Re: Build-Depends listed as a should not a must?
On Sun, Feb 16, 2003 at 03:25:48AM +0000, Colin Watson wrote:
> I don't see how it closes #87510, though. #87510 is a request for
> build-dependencies to be made mandatory: the whole should vs. may thing
> was a digression somewhere. When you said you'd be closing #87510, I
> assumed you meant that you were making build-dependencies a must, which
> I do think is still (unfortunately) premature.
I disagree -- it seems to me that the should vs. may issue is what really
prompted #87510: people wanted to make sure build-dependencies have a strong
basis in policy. The problem is, they tried to achieve it with an overly
disruptive method. :) Unless I missed something, nobody really disputed
this objection in almost two years the bug has been open, so I think it's
fair to assume that such reasoning is correct.
FYI this is what I put in the changelog entry:
* Clarified that using Build-Depends for build-dependencies is not a "may"
but a "should", added proper linking among various sections dealing with
the dependencies so that there is no confusion, closes: #87510.
If any one of those poor, misguided people ;) still thinks that the
requirement should be a "must" (in _our_ meaning, not RFC "MUST"),
please file a new bug, as it's quite unproductive to have to sift
through a 152-page bug log which mostly goes back-and-forth with
explanations how policy should work, occasionally sprinkled with
offtopic stuff, too.
--
2. That which causes joy or happiness.
Reply to: