Re: [Nmh-workers] mhshow: unable to convert character set of...

[Ken Hornstein]

>Because everything else I use works  with my current locale, which means
>I  have to  either change  the locale  for everything  else, or  make an
>exception for nmh, which previously didn't  care what my locale was, and
>it would  just make a best  effort.

Well ... calling it a "best effort" is kind of generous.  Unless of course
by "best effort" you meant, "do nothing".  It didn't matter if it was
ISO-8859-1, KOI-8-R, or UTF-7 ... you got the raw bytes without any
conversion.

But again ... I'm missing the problem with changing your locale,
exactly?  You say "everything" works with your current locale; fair
enough.  But you don't have any other programs other than nmh that
have to do character set conversion, I would gather.  I mean, if your
terminal CAN handle UTF-8, then a locale setting of "C" is wrong,
period.  Really, in this day and age setting a locale of en_US.UTF-8
(or whatever you prefer) should simply just work.  If it causes problems 
I'd be interested in hearing about them.

>I understand that. At any rate, I'm  probably alone in this, so it's not
>worth debating at this point. I will  find a way to make it work (either
>using an environment variable for  invocations of show/next, or removing
>iconv support from nmh).

I know that you compiled nmh without iconv support, but I must caution
you ... one of my plans for future nmh releases was to simply require
the use of iconv.  At this point iconv is a POSIX standard and can
reasonably be assumed to be present on modern systems, and there is
a reasonable open-source alternative; I do not believe there is any
gain in maintaining the conditional #ifdefs require for iconv support
(also, it's a pain to test non-iconv support).  But also in the future
plans is to change the character set conversion routines to always
do substitution, so if you really (for reasons I will admit I cannot
fathom) want to use the C locale, things should work fine.

--Ken