On Sat, Feb 12, 2011 at 08:22:39PM +0100, Bernhard R. Link wrote: > * Don Armstrong <d...@debian.org> [110211 23:01]: > > 3) uniform, known build environments > I think is a major disadvantage of this suggestion.
I'm unconvinced by your (implicit) argument that switching to an uniform
build environment will make the current testing of packages built in
"weird" environments any worse than the present situation.
Package maintainers are already expected to upload packages built in
clean environment. If they are not doing it, then we have a problem of
best practices which are followed by the developer community. Relying on
instances of that problem as test bench for weird build environments
doesn't look sane at all.
> Free Software is about being able to modify what you run. The day a
> user can no longer simply do a
Agreed. But a (binary) distribution is about offering, first of all, an
uniform set of *binary* packages. If we have to balance our priorities
among (1) having that uniformity and (2) ensuring that packages build
properly in weird environment, then I would favor (1) over (2), no
doubt.
But actually no one is saying that we should have that trade-off. We can
work on (2) as well. Maybe you---as the only one who has up to now
voiced against this change---are interested in resurrecting some of the
past initiatives that tried to spot missing Build-Conflicts? If there is
energy to fight that battle, then going that way seems way more
promising than relying in random build environments prepared by
maintainers which are not using {cow,p}builder, when they should.
Cheers.
--
Stefano Zacchiroli -o- PhD in Computer Science \ PostDoc @ Univ. Paris 7
zack@{upsilon.cc,pps.jussieu.fr,debian.org} -<>- http://upsilon.cc/zack/
Quando anche i santi ti voltano le spalle, | . |. I've fans everywhere
ti resta John Fante -- V. Capossela .......| ..: |.......... -- C. Adams
signature.asc
Description: Digital signature
