Date: Wed, 28 Feb 2001 13:53:51 -0700 From: Warner Losh <imp@harmony.village.org> To: John Baldwin <jhb@FreeBSD.org> Cc: cvs-all@FreeBSD.org, cvs-committers@FreeBSD.org, Wesley Morgan <morganw@chemikals.org>, obrien@FreeBSD.org, Will Andrews <will@physics.purdue.edu>, Dag-Erling Smorgrav <des@ofug.org> Subject: Re: cvs commit: src/gnu/usr.bin/binutils/ar Makefile src/gnu/usr Message-ID: <200102282053.f1SKrpd43256@harmony.village.org> In-Reply-To: Your message of "Wed, 28 Feb 2001 12:31:11 PST." <XFMail.010228123111.jhb@FreeBSD.org> References: <XFMail.010228123111.jhb@FreeBSD.org>
next in thread | previous in thread | raw e-mail | index | archive | help
In message <XFMail.010228123111.jhb@FreeBSD.org> John Baldwin writes: : : On 28-Feb-01 Warner Losh wrote: : > In message <xzpn1b6py8b.fsf@flood.ping.uio.no> Dag-Erling Smorgrav writes: : >: Will Andrews <will@physics.purdue.edu> writes: : >: > Why make make(1) statically linked? : >: : >: Because a) you need it to recover from e.g. libc fuckups and b) it : >: forks and execs a *lot*, and according to Bruce (I haven't verified : >: this myself) programs that do that (e.g. shells) perform better and : >: consume less system resources if they're statically linked. : > : > So long as it is not forced unconditionally to be static. : > : > We use make in our embedded devices for a couple of things and having : > it dynamic is a good thing for its space savings. : : And you don't tweak the build at all for your embedded system? :) We do, but my point is that it increases the number of tweaks that I have to do to the system if we start sprinkling these things all through the tree. I'm worried about this being the cammel's nose and the rest of the camel is going to follow. : *shrug* If it were a current only thing that would be fine for me, as : that is where it would be useful. Then again, for a stable -> current : upgrade it might be needed in stable as well one could argue. Also, : the libc sources don't have to be screwed up for libc to be trashed. : Kernel panics at inopportune times due to other bugs can result in : pain. Yes. And if it was a current only thing, then it would do the right thing in upgrades... make builds a new make to be used for the rest fo the build. : > b) I've not seen the numbers for this. If it is only 1% faster, it : > doesn't make sense, even though it sounds good on paper. : : I would only argue for a static make on teh basis of recoverability from : bad juju, not for speed. OK. Of course when my libc was borked (and it was borked at least a dozen times), cc1 and as were what was core dumping, not make. Warner To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe cvs-all" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200102282053.f1SKrpd43256>