Date: Thu, 16 Mar 2006 20:58:17 +0100 (CET) From: Gerald Pfeifer <gerald@pfeifer.com> To: "[LoN]Kamikaze" <LoN_Kamikaze@gmx.de> Cc: freebsd-ports@freebsd.org Subject: Re: lang/gcc41 - libjava Message-ID: <Pine.LNX.4.61.0603162051300.7310@acrux.dbai.tuwien.ac.at> In-Reply-To: <4418A188.8060200@gmx.de> References: <4418A188.8060200@gmx.de>
next in thread | previous in thread | raw e-mail | index | archive | help
On Thu, 16 Mar 2006, [LoN]Kamikaze wrote: > Today I removed the lines > # FIXME: As of 20051202, installing libgcj nearly kills 1GB machines. > WITHOUT_JAVA= yes > > from the ports Makefile. The previous version of the Port gave me > "virtual memory exhausted" messages, even with 800MB of free memory > available. However this time a > > # make makesum build > > Ran through without troubles. > > A "make install" gave me the message again, but forcing an installation > with "make -i install" seems to work. I only compiled a small "Hello > World!", but that is better than nothing, I guess. Thanks for this input! The problem I see is that "make -i" will not only ignore out-of-memory conditions but *any* errors occuring. Given that we generally test updates before releasing to our users we might get away with that, but these builds will not be reproducible: depending on the memory in her build machine, a user will get different results, some of which may match your positive experience, some of which may not. As an alternate approach, Maho-san and me are currently discussing to use the new gmake-devel port which has fixes to address the situation on the gmake side for the time being. Gerald
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?Pine.LNX.4.61.0603162051300.7310>