So, I would like use Gnopernicus with brltty's api with ISO-Latin2 or
> >better UTF-8 characters. One man from gap said that gnopernicus works in
> >utf-8 but actually the braille display shows ISO1 only. Where is
> >problem, how can I do?
> >  
> >
> I think there are two problems.  One is that although gnopernicus uses 
> UTF-8 almost everywhere, it does its own braille dot conversions.  So it 
> can't pass UTF-8 on to BrlAPI at the moment, with its current architecture.

> The other problem is that currently, unless I am mistaken, gnopernicus' 
> braille tables are 8 bit tables, and gnopernicus always assumes Latin-1.
> 
That's right.

> Latin-2 is also 8-bit, so in theory the current gnopernicus architecture 
> could handle this, but it would need to know when to convert from UTF-8 
> to Latin-2 instead of UTF-8 to Latin-1.  I am not even totally sure that 
> gnopernicus is converting from UTF-8 to Latin-8 now, before it does a 
> braille table lookup, but it probably should be.

Gnpernicus calls g_utf8_get_char () for every char to display. If the
result is bigger than a value (256), the undefined code is dispalyed,
otherwise the entry in a table.

> 
> If gnopernicus had a Latin-2 table that worked for your locale, then I 
> think we could solve the problem by including the "character set" in the 
> braille table, at the top, and making sure gnopernicus did the correct 
> UTF-8 to 8-bit conversion for the given table before doing a character 
> lookup.  From there on, the correct dots should be sent to BrlAPI.  
> Remus, please correct me if I have this wrong.
> 
If the new "latin-2" table replaces the current table, it will be used
by gnopernicus. Ther is no way to use both latin-1 and latin2 tables at
the same time. 

> A better solution long-term would not be limited to an 8-bit braille 
> table lookup, but this is the sort of thing 'gnome-braille' attempts to 
> handle.
> 


Regards,
Remus