Date: Tue, 18 Mar 2003 15:46:29 -0500 (EST) From: "Brian Szymanski" <bks10@cornell.edu> To: <allbery@ece.cmu.edu> Cc: <freebsd-stable@freebsd.org> Subject: Re: setting kern.ngroups Message-ID: <1372.192.168.1.5.1048020389.squirrel@wuhjuhbuh.afraid.org> In-Reply-To: <1048011633.2072.1.camel@pyanfar.ece.cmu.edu> References: <4394.192.168.1.5.1048003806.squirrel@wuhjuhbuh.afraid.org> <1048006379.40809.2.camel@rushlight.kf8nh.apk.net> <20030318172857.GA22835@localhost> <1048011633.2072.1.camel@pyanfar.ece.cmu.edu>
next in thread | previous in thread | raw e-mail | index | archive | help
> On Tue, 2003-03-18 at 12:28, Brian Szymanski wrote: >> > Raising the maximum number of groups requires changes all over the >> place; even if you find them all and rebuild the world (yes, libc >> depends on it as well) you'll find that any program that looks at >> the group vector will blow up because it only has space reserved for >> 16 groups. You don't want to go there. >> >> Aren't these programs broken by not using NGROUPS_MAX from >> syslimits.h? > > My point is that changing syslimits.h doesn't help existing compiled > code at all; you need to recompile everything that touches the group > vector, which may be more code than you expect. (Arguably they should > use a runtime method of sizing the group vector but I've seen very few > programs that do.) Ahhh, no problem. Fortunately we have source for all the code we are running on the machine... Which brings up another question. Does anyone know of a good way to rebuild all ports, without dealing with dependency hell? Peace, Brian Szymanski bks10@cornell.edu To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-stable" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?1372.192.168.1.5.1048020389.squirrel>