Date: Thu, 22 Mar 2001 09:33:43 +0200 From: Ruslan Ermilov <ru@FreeBSD.org> To: Bruce Evans <bde@zeta.org.au> Cc: Robert Watson <rwatson@FreeBSD.org>, cvs-committers@FreeBSD.org, cvs-all@FreeBSD.org Subject: Re: cvs commit: src/sbin Makefile.inc src/sbin/adjkerntz Makefile src/sbin/atacontrol Makefile src/sbin/atm/atm Makefile src/sbin/atm/fore_dnld Makefile src/sbin/atm/ilmid Makefile src/sbin/badsect Makefile src/sbin/camcontrol Makefile ... Message-ID: <20010322093343.A53063@sunbay.com> In-Reply-To: <Pine.BSF.4.21.0103221737120.9843-100000@besplex.bde.org>; from bde@zeta.org.au on Thu, Mar 22, 2001 at 05:49:14PM %2B1100 References: <Pine.NEB.3.96L.1010321231155.3074D-100000@fledge.watson.org> <Pine.BSF.4.21.0103221737120.9843-100000@besplex.bde.org>
next in thread | previous in thread | raw e-mail | index | archive | help
On Thu, Mar 22, 2001 at 05:49:14PM +1100, Bruce Evans wrote:
> On Wed, 21 Mar 2001, Robert Watson wrote:
>
> > This change appears to break a checkout of just a single directory (say,
> > src/usr.sbin/watch) followed by a make in that directory, as no
> > ../Makefile.inc exists because that wasn't checked out. In fact, on my
>
> I wish I had thought of that when I reviewed it. This seems to be
> impossible to fix properly without reverting all the changes. There
> just isn't enough context to determine the correct manpage extension
> when the source directory is not in its usual tree.
>
> > -CURRENT box from yesterday, it appears to be broken even with a
> > ../Makefile.inc present. Am I misunderstanding / have an oddly sync'd
> > world/source tree?
>
> You might need Makefile.inc's all the way up to the level below `src',
> but ../Makefile.inc is already at that level for `watch'.
>
> > fledge:/data/fbsd-commit/src/usr.sbin/watch> make
> > Warning: Object directory not changed from original
> > /data/fbsd-commit/src/usr.sbin/watch
> > cc -O -pipe -c watch.c
> > cc -O -pipe -o watch watch.o -ltermcap
> > make: don't know how to make watch.1. Stop
>
> It is an old problem that `make install' doesn't work outside of an
> almost full tree (mainly because BINDIR is not set). Failing at build
> time is more of a problem.
>
The only workaround I can think of without reverting this all is -DNOMAN.
I must admit I overlooked this also. :-(
I am actually thinking of replacing the MAN[0-9] stuff with the single
MAN variable, like in OpenBSD. Hopefully, in the absence of the MAN
specification, I can come up with the code that would check the ${.CURDIR}
for ${PROG}.([0-9]|1aout).
Cheers,
--
Ruslan Ermilov Oracle Developer/DBA,
ru@sunbay.com Sunbay Software AG,
ru@FreeBSD.org FreeBSD committer,
+380.652.512.251 Simferopol, Ukraine
http://www.FreeBSD.org The Power To Serve
http://www.oracle.com Enabling The Information Age
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe cvs-all" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20010322093343.A53063>