<<< Date Index >>>     <<< Thread Index >>>

Re: national chars - libiconv not used?



Alain Bench wrote:
 On Monday, June 12, 2006 at 13:50:24 +0200, Michal Hajek wrote:
Alain Bench wrote:
You would see an uppercase accented U here: "é" (really an e acute).
I do not see it correctly
    What do you see?
You would see an uppercase accented U here: "~B" (really an e acute)

--with-libiconv-prefix=/usr/local/lib
    If you omit this, libiconv is still found? Is it linked?

I do not use this option now (sorry, I forgot to say it, it was one of many
config attempts ...)
Added it again & recompiled: no visible change.

    Could you please pick 2 or 3 standard locales available in the
"locale -a" list, say cs_CZ, en_US, and en_US.UTF-8, export LC_ALL=each
one in turn, and report: "locale charmap" at shell, and

hpuh:/home/hajek>export LC_ALL=cs_CZ.iso88592
hpuh:/home/hajek>locale charmap
"iso88592.cm"
:set &charset ?charset: charset="roman8"

hpuh:/home/hajek>export LC_ALL=en_US.iso88591
hpuh:/home/hajek>locale charmap
"iso88591.cm"
:set &charset ?charset: charset="roman8"

hpuh:/home/hajek>export LC_ALL=en_US.utf8
hpuh:/home/hajek>locale charmap
"utf8.cm"
:set &charset ?charset: charset="roman8"

":set &charset ?charset" in Mutt? After the check, unset LC_ALL, and do
not set it again in any startup scripts (LANG suffices).

| for loc in cs_CZ en_US en_US.UTF-8
| do
|   echo $loc
|   export LC_ALL=$loc
|   locale charmap
|   mutt -nF /dev/null -D | grep ^charset
| done
| unset LC_ALL

aaargh I'd better read whole message before answering ... OK:
(for loc in cs_CZ.iso88592 en_US.iso88591 en_US.utf8)

cs_CZ.iso88592
"iso88592.cm"
charset="roman8"
en_US.iso88591
"iso88591.cm"
charset="roman8"
en_US.utf8
"utf8.cm"
charset="roman8"



| set charset="IBM852"


Pick your system iconv command (not /usr/local/bin/iconv), and try:

| $ printf "Ren\xC3\xA9\n" | /usr/bin/iconv -f utf-8 -t ibm852
| René
| $ printf "Ren\xC3\xA9\n" | /usr/bin/iconv -f utf-8
| René
| $ printf "Ren\xE9\n" | /usr/bin/iconv -f iso-8859-1 -t ibm852
| René
| $ printf "Ren\xE9\n" | /usr/bin/iconv -f iso-8859-2 -t ibm852
| René

It is hard to try, hp-ux printf seems not to support \x things :-(
I will try it somehow later.

Thanks,
             Michal