Date: Fri, 21 Jan 2005 06:51:10 -0600 From: "Conrad J. Sabatier" <conrads@cox.net> To: Edwin Groothuis <edwin@mavetju.org> Cc: Kris Kennaway <kris@obsecurity.org> Subject: Re: Bug in ports system's DISTFILES handling? Message-ID: <20050121065110.100b445f@dolphin.local.net> In-Reply-To: <20050119072607.GJ1033@k7.mavetju> References: <20050117131440.2486ac2a@dolphin.local.net> <41EC175F.8060805@tvog.net> <20050117225702.GA31708@xor.obsecurity.org> <20050119005427.70f7b3ac@dolphin.local.net> <20050119072607.GJ1033@k7.mavetju>
next in thread | previous in thread | raw e-mail | index | archive | help
On Wed, 19 Jan 2005 18:26:07 +1100, Edwin Groothuis <edwin@mavetju.org>
wrote:
> On Wed, Jan 19, 2005 at 12:54:27AM -0600, Conrad J. Sabatier wrote:
> > This just seems less than intuitive, if you ask me, especially given
> > that the += operator does work with other variables without
> > requiring the explicit definition of an initial value first.
>
> Euhm... you need to keep a couple of things seperated:
>
> - The ports framework (bsd.port.mk) consists of two parts:
>
> - One to set the basic variables which are not directly port
> related, for example OSVERSION, WANT_GNOME and LOCALBASE.
> This is called "bsd.port.pre.mk".
>
> - One to set the basic variables which are directly port
> related, for example LIB_DEPENDS and the pre/do/post targets.
> This is called "bsd.port.post.mk".
>
> The definition of DISTFILES itself is defined in the second part.
> Why? I don't know. It shouldn't have, because all the necessary
> information to build DISTFILES (DISTNAME and EXTRACT_SUFX) are
> defined in the "bsd.port.pre.mk" part.
>
> > I mean, if this were something that was consciously decided on,
> > that's one thing, but the lack of consistency would seem to indicate
> > that it's just not as well implemented as it could/should be.
>
> I wouldn't go that far. I don't think you're the first who steps
> into this problem and goes "euhm... this doesn't make sense", but
> I think you're the first who goes "euhm... guys! this doesn't make
> sense!"
>
> For example see biology/garlic:
> DISTFILES= ${DISTNAME}${EXTRACT_SUFX}
> .if !defined(NOPORTDOCS)
> DISTFILES+= ${DISTNAME}-doc${EXTRACT_SUFX}
> .endif
>
> I won't say that
> -DISTFILES= ${DISTNAME}${EXTRACT_SUFX}
> .if !defined(NOPORTDOCS)
> DISTFILES+= ${DISTNAME}-doc${EXTRACT_SUFX}
> .endif
> should work, but I would say that:
> -DISTFILES= ${DISTNAME}${EXTRACT_SUFX}
> +.include <bsd.port.pre.mk>
> .if !defined(NOPORTDOCS)
> DISTFILES+= ${DISTNAME}-doc${EXTRACT_SUFX}
> .endif
> should work.
>
> Or even:
> DISTFILES= ${DISTNAME}${EXTRACT_SUFX}
> .if !defined(NOPORTDOCS)
> +EXTRA_DISTFILES+=${DISTNAME}-doc${EXTRACT_SUFX}
> .endif
> which would work as EXTRA_PATCHES.
>
> Or to make it easiest of all:
> -DISTFILES= ${DISTNAME}${EXTRACT_SUFX}
> +DISTFILES= ${DEFAULTDISTFILE}
> .if !defined(NOPORTDOCS)
> DISTFILES+= ${DISTNAME}-doc${EXTRACT_SUFX}
> .endif
> and then DEFAULTDISTFILE is in the "bsd.port.pre.mk" defined. This
> way it's easy to for the Makefile of chinese/xcin25:
> DISTFILES= ${DEFAULTDISTFILE}
> chewing-2002Jan07-snapshot.tar.gz
> No confusion about things here.
>
> I would go for the last one, and volunteer to send patches when
> there is some consensus.
Yes, I like that last version, too. I say go for it. :-)
--
Conrad J. Sabatier <conrads@cox.net> -- "In Unix veritas"
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20050121065110.100b445f>