Re: tsearch filenames unlikes special symbols and numbers

2007/9/4, Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>:
> "Ben Tilly" <btilly(at)gmail(dot)com> writes:
> > On 9/3/07, Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> wrote:
> >> There seems fairly clear use-case for allowing A-Z a-z 0-9 and
> >> underscore (while CVS head rejects 0-9 and underscore).
>
> > The problem with allowing uppercase letters is that on some
> > filesystems foo and Foo are the same file, and on others they are not.
> > This may lead to obscure portability problems where code worked fine
> > on Unix and then fails when the database is running on Windows.
>
> Yeah, good point. So far it seems that a-z 0-9 and underscore cover the
> real use-cases, so what say we just allow those for now? It's a lot
> easier to loosen up later than tighten up ...
>
> regards, tom lane
>

It's system specific. I prefere a-z and A-Z. Clasic name for
dictionaries combine lower and upper characters .. for czech
cs_CZ_UTF8 etc.

dictfile = cs_CZ_UTF8 ... automatic convert to cs_cz_utf8.dict
dictfile = 'cs_CZ_UTF8' .. check and use cs_CZ_UTF8

Regards
Pavel Stehule

p.s. it's important on UNIX platforms and without any efect on windows.