Hi, On Tue, Feb 22, 2005, Nicolas Boullis wrote: > Well, I gave you my reasoning why I think circular dependencies > (including self-dependencies) are broken. Where do you think my > reasoning is wrong?
You did not prove by "a + b" that circular dependencies are wrong, you quoted the policy on the way depends are handled (and I had already read it in my previous message). The policy doesn't tell in 7.2 how circular dependencies are handled, I've grepped on "circular" and couldn't find any chapter where this is used. What followed in your previous message was your interpretation on why circular dependencies make packages uninstallable. That's plain wrong, there are a lot of packages with circular dependencies in Debian, and I pointed galeon / galeon-common as an example thereof: galeon depends on galeon-common which depends on galeon (I've picked this example because I'm the maintainer of galeon, but apt-get search -- -common for more, such as vim / vim-common). But I completely agree that self-dependencies are plain wrong, and as I said this will be solved by next gtk upload (removing the shlibs.local). You might be interested in reading this thread from january: <http://lists.debian.org/debian-devel/2005/01/msg01784.html> I presume the problem is the same as the one you mentionned. > Moreover, although I don't know much about the galeon/galeon-common > situation, I don't think foo-common depends on foo any more than a > library depends on a binary that uses it... I didn't understand that, could you please rephrase it? Regards, -- Loïc Minier <[EMAIL PROTECTED]> "Neutral President: I have no strong feelings one way or the other."