Date: Sun, 3 Apr 2011 01:57:04 -0400 From: Sahil Tandon <sahil@FreeBSD.org> To: Doug Barton <dougb@FreeBSD.org> Cc: Dirk Meyer <dinoex@FreeBSD.org>, cvs-ports@FreeBSD.org, cvs-all@FreeBSD.org, ports-committers@FreeBSD.org Subject: Re: cvs commit: ports/graphics/netpbm Makefile Message-ID: <20110403055703.GA81066@magic.hamla.org> In-Reply-To: <4D978D14.406@FreeBSD.org> References: <201104021205.p32C5Y8g082718@repoman.freebsd.org> <20110402155230.GA80090@magic.hamla.org> <4D978D14.406@FreeBSD.org>
next in thread | previous in thread | raw e-mail | index | archive | help
On Sat, 2011-04-02 at 13:54:44 -0700, Doug Barton wrote: > On 04/02/2011 08:52, Sahil Tandon wrote: > >On Sat, 2011-04-02 at 12:05:34 +0000, Dirk Meyer wrote: > > > >>dinoex 2011-04-02 12:05:34 UTC > >> > >> FreeBSD ports repository > >> > >> Modified files: > >> graphics/netpbm Makefile > >> Log: > >> - drop CONFLICTS with megetty > >> Submitted by: olgeni > > > >Does PORTREVISION need to be bumped when we change CONFLICTS? > > When adding CONFLICTS I would say definitely yes, since they are > recorded in the package. When dropping a conflict I personally would > be less inclined to bump it, unless I knew that there was an > existing problem that is being solved by the update. I share your rationale for the most part, but I am still unclear about what some might call an 'edge' case. To wit, let us assume that I already have port foo installed; foo conflicts with bar (as noted in CONFLICTS within bar's Makefile). Therefore, whenever I try to install bar, I am notified about the conflict and unable to proceed. A few months pass, and for whatever reason, the bar maintainer removes the conflict with foo, but does not bump PORTREVISION. The default bar package, from whichever repository I had originally obtained it, will still contain the (now incorrect) conflict with foo. Is my understanding or interpretation of this example flawed? Does this type of situation occur so seldom that it is not worth the bump in the majority of cases when a conflict is dropped? Should we similarly not bump PORTREVISION if something is dropped from RUN_DEPENDS (which, like CONFLICTS, is recorded in the package)? Or perhaps I've missed a nuance that makes my question foolish. In any case, please bludgeon me with a cluebat so I stop pestering you with these questions. :-) -- Sahil Tandon <sahil@FreeBSD.org>
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20110403055703.GA81066>