Date: Thu, 3 Jul 2014 12:47:56 +0200 From: Tijl Coosemans <tijl@FreeBSD.org> To: Kubilay Kocak <koobs@FreeBSD.org> Cc: Randy Bush <randy@psg.com>, python@FreeBSD.org Subject: Re: ports/169276 Message-ID: <20140703124756.5177f63b@kalimero.tijl.coosemans.org> In-Reply-To: <53B52633.9000000@FreeBSD.org> References: <m238ejrm4r.wl%randy@psg.com> <53B4FC59.9000706@FreeBSD.org> <20140703112112.120f0db3@kalimero.tijl.coosemans.org> <53B52633.9000000@FreeBSD.org>
next in thread | previous in thread | raw e-mail | index | archive | help
On Thu, 03 Jul 2014 19:45:23 +1000 Kubilay Kocak wrote: > The cause of the module failures was due to a > CFLAGS="-I/usr/local/include" in /etc/make.conf so that parts all > sorted. This is why we kept the comment in there about using CPPFLAGS vs > CFLAGS (removed in your change) I see, that will probably cause subtle problems in other ports as well. The reason I removed the comment about CPPFLAGS is because it seemed redundant to me. Flags like -I and -D should always go into CPPFLAGS. Adding them to CFLAGS is the special case. > I understand the rationale for the move from LDFLAGS -> LIBS, but I am > still concerned due to the plethora of flags based issues we've had with > Python in the past > > This is especially with regards to the right flags turning up in the > right places for shared extensions (such as within python-config output) > and doubly-so for the substantial number of workarounds that we've had > to retain and maintain locally in the port (see the post-configure and > pre-patch targets for lang/python27 for example) Ok, I think everything is ok at the moment, but feel free to contact me if some issue comes up.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20140703124756.5177f63b>