Subject: Re: pkg/32202
To: None <rillig@netbsd.org, gnats-admin@netbsd.org,>
From: None <joerg@britannica.bec.de>
List: pkgsrc-bugs
Date: 12/03/2005 21:10:02
The following reply was made to PR pkg/32202; it has been noted by GNATS.
From: joerg@britannica.bec.de
To: gnats-bugs@netbsd.org
Cc:
Subject: Re: pkg/32202
Date: Sat, 3 Dec 2005 22:08:46 +0100
On Sat, Dec 03, 2005 at 09:05:02PM +0000, Krister Walfridsson wrote:
> The following reply was made to PR pkg/32202; it has been noted by GNATS.
>
> From: Krister Walfridsson <cato@df.lth.se>
> To: gnats-bugs@netbsd.org
> Cc:
> Subject: Re: pkg/32202
> Date: Sat, 3 Dec 2005 22:04:50 +0100 (CET)
>
> On Sat, 3 Dec 2005 joerg@britannica.bec.de wrote:
>
> > On Sat, Dec 03, 2005 at 08:15:07PM +0000, Krister Walfridsson wrote:
> > > Hmm. mk/bulk/printdepends is not completely happy:
> >
> > That's a different problem. Those packages don't have a working
> > "show-depends-dirs", since e.g. PKG_FAIL or so is set. The message is a
> > bit misleading, but the behaviour correct.
>
> Well, this behavior is fallout of fixing the PR.
>
> And I claim it is a bug -- the mk/bulk/printdepends should not
> write any errors unless the errors are real problems that
> should be fixed. I have no problem with "show-depends-dirs"
> writing errors in this case, but then "printdepends" need to
> suppress the errors, or use some other mechanism.
It is a bug, the question is just of which part of the infrastructure.
For example, it did make the missing default option in chat/unrealircd
easy to spot and I'm inclined to consider the other packages buggy as
well for the same reason.
Joerg