Bug#432563: locales: norwegian locale has started treating aa as Е in regex

[Oleg Verych <olecom@flower.upol.cz>]

* Haavard Moen (Tue, 10 Jul 2007 17:22:44 +0200)

> $echo "petrus.haavard.name" | sed 's/[^-0-9a-zA-Z\/\.!*@_~:;< ]/_/g'
> petrus.h_vard.name
> $export LC_ALL=C
> $echo "petrus.haavard.name" | sed 's/[^-0-9a-zA-Z\/\.!*@_~:;< ]/_/g'
> petrus.haavard.name

IMHO this is subject of the `sed` package and its maintainer(s), because
of experience, documentation they may have know better.

> Aa should be treated as Е when sorting, but this behavior seems wrong.
> See also bug 430391.

As Marc noted there:

| This looks like a sed or libc issue for me:

Thus, this bug.

But lets look closer to `update-exim4.conf'. That *ASCII* checking routine
doesn't set LANG=C and LC_ALL=C before its processing, and this is a bug
instead.
____