ports/150235: sysutils/smartmontools build system bug

Giorgos Keramidas keramida at freebsd.org
Sun Sep 5 19:40:05 UTC 2010

The following reply was made to PR ports/150235; it has been noted by GNATS.

From: keramida at freebsd.org (Giorgos Keramidas)
To: Garrett Wollman <wollman at freebsd.org>
Cc: Alex Samorukov <samm at os2.kiev.ua>, bug-followup at freebsd.org,
        developers at freebsd.org
Subject: Re: ports/150235: sysutils/smartmontools build system bug
Date: Sun, 05 Sep 2010 22:16:46 +0300

 On Sun, 5 Sep 2010 14:53:10 -0400, Garrett Wollman <wollman at FreeBSD.org> wrote:
 > <<On Sun, 05 Sep 2010 20:24:59 +0200, Alex Samorukov <samm at os2.kiev.ua> said:
 >> Very funny.  I don`t think so. Try to argue your position next time,
 >> this may help.
 > Only if someone is willing to actually engage in an argument.
 >> BTW - this "bug" was in smartmontools from early beginning and only
 >> you hit this condition - i think that is very good indicator of its
 >> importance, isn`t it?
 > A bug is a bug.
 >> Your commit will break smartmontools functionality, as i mentioned
 >> before. Thats why i`m against it. Try to read before posting.
 > What commit?  I haven't committed anything.
 >> The required file was never included in the /usr/include. And it was
 >> changed more then once from the beginning.
 > If anything in userland links against a kernel header, then that
 > header needs to be installed in /usr/include, end of story.
 > -I/usr/src/sys is never acceptable in userland code.
 We support building the kernel itself from arbitrary locations, even
 using arbitrary OBJDIR locations.  I don't think userland code should
 depend on /usr/src or /usr/obj as absolute paths.  They are not part of
 the 'published interface' of the kernel and they should never be, as
 long as we want to support building e.g. with MAKEOBJDIRPREFIX set to
 something like '/home/keramida/work/freebsd/obj.i386'.

More information about the freebsd-ports-bugs mailing list