Re: [Nmh-workers] General question - unsupported charset conversion

[Oliver Kiddle]

On 28 Feb, Ken Hornstein wrote:
> FWIW, I wanted to see what other MUAs do, so I decided to look at mutt;
> if iconv() fails and the target character set is UTF-8, it substitutes
> U+FFFD, otherwise it substitutes '?'.

For what it's worth, that seems entirely sensible to me. It is also the
behaviour we already have when presenting header fields such as Subject
and From. (possibly without the U+FFFD addition)
As a general rule, I'd prefer an approach that makes it easy for me to
read an e-mail even where there are some errors. A few ?s where there
would be accented characters is still very readable.

In my existing script, I use the suffix //TRANSLIT on the target
codeset. With GNU iconv, this will do some things like convert € to EUR.
The trouble is it isn't portable. There's also limitations to the
approach as for a German text it would make sense to convert ä to ae but
for other languages, this makes no sense at all.

For HTML e-mails I rely on the -I and -O options to w3m. I think this is
better than having nmh do the conversion. Especially if the HTML defines
a charset in meta tags.

Oliver