KernelTrap has a new story showing how to upgrade your GNU/Linux OS from the 2.4 stable kernel to the latest 2.6.0-test4 development kernel. Included among the the eight detailed steps are a number of screen shots.
HowTo Upgrade To The 2.6 Kernel
About The Author
Eugenia Loli
Ex-programmer, ex-editor in chief at OSNews.com, now a visual artist/filmmaker.
Follow me on Twitter @EugeniaLoli
Are there anyone person(s) out there who have given Linux 2.6.0-test4 a try on their Mac yet? I’m looking at downloading/purchasing Yellowdog Linux out of curiousity to see what the performance is like using Linux on PPC.
what I really want to see is a Gnu/BSD system. A BSD kernel but with GNU utils (such as gmake and not make) and maybe glibc and lot libc , if possible.
I would also like to see a gnu/hurd gentoo system.
Ive never actually used it, but dont debian do a version with the debian userland ontop of a BSD kernel?
I cant vouch for its readyness etc, but afaik it exists….
what would be the advantages of a BSD kernel over a Linux kernel?
NetBSD is ported to many more platforms than Linux, but most people use x86 or PowerPC anyway.
> what I really want to see is a Gnu/BSD system. A BSD kernel but with
> GNU utils (such as gmake and not make) and maybe glibc and lot libc
> if possible.
GLIBC? Wouldn’t you prefer libc? After all, GLIBC is one very slow bloated piece of code in comparison and one that re-uses a lot of code across platforms (not very portable). I can understand your reasons for wanting gmake, bash and whatever else (which you can already have on BSD systems) but your wanting of GLIBC surprises me. What advantages does GLIBC have over libc for you?
> I would also like to see a gnu/hurd gentoo system.
GNU/Hurd is still very immature. A port could be done but it couldn’t be recommended for production.
CooCooCaChoo wrote:
curiousity to see what the performance is like using Linux on PPC[i]
instead of purchasing a linux distro, you can always use debian ppc, it rocks on stability & speed (much faster than macosx). You can always use the macosx interface via <a href=”http://www.maconlinux.org/“>MacOnLinuxLinux 2.6 PPC is my main desktop system.
[i]Nadav wrote:
A BSD kernel but with GNU utils (such as gmake and not make) and maybe glibc and lot libc , if possible.
you are not alone, BSD sucks on one important thing: userspace.
There is were debian BSD excels, a BSD kernel with GNU userspace.
Matthew Baulch wrote:
GLIBC is one very slow bloated piece of code in comparison and one that re-uses a lot of code across platforms (not very portable)
haha!! FUD! haha!! Slow? i use it on a very old sparc32 and is fast and reliable. You probably doesn’t know what means memory management, shared objects and load on demand on modern operation systems. Big is not the same as inefficient and slow anymore, just put your mouth where the facts are.
[i]What advantages does GLIBC have over libc for you?</>
In fact the diference between glibc and ‘others’ libc is that glibc support modern userspace requirements: i18n, unicode, portability, posix threads, etc., etc. I don’t know other libc that fullfills these requirements, evens slowlaris libc, not to mention BSDs. Ulrich Drepper is a great coder and he maintains glibc. His work for the community can not be destroyed with your FUD.
In fact the diference between glibc and ‘others’ libc is that glibc support modern userspace requirements: i18n, unicode, portability, posix threads, etc., etc. I don’t know other libc that fullfills these requirements, evens slowlaris libc, not to mention BSDs. Ulrich Drepper is a great coder and he maintains glibc. His work for the community can not be destroyed with your FUD.
FreeBSD libc has all those features. POSIX threading it provided by libc_r, however, the kse project is going to bring libpthreads. I am sure there are lots of other standards that both support, untimately, let the user decide and stop trying to ram things down peoples throats.
RE: Nadav (IP: —.red.bezeqint.net)
Ick. If you *really* want it, Debian is working on it. The (misguided) idea was to port the GNU libs and userland to the FreeBSD kernel. It was proving so problematic that they forked that project into a second competing one using the native FreeBSD libc instead of GNU’s.
Personally I think that the whole thing is a supreme waste of time, as all of the GNU utils (AFAIK) have been ported (natively) to all three of the open source BSDs. Furthermore, the BSD userland and kernels tend to be kept in sync, as complete, tightly integrated, working systems, and welding bloated GNU tools on top is a pointless endeavour.
GNU programs are bloated with “proprietary” (to GNU, not caring much for POSIX standards (as GNU’s *Not* Unix after all)) extensions, and are not well integrated with the various BSD kernels. The whole idea strikes me as potentially the biggest possible waste of time for the OSS and FS communities.
If you want GNU userland on BSD, goto Debian. Those folks have long baffled me with their crazy ideas, and it seems like they’ve got a crazy idea for every one on the planet from what I’ve seen. If you want “Unix done right”, stick with BSD, whatever the flavor. They all have their specialties (and I myself am partial to FreeBSD) but they’re all pretty similar.
As far as you wanting glibc, I would recomend against that outright. It’s far too large, far too slow, it’s math functions suck compared to one found in the *old* BSD libc, and is absolutely blown away by the new BSD math library derived from Sun’s. The BSD libc is smaller, faster, more POSIX compliant, and generally more well thought out and tested. It has also been ported to far more architectures than has glibc. The layout of the source tree is also much neater, which helps newbies read through it.
Yeah, I might come off as a BSD zealot here, but I have used both GNU and BSD for quite a while now, and and constantly comparing and contrasting both. I test things, and I read the sources, and talk at length with others regarding their real world experiences with both systems. BSD generally wins over GNU in the areas of maturity, stability, security, portability, documentation (GNU documentation just plain sucks), usability and elegance. There are good GNU programs to be sure, but as far as basic Unix utils go, BSD provides the best.