java/openjdk6 fails to build..
glewis at eyesbeyond.com
Sat Dec 29 00:46:40 UTC 2012
On Fri, Dec 28, 2012 at 03:49:56PM -0500, Eitan Adler wrote:
> On 28 December 2012 15:24, Greg Lewis <glewis at eyesbeyond.com> wrote:
> > On Fri, Dec 21, 2012 at 07:27:22PM -0500, Eitan Adler wrote:
> >> On 21 December 2012 19:16, David Demelier <demelier.david at gmail.com> wrote:
> >> > Yes, I wanted to answser, but the man says that MAKE_JOBS are disabled by
> >> > default, I tried to build with MAKE_JOBS=1 and it works, the problem did
> >> > appear when MAKE_JOBS=6 was set. Can you reproduce the issue?
> >> I want to mark this MAKE_JOBS_UNSAFE=yes and remove the current hack
> >> Any objection from java@ ?
> >> > I had MAKE_JOBS=6 in my /etc/make.conf instead of MAKE_JOBS_NUMBER though
> >> Exactly. MAKE_JOBS is not user-settable. MAKE_JOBS_NUMBER is. IMHO
> >> MAKE_JOBS should start with an _.
> > I think that's right. I have no objection. I think whats in openjdk7
> > is closer to what it should be.
> What is in openjdk7 seems bogus and it isn't clear why it is there:
> MAKE_JOBS_UNSAFE= yes
> is defined implying that MAKE_JOBS should always be 1 (and -j never set)
> but it does some weird things with the global variables
> MAKE_JOBS_NUMBER and DISABLE_MAKE_JOBS.
> Can the generic handling in b.p.m not work?
> IMHO MAKE_JOBS_SAFE=yes should be defined and either b.p.m or b.java.m
> should be handling the special build code.
MAKE_JOBS_SAFE=yes should not be defined since neither of the ports are
safe for multiple make jobs. As an exception, the HotSpot portion of the
build is and that is what the other portion handles.
Greg Lewis Email : glewis at eyesbeyond.com
Eyes Beyond Web : http://www.eyesbeyond.com
Information Technology FreeBSD : glewis at FreeBSD.org
More information about the freebsd-ports