Re: Best way to depend on the user's charset

[Bruno Haible]

Hi Akim,

> I would like to use "•" in Bison's output messages when this character is 
> available.  For a similar purpose, we already have this:
> 
> #. TRANSLATORS: This is used when a line is too long, and is
> #. displayed truncated.  Use an ellipsis appropriate for your
> #. language, remembering that "…" (U+2026 HORIZONTAL ELLIPSIS)
> #. sometimes misdisplays and that "..." (three ASCII periods) is a
> #. safer choice in some locales.
> #: src/location.c:252
> msgid "..."
> msgstr "…"
> 
> So I plan to map _(".") to "•".
> 
> But I see that currently English does not even show on the translation 
> project for Bison (likewise for gnulib).  Besides, I don't know what 
> English-speaking people usually use for their locale.  I guess C.UTF-8 is not 
> unfrequent.  But it's a pity not to use "•" when it's available and to fall 
> back uselessly to ".".
> 
> So my question is: what is the recommended way to address such a case?

Gnulib has a module 'unicodeio' for this purpose. You would use it like this:

  print_unicode_char (stdout, 0x2022, 0);

This module is tested "in the field" by GNU coreutils.

Bruno