Re: [Gnumed-devel] Country zones and i18n

[Karsten Hilbert]

On Thu, Nov 17, 2011 at 11:17:37PM +0000, Jim Busser wrote:

> >> Wondering how "generic" would work. Say a user was configured as one of
> >> 
> >>    de_DE
> >> 
> >> … would their phrasewheels be populated by matches among
> >> 
> >>    base values
> >> +  de_DE
> >> +  generic (even without configuration) ?
> > 
> > No. It would be populated by
> > 
> > dem.* + de_DE
> >                     or de
> >                             or geneirc [sic]
> 
> So, in the case of the portuguese city and street names, shall I insert the 
> accented values into
> 
>       dem*
> 
> and insert the unaccented ones into
> 
>       i18n.translation.trans
> 
> with lang = 'pt'

No, because that would make unaccented strings to be *shown*
regardless of how they are *matched and selected*. We don't
want that. Which is why I ...

> ... indicated a few mails ago was *not* what you were suggesting?

As I said, it's a crutch at best and the solution is to
provide unaccent() and identify the queries that need it.

Karsten
-- 
GPG key ID E4071346 @ gpg-keyserver.de
E167 67FD A291 2BEA 73BD  4537 78B9 A9F9 E407 1346