Re: Screwball characters
On Thursday, October 28, 2004 at 5:15:58 PM -0500, Kyle Wheeler wrote:
> set charset=`printenv | grep '^LC_' | grep -i utf-8 2>&1 1>/dev/null \
> && echo utf-8 || echo //TRANSLIT`
Hum... "egrep '^(LANG|LC_)'" maybe catch better. Or more specific
"echo ${LC_ALL-${LC_CTYPE-${LANG-C}}}". Or more adapted but less
portable "locale charmap". Or simply...
> if utf-8 support is available, use it, otherwise use //TRANSLIT
I'm not sure it's always true with every iconv library out there,
but AFAICS it doesn't hurt to have always static $charset=//TRANSLIT,
even in an UTF-8 locale (in your +HAVE_LANGINFO_CODESET case).
The problem with permanent transliteration is more that you never
can't be sure if you see (and quote) a real dash or a faked em-dash. I'd
wish Mutt hints us, say with a different background color for faked (and
masked and \octalised) chars. Why not reusing "color markers" object.
In the meantime I mostly don't use //TRANSLIT by default, but
activate it on-demand™ with a macro. Unfortunately runtime $charset
changes need a <change-folder> reread to apply completely.
[1252 labelled L1]
> the bad MUA this time was Horde's IMP webmail
Thanks! Granted, there is no good webmail. But IMP seemed one of the
rare not so bad. :-(
Bye! Alain.
--
[There is no good webmail. A good webmail is a dead webmail.]