Re: Impossible dependency: slang1-dev vs. slang1-utf8-dev
>Hi, Colin Watson wrote:
>
>OK, then I guess I'll have to hurry up and write a first version of that
>program I'm planning. Of course, I can't run it, or test it, or even
>submit it to Debian: it'd not be buildable. Since Debian doesn't accept
>source-only submissions, this neatly sidesteps any possible RC-ness of
>the problem. :-/
>
>So. Next step: Create newt-8bit{,-dev}? (Quick and somewhat dirty)
>
>Or (cleaner, but much more work) create -utf8 versions of everything that,
>directly or indirectly, build-depends on slang1-dev?
>
>To do nothing and release Sarge with the conflict is, IMNSHO, not
>an option here, RC-bug-when-literally-interpreting-policy or not.
I am testing a solution that adds ELF dependencies to the
slang libraries, following Steve Langaseks paper at
debconf4:
http://www.debconf.org/debconf4/talks/dependency-hell/index.html
basically, version the symbols differently, with
SLANG_1.4.9_UTF8 in the UTF-8 version,
SLANG_1.4.9 in the other;
then a recompile of the packages (especially the libraries)
that depend on slang:
apt-cache rdepends slang1 | grep lib gives:
libterm-slang-perl
libdv-bin
libavifile-0.7c102
aalib1
aalib-bin
and it should be fine.
However, I believe packages should just be rebuilt against
the slang1a-utf8 and slang1-utf8-dev and slang1-dev dropped:
the only reason slang1 should remain is for support for
non-Debian binaries. I was intending to propose this when
sarge ships, but now appears to be a good time.
Regards
Alastair
>--
>Matthias Urlichs
>
Reply to: