On Tuesday 02 July 2002 21:04, Jon LaBadie wrote: >On Wed, Jul 03, 2002 at 09:43:08AM +1000, Robert Kearey wrote: >> Gene Heskett wrote: >> >The majority opinion of those on this list is that any rpm's >> > for amanda probably should be subjected to an intentional rm >> > -f, they >> >> ... >> Incorrect. Where did you get that idea? > > ... > >> A little research does wonders, you'll find. >> >> >Generally speaking, making an rpm installation work is going to >> > be 10x more difficult than making a home built install work. >> > There >> >> I'd dispute that most strongly. It's all there in the spec file, >> if you'd care to look. > Robert: It's in the spec file you say. Ok, I might buy that, but where in all that rpm related stuff is a pointer that says for the new bee to read the spec file. I've been under the impression it was strictly for rpms instructional useage when building it all this time. > ... > >Sounds like you took it personally.
Which most certainly was not my intention. >Gene was commenting from a point having viewed many bad packages. >Perhaps yours is wonderous. That has not been the general > experience of long time contributers to this list. Amen, with some reservations, to wit: I have begun to appreciate that where system related stuff is concerned, the rpm path is a very good one. It tends to keep dependency wolves at bay. OTOH, if the system stuff is solid, then I'd druther build accessory stuff like amanda, cups, gimp-print, sane, even gimp from scratch because the configure scripts, generally speaking, are very good at sorting out the diffs between RHat and everybody else and giving one good usable code in all situations, something the rpm's can't begin to discuss because they are so "system" specific. Amanda is intended to run on many different systems and cpu's. No one can build a single rpm, binary or source that can do that. I've tried many times to build from a source rpm, and have in 3 years, succeeded maybe 3 or 4 times. Usually my compiler is the wrong version, or my rpm is the wrong version to even unpack the thing, or any one of 20 other reasons for the compatibility to go out the window. I was rather surprised on building the 20020610 snapshot of 2.4.3b3. I'd installed gcc 3.1, and most of the code you can download today will cause it to puke all over itself, requireing editing of makefiles to convert them to use gcc296. I didn't realise I'd used 3.1 on it till after it was built, installed, and running. Like the tv commercial says, These guys are good! Rant mode on again if the old timers can tolerate me one more time. I've made a living out of chaseing electrons for about 53 years now. I grew up with tubes, then transistors, then integrated circuits that gradually got smarter until they can do what they can do today. My own personal computer expertise if one would call it that, goes back to being pretty intimate with several of the old 8 and 16 bit cpu's and assembly or pure K&R C to program them. That includes walking around in their os's fixing bugs, or writing programs that were in constant use for 10+ years. Unforch, that was a decade or more ago and time marches on, in this case at a faster pace than I can now manage at 67 and counting. Linuxwise, I started with RH5.1. I guess what I'm trying to say is that I've lived long enough to get a feel for what works, and what doesn't. For amanda, rpms don't very often, and generate too much "help me" traffic, all out of proportion to the numbers of rpms actually being used. Its easier to instruct on how to build it right, plus the end user then learns far more about how it works, and is far better equipt mentally to handle things when they go bump in the night, or when he adds another machine or drive to his system. Which eventually they will unless we can finally cuff & stuff that Mr. Murphy that wrote all those laws... BTW, my way isn't the only way to do it, far from it. Consistency in how you do it should be your way as long as it works. Thats why I emphasize useing a script to configure the girl. -- Cheers to all, Gene AMD K6-III@500mhz 320M Athlon1600XP@1400mhz 512M 99.04% setiathome rank, not too shabby for a WV hillbilly
