Re: gcc 3.2 transition in unstable
On Tue, 7 Jan 2003, Adrian 'Dagurashibanipal' von Bidder wrote:
> Packages should build-depend on what they should build-depend.
> -- Santiago Vila on debian-devel
I meant to say: packages should build-depend on their real build-dependencies.
It does not make sense to build-depend on build-essential (>= someversion)
if the real build-dependency (what the package *really* needs) is g++-3.2.
A build-dependency like build-essential (>= someversion) does nothing
but obfuscate the real meaning and it's artificial.
Suppose someone compiles g++-3.2 for woody and tries to compile a
source package from unstable which needs g++ 3.2. Why in earth would
someone need to install an empty package from unstable, one that
has no functionality by itself, to satisfy a build-dependency?
Reply to: