Date: Tue, 26 Jul 2005 22:17:50 -0700 From: Adam Weinberger <adamw@magnesium.net> To: Mikhail Teterin <mi+mx@aldan.algebra.com> Cc: ports@freebsd.org, Mark Linimon <linimon@lonesome.com> Subject: Re: a portlint dilemma Message-ID: <42E718FE.4070506@magnesium.net> In-Reply-To: <200507262327.24793.mi%2Bmx@aldan.algebra.com> References: <200507262156.43233.mi%2Bmx@aldan.algebra.com> <20050727024231.GA18447@soaustin.net> <200507262327.24793.mi%2Bmx@aldan.algebra.com>
next in thread | previous in thread | raw e-mail | index | archive | help
Mikhail Teterin wrote:
>> There are some chicken-and-egg problems like this in bsd.port.mk (well,
>> the part that is invoked by bsd.port.pre.mk). At one time I understood
>> which each of them was but I haven't looked at them in a long time.
>
> Actually, I think, I found the way. Look at the new GraphicsMagick port.
>
> portlint would've raised hell about this trick (if it new about variable
> modifiers), but that would be groundless. ARCH is not expanded until OPTIONS
> is used (by which time ARCH is already known), so things work just fine.
>
> A more straightforward way would be to:
>
> OPTIONS= MEOW "MEOW?" ${MEOW_on_off}
>
> .include <bsd.port.pre.mk>
>
> .if ${ARCH} == "i386"
> MEOW_on_off= on
> .else
> MEOW_on_off= off
> .endif
>
> .include <bsd.port.post.mk>
>
> This is also fooling portlint -- OPTIONS are, in fact, modified after
> including the ...pre.mk. And there is nothing wrong with that :-)
>
> -mi
That is indeed a very good - albeit sneaky - trick!
The main problem that I find with the OPTIONS chicken-and-egg problem is
that you can't modify OPTIONS based on the presence of files because,
when OPTIONS processing takes place, LOCALBASE and X11BASE aren't
defined. Do you have a good sneaky trick for that case?
# Adam
--
Adam Weinberger
adamw@magnesium.net || adamw@FreeBSD.org
adamw@vectors.cx || adamw@gnome.org
http://www.vectors.cx
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?42E718FE.4070506>