home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.os.linux
- Path: sparky!uunet!ukma!darwin.sura.net!zaphod.mps.ohio-state.edu!cs.utexas.edu!mercury.unt.edu!sol.acs.unt.edu!kenc
- From: kenc@sol.acs.unt.edu (Ken Corey - Operator)
- Subject: Re: Whining...I don't want to hack on the kernel
- Message-ID: <kenc.714229866@sol.acs.unt.edu>
- Sender: usenet@mercury.unt.edu (UNT USENet Adminstrator)
- Organization: University of North Texas
- References: <1992Aug18.011305.27223@access.usask.ca> <1992Aug18.212007.28991@unislc.uucp> <1992Aug19.063722.9778@nwnexus.WA.COM>
- Date: Wed, 19 Aug 1992 13:11:06 GMT
- Lines: 85
-
- vince@halcyon.com (Vince Skahan) writes:
-
- >[...semi-soapbox on...]
- >There *ARE* people who want to investigate Linux as an out-of-the-box
- >unix system that they can do stuff with, and who are quite willing
- >(and grateful) to have somebody get their jollies by compiling all the
- >sources and making a kit that they can use (to get their jollies by getting
- >their application(s) running, serving as real-world users and testers, etc.).
- >I guess it bugs me that the folks who want a stable o/s and compiler
- >that they can use to build (or port) software get pounded on in c.o.l.
- >as being 'beneath' the folks who are building functionality into the
- >kernel. Both are equally valid users of the o/s.
-
- This is *very* true! There *are* equally valid, yet different users of
- Linux. However, if you'll take a look at most of the pounding messages in a
- slightly different light, most of the pounding came when people wanted
- conflicting things:
-
- 1) Stability for use in non-hacker projects.
- 2) The absolute most latest whiz-bang kernel/GCC
- 3) The once a week update to program xxxx.
-
- It has been said many times that if you want stability, then find a version
- that does what you need it to do, and stick with it for a little while, and
- don't try to "keep up with the joneses" as far as versions go.
-
- Yes, I'm following my own medicine. I'm still using GCC 2.2.2b, and Linux
- .96pl4 (I think, been a while since I looked) because of the fragility of
- the ST01/02 support code right now. I don't feel like hacking, and this
- kernel does what I need. Great. End of statement. I won't upgrade for a
- while. Why moan about not being "current"? Because .a files won't easily
- compile? Heck, they're not any more reliable than regular source code.
- (see my next paragraph.)
-
- >The guys who have built the various canned installable sets of goodies
- >such as mcc-interim, SLS, the 'rootimage' disk, etc. have done a great
- >job. I'm willing to use the fruits of their labor (as a very grateful
- >user) and spend my limited time trying to port the stuff I want to
- >run, or to put all the pieces together as a system that does something
- >I want it to do.
-
- Yeah, this is a nice idea, but out of all the .a files I tried to download
- (probably 20 or so), I only got one to compile and link properly. The rest
- didn't work, and I ended up getting the sources and recompiling from scratch
- myself. I *know* things work that way.
-
- >Building kernels is not necessarily great fun. Once you do a few, it
- >gets kind'a boring doesn't it? I think making patched compiled kernels
- >available to stick over the latest mcc-interim (or whatever) is a great
- >idea.
- >I don't *WANT* to be the local expert in all the obscure non-standard
- >gcc flags, just so I can build a darn program. I want plug'n'play tools
- >that I can assemble (ie...mcc-interim+new_kernel+mailpak+perl) to get
- [...]
- >but I'm also very willing to use mailpak, etc. and say "thanks!!!".
-
- Of course it's not fun. However, with all the changes floating around,
- there's no way to guarantee compatibility for all the old library calls and
- whatever, SO, when a kernel is changed, or a GCC and its library is
- changed, it's gonna break things in other ways with stock releases like MCC.
- Like it or not, if you're gonna use Linux, you're almost going to HAVE to do
- a little hacking to get your setup just the way you want.
-
- Also, though kernel hacking is no fun, it's much more fun than setting up
- the "current" version of Linux, and recompiling ALL the utilities desired,
- and then uploading that, and then finally makeing it publicly available.
- I know that for now, I've other things to waste time on...;)
-
- >The fact that those folks did a great job doing their parts in no
- >way means that I'm less of a valid user for saying "thanks a million"
- [...]
- >man pages, etc. They're just as important...
- >[...semi-soapbox off...]
-
- Yeah, this is true. Everyone is important. The only problem is: who is
- going to provide the labor to make everything stable between versions
- (especially when those versions are coming out every week)?
-
- I *still* say, Stay a few releases behind the "cutting edge", and compile
- from sources, not .a files, and you'll be much happier.
- --
- Ken == kenc@sol.acs.unt.edu == kenc@vaxb.acs.unt.edu
- Hurewitz's Memory Principle:
- The chance of forgetting something is directly proportional
- to ..... to ........ uh ..............
-