Please block sed 4.2-1

Adeodato Simó dato at net.com.org.es
Sat May 9 10:07:19 UTC 2009


+ Andreas Metzler (Sat, 09 May 2009 09:05:02 +0200):

> sed's behavior has changed in respect to using non-ASCII chars as a
> pattern delimiter. Previously sed -e 's§text§replacement§' worked
> no matter which locale was used. (The "§" being represented in
> ISO-8859-1 encoding, not as a two-byte UTF-8 character.)

> Now with version 4.2 sed's unicode support seems to grown, it now sees
> the "§", realizes it is neither ASCII nor a valid UTF-8 character and
> throws an error message:
> sed: -e expression #1, char 2: delimiter character is not a single-byte character

> This breaks exim4 for anybody running an UTF-8 locale who has
> fine-tuned the locale configuration by setting not just LANG but
> LC_CTYPE (or LC_ALL). See #527445.

> Please stop sed from migration to testing until a to-be-uploaded
> version of exim4 (i.e. >= 4.69-10.1) that has a workaround for the bug
> has reached lenny.

Could sed 4.2 gain an appropriate "Breaks: exim4-whatever (<< foo)"?
But, since britney does not grok Breaks yet, I've added a block, okay.

And, in any case, can exim4 stop using non-ASCII as a delimiter? Or is
there really *no* ASCII delimiter left that would be used for the
expression at and?

Thanks,

-- 
- Are you sure we're good?
- Always.
        -- Rory and Lorelai




More information about the Pkg-exim4-maintainers mailing list