Re: Aging Dependencies?

>As far as I can see, there wasn't any non-ascii characters in the
>email, only some ? where they were supposed to be. I didn't analyze it
>character by character, though...

Well, no, of course you didn't ... why would you? :-)

Not trying to pick on Andreas ... I've just become more aware of these
things since I've been trying to improve how nmh handles MIME replies.
Handling all of the corner cases ends up being difficult, and when I run
into them I want to try to understand exactly what is going wrong.

Specifically, Andreas's signature contained:

Dortmund, HRB 14952 - Geschäftsführer: Detlef Eppig - Vorsitzender des

At least that's what it was supposed to say (I've converted that into UTF-8).
But because it was marked as us-ascii nothing handled it quite right.

In my particular case, my locale is set to UTF-8.  If you look at
it with "show" or "mhshow", what you see for the word in question is:

	Gesch<E4>ftsf<FC>hrer

What happened here is that since it was marked as us-ascii nmh just sent the
message to "less", who realized that 0xE4 and 0xFC are not valid UTF-8, so
it ended up hilighting the invalid bytes in question.

Exmh did something even weirder.  What I see there is:

	Geschäftsführer

What I guess happened THERE is something assumed that the characters
in question were ISO-8859 and converted them to UTF-8.  But since
exmh thought that the text was in us-ascii, it ended up simply
displaying the bytes using a "us-ascii" font and that had the
practical effect of showing the UTF-8 bytes as ISO-8859.  (The UTF-8
encoding of U+00E4, “ä”, is 0xc3 0xa4, which ends up being displayed
as U+00C3, “Ô, and U+00A4, “¤”).  In theory those probably should
have been displayed as U+FFFD, “�”, but I'm not sure who exactly
should take care of that.

--Ken

_______________________________________________
Exmh-users mailing list
[email protected]
https://www.redhat.com/mailman/listinfo/exmh-users