From owner-svn-ports-all@freebsd.org Mon Nov 9 04:26:07 2015 Return-Path: Delivered-To: svn-ports-all@mailman.ysv.freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:1900:2254:206a::19:1]) by mailman.ysv.freebsd.org (Postfix) with ESMTP id 03F72A29D34; Mon, 9 Nov 2015 04:26:07 +0000 (UTC) (envelope-from danfe@freebsd.org) Received: from freefall.freebsd.org (freefall.freebsd.org [IPv6:2001:1900:2254:206c::16:87]) by mx1.freebsd.org (Postfix) with ESMTP id E9E9D122D; Mon, 9 Nov 2015 04:26:06 +0000 (UTC) (envelope-from danfe@freebsd.org) Received: by freefall.freebsd.org (Postfix, from userid 1033) id E7C131FF9; Mon, 9 Nov 2015 04:26:06 +0000 (UTC) Date: Mon, 9 Nov 2015 04:26:06 +0000 From: Alexey Dokuchaev To: Adam Weinberger Cc: Gerald Pfeifer , Guido Falsi , ports-committers@freebsd.org, svn-ports-all@freebsd.org, svn-ports-head@freebsd.org Subject: Re: svn commit: r388648 - in head/net: asterisk11 asterisk13 Message-ID: <20151109042606.GA81532@FreeBSD.org> References: <201506052330.t55NUBZL020965@svn.freebsd.org> <20151109035010.GB64153@FreeBSD.org> <8160653D-E567-42C4-9741-94C593DD62E5@adamw.org> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <8160653D-E567-42C4-9741-94C593DD62E5@adamw.org> User-Agent: Mutt/1.5.24 (2015-08-30) X-BeenThere: svn-ports-all@freebsd.org X-Mailman-Version: 2.1.20 Precedence: list List-Id: SVN commit messages for the ports tree List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Mon, 09 Nov 2015 04:26:07 -0000 On Sun, Nov 08, 2015 at 09:04:13PM -0700, Adam Weinberger wrote: > > On 8 Nov, 2015, at 20:50, Alexey Dokuchaev wrote: > > I've thought of how to word GCC_DESC to make it generic, and had failed. > > IMHO it can mean lots of things throughout the ports and not necessarily > > "want GCC from ports". > > What are some of the other common intentions of a GCC option? Hmm, right now we have about a dozen ports than define GCC_DESC, and most of them in fact mean "build with GCC (from ports)" and some phrase it as "use current GCC". But `cad/ghdl' spells it more specific as "code generator is GCC (mature)". With recent work in GCC camp and it becoming more modular there might be place for GCC_DESC to mean something other than "use GCC from ports". OTOH, those cases should be a minority, so having sensible defaul GCC_DESC makes sense after all. Shall it be "Use GCC from ports" or "Build with current GCC"? > I am a huge, huge believer in options should mean the same thing everywhere. > For example, DOCS should mean "go ahead and install supplied documentation." > It should never ever mean "go ahead and install the entire TeX suite so I can > build the documentation after 45 minutes and 5GB of dependencies." +1 on the anger when DOCS require TeX suite or those annoying Python modules because of Sphinx. DOCS_DESC reads as "build and/or install documentation" right now, but the more ports I have to patch locally the more I'm thinking that DOCS (as implied option) should only install things (or build them with only tools already available), but not be allowed to pull heavy deps. I'm far more in favor of explicit DOXYGEN, SPHINX, TEX2PDF options that are off by default (or conditionally on for batch (cluster) package builders). ./danfe