Date: Wed, 09 Aug 2000 21:18:43 -0600 From: Warner Losh <imp@village.org> To: Matt Dillon <dillon@earth.backplane.com> Cc: Gunnar Flygt <gunnar@pluto.sr.se>, Ruslan Ermilov <ru@sunbay.com>, freebsd-stable@FreeBSD.ORG Subject: Re: agp_if.c Message-ID: <200008100318.VAA53491@harmony.village.org> In-Reply-To: Your message of "Wed, 09 Aug 2000 19:52:35 PDT." <200008100252.TAA74174@earth.backplane.com> References: <200008100252.TAA74174@earth.backplane.com> <200008090743.AAA50631@earth.backplane.com> <134950593991.20000808104445@corbina.net> <20000808104924.E11454@sunbay.com> <20000808123728.D92082@sr.se> <200008092038.OAA50466@harmony.village.org>
next in thread | previous in thread | raw e-mail | index | archive | help
In message <200008100252.TAA74174@earth.backplane.com> Matt Dillon writes: : : :BTW, I couldn't find the phrase 'use the magic kernel building : :mechanism and it ought to work' in my UPDATING file. Such a phrase is : :patronizing and in appropriate for UDPATING. Can you tell me which : :version you found it in? : : : :Warner : : I think it's well deserved, actually. All the updating file says is : to essentially rebuild the world. That's a cop-out and a huge : annoyance if someone just wants to get a blasted kernel compiled up : and doesn't have 2 hours to waste waiting for a buildworld. Matt, you put quotes around it. Implying that it was in the file. It is your own damn fault that it took you 2 hours to find out. There's an entry on the binutils problem in the -current UPDATING. In addition, had you followed the detailed instructions, exactly, YOU wouldn't have wasted that time. Don't blame me for YOU not reading -stable and -current mailing lists. I'm sorry you got burned, but the instructions are accurate. : UPDATING needs to contain more detailed instructions. For example, if : all you need to do to get the new binutils is to compile up : /usr/src/gnu/usr.bin/binutils, then UPDATING should damn well SAY that. But that's not all you need to do. Sometimes you need to do other things. That's why we hid it behind the safe target. I don't have time to deal with all the stupid details, and our users don't want to know. They just want something to work. If you don't like it, that's your problem, not mine. Don't try to blame me that you got hurt when you went off the beaten path. : Same goes with other breakages. If UPDATING had simply stated that all : one needed to do was to 'cd /usr/src/share/mk; make install', rather then : build the fraggin world, it would have saved me over an hour of hunting : around trying to figure out why the kernel compile was breaking. But that isn't always the case. Sometimes it is hard to know what all the cases are. I'll be adding an entry in UPDATING about share/mk since it is turning into a FAQ. : In the last month I have wasted at least 8 hours working around kernel : breakages that two simple paragraphs in UPDATING could have prevented. One of the two paragraphs was there. You missed it. Not my fault. : You bet your ass I'm pissed off! Stop trying to minimize UPDATING and : put something *DETAILED* in it the next time someone commits something : that will break an existing tree! Well, there's the following paragraph: 20000522: A new set of binutils went into the tree today. Anybody building a kernel after this point is advised that they need to rebuild their binutils (or better yet do a buildworld/installworld) before building a new kernel. So the details are there if you wanted them. You didn't bother to read it. Not my fault. Don't blame me. The cookbook instructions are for people that want simple and consistant rather than hightly optimized and fast. Sure, I didn't back port this to -stable, but that was a conscious decision. The details are in -current, since that's supposed to be more detailed, dangerous place. -stable users don't want kludges, they want simple solutions. I'm sorry that you didn't bother to read them, or understand the implications of them. I didn't put anything in about the mk stuff for two reasons. 1) I didn't know about it until a few days ago and 2) I didn't think it was needed because of the binutils issue. I'll add an issue. So please, pull the thorn out of your paw. Warner 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?200008100318.VAA53491>