On Thu, 02 Nov 2006 11:06:21 +0100, Goswin von Brederlow <[EMAIL PROTECTED]> said:
> "Steinar H. Gunderson" <[EMAIL PROTECTED]> writes: >> Hi, >> >> It has recently come to my attention that nfs-utils (which is >> priority standard) cannot depend on ucf, since ucf is of priority >> optional. >> >> I can only see four solutions for this: >> >> a) Ignore the problem for etch, figure out what do to afterwards. >> b) Downgrade nfs-utils' priority (but I don't think this is a very >> good >> idea). >> c) Rip out the ucf dependency (possible, but far from ideal; using >> ucf >> instead of regular conffile handling allowed me to close a few >> bugs, at least one of them RC). >> d) Upgrade ucf to priority standard. >> >> I'd personally go with d); would anybody have objections to this? >> (Cc-ing Manoj as the ucf maintainer, even though I'd believe he >> reads -devel.) >> >> /* Steinar */ > e) Test for ucf presence and emulate its behaviour if not. This is pointless duplication of functionality. Having every standard package with non-conffile configuration files re-invent the wheel endlessly seems like bad systems integration design. I posit that we do need a standard, well defined, policy driven mechanism that packages can depend upon to provide debconf based handling of non-conffile configuration files; if there are deficiencies in ucf that should be pointed out, and fixed, or a standard mechanism implemented from scratch if ucf is deemed too buggy to fix. ucf has a size of 55KB. I think we can easily find more than 55KB we can yank out of standard post-etch (I have my eye on some packages already, and I am sure d-i folks have some ideas too). manoj -- If God had intended Man to Smoke, He would have set him on Fire. Manoj Srivastava <[EMAIL PROTECTED]> <http://www.debian.org/~srivasta/> 1024D/BF24424C print 4966 F272 D093 B493 410B 924B 21BA DABB BF24 424C -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]