Date: Wed, 09 Oct 1996 10:11:16 -0500 From: Steve Price <sprice@hiwaay.net> To: Wolfram Schneider <wosch@freebsd.org> Cc: current@freebsd.org Subject: Re: .depend Message-ID: <325BC094.1F74CF80@hiwaay.net> References: <199610082021.WAA06329@campa.panke.de> <l0301051cae80a82b2968@[208.2.87.4]> <199610091312.PAA01354@campa.panke.de>
next in thread | previous in thread | raw e-mail | index | archive | help
Wolfram Schneider wrote: > > Richard Wackerbarth writes: > >>I think we should add an enviroment variable (e.g. DEPENDFILE) to > >>make(1) for `.depend'. mkdep -f foo support different depend files. > > > >Why change "make"? > > Because .depend is wired in make(1). There is no way > to stop make(1) to read .depend if exists. There is no way > to use an other depend file, e.g. `.depend.i486' or > `.depend.hostname'. > > Example > $ mkdep <cpp options> <files> > $ mkdep -f .depend.debug -DDEBUG <cpp options> <files> > $ mkdep -f .depend.i386 -Di386 <cpp options> <files> > > You can now include .depend.debug in ./Makefile, but make(1) also > read .depend because it exists. > > Wolfram Then the best thing would be to take the hard-coded value of .depend out of make and put a rule in one of the *.mk files. IMHO including .depend as a rule in make is not good. make should be a generic tool that can be used in any environment for doing everything from writing books to writing software. Adding another hard-coded rule/value in make(1) makes it even more biased in the build/maintenance of software programs. I would be happy (if there are no strong objections) to remove the rule from make and put it in one of the *.mk files, but I really don't think always including .depend* files from within make are the best answer for a generic build tool. Just my opinion, Steve
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?325BC094.1F74CF80>