On Mon, 2003-06-16 at 19:15, James Troup wrote: > Colin Walters <[EMAIL PROTECTED]> writes: > >> If Build-Depends-Indep were > >> installed to satisfy 'build' their entire raison d'etre would be > >> voided. > > > > The buildds[...] > > invoke 'dpkg-buildpackage -B'.
Ah, OK. > > This all goes for dpkg-buildpackage too, of course. > > Fine and dandy; feel free to talk to the dpkg folks. I'm sure you'll > get somewhere before 2038, if not a year or two sooner. Sigh...I wish I could just pretend you were joking and laugh, but... > >> That's not necessarily a bad thing, but it's orthogonal to > >> this bug. > > > > I don't think it's orthogonal at all. > > That == "invalidating build-depends-indep's purpose in life" - which > is orthogonal. I don't think we're invalidating its purpose; Build-Depends-Indep is still useful for invoking debian/rules binary-indep. > As I've said before: Policy is either meant to document current > practice or it's not. If it is, then it's wrong because the de facto > implementation of build-depends disagrees with it _and always has > done_ (or at least, it's not the buildds which have changed). If it's > not, well... Well, probably the main reason we haven't hit this before is that most people don't bother to do a proper arch/indep split in their packages, and just stuff everything into Build-Depends.
