Mark Knecht posted on Mon, 24 Apr 2017 08:36:43 -0700 as excerpted: > I certainly could chroot a specific copy of Gentoo and build on my > machine. I might also be able to build binary packages on my fast > machine and then do an emerge -k type install and see if it works. > > However, in the end how much do I gain for all that work vs installing > Kubuntu?
There's some advantage in learning one distro, learning it well, and using it on everything. That's what you gain, assuming you're keeping everything else on gentoo, as you then don't need to keep track of the many distro differences. I learned the difficulty of dealing with multiple distros here with my current router, still an old Linksys wrt54gl (which as I said I intend to eventually upgrade to an amd64, so I can build for it at the same time as the rest of my systems, can configure it using the same methods and tools, etc.), running openwrt. I had a horrible time trying to configure its networking system the way I wanted to, basically having to read a bunch of its init system scripts and config to figure out what started what, in what order, what and how to modify that to my liking, etc, pretty much just to figure out what config file to edit to change a few settings I wanted to change. Even then, I felt like I wasn't getting the most out of it, because in ordered to do that I'd have had to read and understand pretty much the entire init system. So mainly I just stuck with the defaults instead of really getting it to work how I wanted, and I never did really /truly/ understand it. Now that version is now long outdated, but I don't want to update or indeed, to really change the config as I set it up back then, because in ordered to do so I'm going to have to dive back into things and figure all that stuff out again. But I'll only be using it on that one thing; the info and skills gained won't really transfer to anything else, unless I decide to standardize on openwrt for everything, including my main machines! By contrast, if it was gentoo, I would have already known the basics and could have gotten right to the task at hand. And I could have and likely would have done far more with it, because I really do understand the openrc setup (this was before systemd went mainstream). These days of course most distros are standardized on systemd for init, so learn it once, use it on all. And that's one of the reasons why I eventually switched to systemd on gentoo. Except, particularly for that old thing with its extremely limited system image and RAM sizes, I don't think systemd would fit. Which is probably a good share of the reason that last I heard anyway, openwrt wasn't switching to systemd. Between my dissatisfaction with not being able to truly master the openwrt system in the time I was willing to devote to it as a one-off, and my dissatisfaction with having to build separately for my netbook, even if it was gentoo, I resolved, as I explained, that next time I upgraded things, I'd standardize on amd64 (Intel or AMD chips either one), and try to keep things similar enough that at least for most packages, I could use the same C(XX)FLAGS and USE flags for everything, and just do binpkg- only emerges on systems other than my primary, for most packages. That way both the packages and the setup would be the same across everything, except where I had actual reason to make it different. And I'd really understand both that setup, and how to change it to accomplish what I wanted to do, if necessary. Now I'm into customizing enough that I've never met a desktop that I liked as it was shipped, and I expect I never will. And at least as I envision things, even if I'm 80 (30 years from now as I just turned 50 this year) and in a nursing home, if I'm still of sound enough mind and body to be running computers, now that I know the level to which I can efficiently customize gentoo, I really can't see myself being happy within the limitations of a normal binary distro an longer. It's not as emphatic a "won't ever happen" as the idea of me switching back to something proprietary like MS Windows or Apple OSX, but for me it would certainly feel like going in the same direction, and would thus feel like defeat. At that point, if I can't any longer do gentoo or at least arch, I may well simply turn in the keyboard and mouse, and if I do that, I can't imagine I'd have much else to do to keep me happy, so realistically, I might well wither and die within a few months, figuring I have little to nothing remaining to live for. Now I'm /not/ saying the answer has to be the same for you. Far from it! In fact, the above sounds like you may be tilting the other way, toward making everything (k)ubuntu, and giving up on gentoo. If you're satisfied with (k)ubuntu, standardizing on it would equally as effectively solve the problem of having to deal with two different distros with wildly different ways of doing things. And that may work very well for you. But it definitely wouldn't work for me. I couldn't be happy on (k)ubuntu, or fedora, or... I left those limitations behind in 2004 when I left mandrake for gentoo, much as I left the limitations of proprietaryware behind in 2001, when I left MS as eXPrivacy crossed a line I couldn't and wouldn't cross, for the land of Linux freedomware, where I'd not be /asked/ or /expected/ to cross such a line in the first place. Of course doing a split across multiple distros is possible too, but it does have its negatives, which I'm trying to point out here, and for me anyway, those negatives were high enough that while I lived with them while I had to, I resolved that when I got new hardware, I wouldn't have to any longer. But of course perhaps that too you'll find less of a problem than I did. I just don't like being jack of all distros and master of none, is all, and would prefer to master one distro, ideally a really flexible one like gentoo, knowing it well enough to comfortably make it do what I want, and use it everywhere. -- Duncan - List replies preferred. No HTML msgs. "Every nonfree program has a lord, a master -- and if you use the program, he is your master." Richard Stallman