Re: pkg/57974: Depends conflict for firefox AND postgres

["David H. Gutteridge" <david%gutteridge.ca@localhost>]

The following reply was made to PR pkg/57974; it has been noted by GNATS.

From: "David H. Gutteridge" <david%gutteridge.ca@localhost>
To: gnats-bugs%netbsd.org@localhost
Cc: 
Subject: Re: pkg/57974: Depends conflict for firefox AND postgres
Date: Sat, 09 Mar 2024 20:46:40 -0500

 On Thu, 2024-03-07 at 23:20 +0000, Joe Mifsud wrote:
 > The following reply was made to PR pkg/57974; it has been noted by
 > GNATS.
 >=20
 > From: Joe Mifsud <jmifsud%posteo.net@localhost>
 > To: gnats-bugs%netbsd.org@localhost
 > Cc:=20
 > Subject: Re: pkg/57974: Depends conflict for firefox AND postgres
 > Date: Thu, 07 Mar 2024 23:16:59 +0000
 >=20
 > =C2=A0On 2024-03-07 01:25, David H. Gutteridge wrote:
 > =C2=A0> The following reply was made to PR pkg/57974; it has been noted b=
 y=20
 > =C2=A0> GNATS.
 > =C2=A0>=20
 > =C2=A0> From: "David H. Gutteridge" <david%gutteridge.ca@localhost>
 > =C2=A0> To: Gnats Bugs <gnats-bugs%netbsd.org@localhost>
 > =C2=A0> Cc:
 > =C2=A0> Subject: Re: pkg/57974: Depends conflict for firefox AND postgres
 > =C2=A0> Date: Wed, 06 Mar 2024 20:23:41 -0500
 > =C2=A0>=20
 > =C2=A0>=C2=A0 On Tue, 05 Mar 2024 at 23:14:36 +0000, Joe Mifsud wrote:
 > =C2=A0>=C2=A0 > Thank you for your reply and consideration. This is my mk=
 .conf:
 > =C2=A0>=C2=A0 >
 > =C2=A0>=C2=A0 >=C2=A0=C2=A0=C2=A0 UNPRIVILEGED=3Dyes
 > =C2=A0>=C2=A0 >
 > =C2=A0>=C2=A0 > PREFER_PKGSRC is not set and pkgin is not installed.=C2=
 =A0 Maybe the=20
 > =C2=A0> issue
 > =C2=A0>=C2=A0 > is that I am not using an official release of pkgsrc? On =
 my
 > system,
 > =C2=A0>=C2=A0 > security/heimdal is pulled in from this graph branch:
 > =C2=A0>=C2=A0 >
 > =C2=A0>=C2=A0 > www/firefox -> x11/gtk3 -> print/libcups (defaults w kerb=
 eros)
 > =C2=A0>=C2=A0 >=C2=A0=C2=A0=C2=A0 -> mk/kerb5.buildlink.mk -> security/he=
 imdal
 > =C2=A0>=20
 > =C2=A0>=C2=A0 Hmm. It shouldn't be that you're tracking -current, and not=
 hing to
 > do
 > =C2=A0>=C2=A0 with pkgin. I'm also testing with -current.
 > =C2=A0>=20
 > =C2=A0>=C2=A0 A couple of questions:
 > =C2=A0>=20
 > =C2=A0>=C2=A0 1. Does security/heimdal show other packages registered as
 > dependants?
 > =C2=A0>=C2=A0 Or just print/libcups?
 > =C2=A0>=C2=A0 2. Are /usr/include/krb5.h and /usr/include/krb5/krb5.h ins=
 talled?
 > I'd
 > =C2=A0>=C2=A0 imagine they must be, as they're part of the "comp" set tha=
 t
 > provides
 > =C2=A0>=C2=A0 headers.
 > =C2=A0>=20
 > =C2=A0>=C2=A0 If you go into print/libcups, you can run tests like:
 > =C2=A0>=20
 > =C2=A0>=C2=A0 $ make show-var VARNAME=3DIS_BUILTIN.heimdal
 > =C2=A0>=C2=A0 yes
 > =C2=A0>=C2=A0 $ make show-var VARNAME=3DH_HEIMDAL
 > =C2=A0>=C2=A0 /usr/include/krb5/krb5.h
 > =C2=A0>=20
 > =C2=A0>=C2=A0 (Those are my results on 9.3_STABLE. I don't have UNPRIVILE=
 GED
 > set,=20
 > =C2=A0> and
 > =C2=A0>=C2=A0 don't see how it would make any difference from code inspec=
 tion,
 > but I
 > =C2=A0>=C2=A0 set it to "yes" to test that, and those built-in tests beha=
 ve the
 > same
 > =C2=A0>=C2=A0 for me either way.)
 > =C2=A0>=20
 > =C2=A0>=C2=A0 Dave
 > =C2=A0
 > =C2=A0Answers:
 > =C2=A0- 1 -
 > =C2=A0I don't know the best way to check that. Heres a guess:
 > =C2=A0$ pkg_info -r heimdal
 > =C2=A0Information for heimdal-7.8.0nb7:
 > =C2=A0
 > =C2=A0Full required by list:
 > =C2=A0
 > =C2=A0
 > =C2=A0$
 > =C2=A0It's interesting that it is not recorded as a dependency of=20
 > =C2=A0print/libcups.
 > =C2=A0
 > =C2=A0
 > =C2=A0- 2 -
 > =C2=A0Yup, they're located where you indicated
 > =C2=A0$ ls /usr/pkg/include/ | grep krb5
 > =C2=A0krb5
 > =C2=A0$ ls /usr/include/ | grep krb5.h
 > =C2=A0krb5.h
 > =C2=A0$ ls /usr/include/krb5 | grep krb5.h
 > =C2=A0krb5.h
 > =C2=A0
 > =C2=A0And the output of these commands are the same as the ones you=20
 > =C2=A0demonstrated.
 > =C2=A0$ cd /usr/pkgsrc/print/libcups/
 > =C2=A0$ make show-var VARNAME=3DIS_BUILTIN.heimdal
 > =C2=A0yes
 > =C2=A0$ make show-var VARNAME=3DH_HEIMDAL
 > =C2=A0/usr/include/krb5/krb5.h
 > =C2=A0
 > =C2=A0Curious what you think is happening
 
 pkg_info is telling us that heimdal is installed from pkgsrc, but in
 fact, it has no dependants. That means you should simply be able to
 pkg_delete it, and could probably rebuild with libcups depending on
 the base version of heimdal, avoiding the conflict without even
 resorting to non-default options. (Should you choose to do so.)
 
 I'm not sure how much time you (or I) want to spend examining this,
 but another thing you could look at is why heimdal was built from
 pkgsrc in the first place.
 
 I don't know where your pkgdb is located (/usr/pkg/pkgdb or /var/db/pkg
 or somewhere else), but you could find the heimdal directory there,
 "ls -l" its contents, and see if there's a "+REQUIRED_BY" file
 (presumably with a current size of zero) present. If so, that would tell
 you at some point, pkgsrc heimdal was built as a dependency of something
 else, and that something else was subsequently deleted. It doesn't seem
 the dependency chain you indicated before is actually the cause of it
 being built. (There may have been another package that had some chain
 of dependencies that weren't satisfied by base heimdal, or base openssl,
 or such.)
 
 Dave