[BUG] speechd-up fails to detect one-character messages

[Alexander E. Patrakov]

2008/1/11, Hynek Hanke <hanke at brailcom.org>:

> I tried to cleanup the code in CVS (http://www.freebsoft.org/speechd-up)
> and make the single character handling a bit more general, so now this
> should work and not interfere with controll commands. Could you
> please test if it is fixed now?

This particular bug no longer exists with the English language.
However, the use of isprint() is questionable for other languages:
speechd-up never calls setlocale(), and this means that all non-ASCII
(i.e., 0x80..0xff) characters will be omitted from being counted
(because the default "C" locale is used).

Either the setlocale() call has to be added and the "correct locale
for speechd-up bootscript" requirement has to be documented (note:
most distros run initscripts in the "C" locale), or the check whether
to count a character has to be rewritten.

-- 
Alexander E. Patrakov