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

[David Levine]

Andy wrote:

> The previous version of nmh on my system didn't bother with
> conversion (at least not to my knowledge) and would naively just
> output the text/plain portion and let my terminal figure out what to
> do with it (which occasionally left some characters unreadable, but
> it was sufficient for my needs).

That's the key:  that was bad behavior.  Some characters could
have had worse effects than just being unreadable.

> It seems now that nmh is trying to be more selective about what it
> shows rather than just being liberal about it and showing me what is
> in the file.

As noted above, that's a good thing.  If you want to see the raw
file, there are ways to do that without using mhshow, of course.

> Either that or previously nmh was more successful at converting the
> UTF-8 to US-ASCII.

No, it didn't convert before, by default.

> Maybe it has something to do with this particular section I found in
> the man page:
>
>        If mhshow was built with iconv(3), then all text/plain parts
>        of the message(s) will be displayed using the character set
>        of the current locale.
>
> I do indeed apparently have iconv support built-in
>
> which is perhaps why it ignored the mhshow-charset-utf-8 directive?

Yes.

> Shouldn't mhshow honor that setting regardless of whether or not
> iconv is supported?

The goal was to remove the need for that kind of manual configuration.
If someone wants to show using a character set that doesn't match
their locale, that's beyond what nmh supports.

> It does look like my only option is to set the locale for mhshow, or
> to remove iconv support.

Or just set LC_ALL (or LANG) in your environment?  LANG gets set
for me by default on Linux, so I don't have to manually.  In fact,
I had to remove an old LC_CTYPE setting from my .bashrc.

And, it gets set to a UTF-8 locale for me.  The only time I see
conversion failures is with broken emails that have an incorrect
charset label.  Though I don't receive much (any?) text that can't
be converted to UTF-8.

David