Date: Fri, 12 Nov 2004 13:01:54 -0800 From: Marcel Moolenaar <marcel@xcllnt.net> To: Poul-Henning Kamp <phk@phk.freebsd.dk> Cc: current@freebsd.org Subject: Re: [TEST] make -j patch [take 2] Message-ID: <20041112210154.GA63387@ns1.xcllnt.net> In-Reply-To: <12448.1100289673@critter.freebsd.dk> References: <20041112195030.GA63153@ns1.xcllnt.net> <12448.1100289673@critter.freebsd.dk>
next in thread | previous in thread | raw e-mail | index | archive | help
On Fri, Nov 12, 2004 at 09:01:13PM +0100, Poul-Henning Kamp wrote: > In message <20041112195030.GA63153@ns1.xcllnt.net>, Marcel Moolenaar writes: > >On Fri, Nov 12, 2004 at 08:57:21PM +0200, Ruslan Ermilov wrote: > >> > >> I now feel like everything that should have been said > >> was said, and this thread should die. ;) > > > >Not quite. It hasn't been said that phk@ is obstinate, disrespectful > >and arrogant. For someone who's advocating mechanisms over policies > >he's also been trying very hard to shove his policies down your > >throat. This also makes him promiscuous (in a non-sexual manner). > > Uhm, it's actually me who is trying to prevent Ruslan for enforcing > his policies on the "make universe" target... He likes a feature so he can control his make universe in a certain way. He isn't forcing his policy upon anybody, just a way for him to do it his way for himself. Now, if his way was totally bogus then I see grounds for unilaterally denying his request. This I don't see. A mechanism to allow a submake to become the first in a new set is an elementary feature that allows us to implement the old behaviour, good or bad, and if nothing else is just a simple way to cover our asses. Since the implementation would not complicate matters to the extend that maintenance becomes impossible, I see no justification for the way the thread evolved. The default behaviour, what you've implemented, is the perfect behaviour to minimize wall-clock time and maximize resource utilization, constrained by the -j setting, but it's not the only way you can allocate resources and spawn jobs. This is where it becomes a policy issue and denying sub-optimal scheduling is therefore a policy decision as well. -- Marcel Moolenaar USPA: A-39004 marcel@xcllnt.net
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20041112210154.GA63387>