Date: Thu, 30 Apr 1998 22:38:58 -0600 From: Warner Losh <imp@village.org> To: Wolfram Schneider <wosch@cs.tu-berlin.de> Cc: Bruce Evans <bde@FreeBSD.ORG>, current@FreeBSD.ORG Subject: Re: cvs commit: src/share/mk bsd.man.mk Message-ID: <199805010438.WAA04983@harmony.village.org> In-Reply-To: Your message of "Wed, 29 Apr 1998 11:02:29 %2B0200." <19980429110229.A8210@freno.cs.tu-berlin.de> References: <19980429110229.A8210@freno.cs.tu-berlin.de> <199804271508.IAA14836@freefall.freebsd.org>
next in thread | previous in thread | raw e-mail | index | archive | help
In message <19980429110229.A8210@freno.cs.tu-berlin.de> Wolfram Schneider writes: : A non-recursive make should perform make -jN much better. I don't buy much of what that paper says. There are often times that some files need to be compiled with some options, while others in the tree need to be compiled with others. A single, huge makefile, even if it is disguised as lots of includes, makes it hard to say I want these CFLAGS for this set of files, and those CFLAGS for that set of files. And it tends to force files to be uniquely named, which experience has shown is a bad thing. The paper does bring up several valid, good points. However, I don't think it supports its thesis very well. It also ignores make depend as a solution to the problems it describes. Then again, I'm biased. I've written makefiles and build systems for six different projects now and I've been burned, in one way or another, by all possible problems. So far gmake and/or bmake suck less than anything else I've used (with Imake sucking only a little bit more than these two, but its learning curve is much steeper). Warner To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-current" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199805010438.WAA04983>