On Tuesday, February 26, 2002, at 06:18 , David R. Morrison wrote: > Under your system, let's say I introduce foo-3.0.0 which is not > backwards > compatible. Now all of the other developers need to revise their > packages > immediately, because their packages were saying > Depends: foo (>= 2.0.0) > but now they need to say > Depends: foo (>= 2.0.0 << 3.0.0) > See the problem?
I was proposing that the second one be required in our packaging policy, or implemented by fink (Maybe 'Depends foo (== MAJOR)'?) Then all packages that use the new system would have that to start with the lower one. > Under your system, let's say that there is already a package named foo > which has binaries, headers and libraries. Now we revise, creating > new packages foo-bin, foo, foo-shlibs. Now all of the other developers > need to revised immediately, because their packages said > Depends: foo > but (under your proposal) the new foo only depends on foo-shlibs, not > on foo-bin. So if their package actually uses foo-bin, it will break See below > The second one is what I was trying to address with this recent email. > Yes, we want the other developers to revise, but we can't assume they > will do so right away. So to *temporarily* put > Depends: foo-shlibs, foo-bin > into the foo package makes sense to me. Once all other developers have > revised, the foo package can be revised to say "Depends: foo-shlibs". I agree with this, that for the interim period, that should happen, but I think that BuildDepends would be better for foo-bin, because that way, foo-shlibs gets built and installed, then foo-bin, then foo. After all are built, the user could choose to remove foo-bin. But never mind, that won't work for apt-get, sorry! =) > -- Dave Kyle Moffett _______________________________________________ Fink-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/fink-devel
