Re: gcc 3.2 transition in unstable
On Mon, 2003-01-06 at 13:00, Matthias Klose wrote:
> yes, the build-depends should include "gcc (>= 3:3.2)", and
> build-essentials should be changed as well.
That's what I thought too, but Ryan and Colin Watson disagreed. Can you
guys elaborate?
IIRC, Colin's argument was that most programs could equally well be
compiled with gcc 2.95 or whatever, so technically the program doesn't
depend on gcc 3.2. However, my response to that is that it is the
explicit intention of the package maintainer to have the program
compiled with gcc 3.2 (especially important for C++), so that should be
expressed in the Build-Depends.
Actually, probably a better alternative than adding gcc directly to the
Build-Depends is to say:
Build-Depends: build-essential (>= 9-1)
I am open to further discussion about this, but right now I am inclined
at least to do so in my C++ packages (and also to update
build-essential).
Reply to: