Alex Smith wrote: > On Thu, 2006-11-02 at 18:24 +0100, Gabriel C wrote: > >> Brrrrrrr. >> 'with the sysvinit emulation shit based on runlevels* you don't need it >> , *I said if you want a full based even one*. >> This emulation fu** is *what* ununtu is running LOL!. Then again why to >> hell use something >> , *uses* the INIT|INIT STYLE WE DON'T WANT! brr brr brr >> > > 17:21 < Keybuk> right, we just replaced the stuff that was in inittab > directly > 17:21 < Keybuk> it doesn't make sense to *not* be backwards compatible > with /etc/rc*.d > 17:21 < Keybuk> we need to support that for most software to work > without modification > 17:21 < Keybuk> otherwise on day #1, we need to write upstart jobs for > everything > > 17:22 < Keybuk> by supporting it, we can write upstart jobs one at a > time (until they are all done) >
This sounds like a rushed 'we need get it in for some release' sorry. But anyway they can do whatever they *want* . We *Frugalware* want replace *sysvinit* and not add some whatever may be good or bad 'sysvinit EMULATION|COMPAT' back. Does not make sense *for* Frugalware ( It may for Ubuntu ). >>> >>> >>>> Pls don't tell me now *but it can be done using sysvinit* I know this >>>> but the point is *we want to change sysvinit* and not replace by >>>> something add sysvinit back ... >>>> >>>> >>> It's not a sysvinit hack... There is *NO* sysvinit code in there. Ok, it >>> was started from sysvinit, but gradually it was replaced with better >>> code. >>> >> Whatever better code or not is a sysvinit *fork*. >> > > So uhh, InitNG is a fork of sysvinit? Minit is a fork of sysvinit? Runit > is a fork of sysvinit? <insert init system here> is a fork of sysvinit? > I don't think so... > > init-ng is based ( well the idea ) on minit. Runit is based|rework on|from daemontools , and upstart on sysvinit. Everything starts with whatever code_from_foo is first a fork.Later you can call it based_on_foo if this sounds better for you. >>> Actually reading about upstart in detail really helps. This >>> LugRadio episode might help - http://www.lugradio.org/episodes/61 - go >>> to 57 minutes into it, the Upstart developer is talking there. >>> >>> >> .... >> > > ... What? > Na what for ? I don't need and want to waste my time with this. >>>> And again just do it , I don't want flame about I just think a 'init >>>> system' should be small , fast , ( not having weird depends ) , easy to >>>> understand ( not only for devels , think on server admins as example ) , >>>> of course is my opinion. >>>> >>>> >>> No weird depends, depends on glibc, that's all. >>> >> In 'emulation modus' yes. >> > > There is nothing, I repeat, nothing that needs DBUS or Hal. The Upstart > devel dislikes dbus anyway > Ok then pls forget this emul | compat modus and get everything working with jobs|events. Does everything works yet ? > >> >>> It's fast - at least, >>> performs better than sysvinit here. >>> >> I don't really care whatever foo_init boots 2 or 3 secs faster or >> whatever not. >> At last on ubuntu is *slow* as hell. >> > > That's because they're using the sysvinit compat stuff and have lots of > uneeded stuff at boot > > Which everyone will need to use. Well I got 3 daemons more on boot on Ubuntu compared with sysvinit on Frugalware. Maybe is because Ubuntu is i386 but ... Ohh well slow or not this is not the problem ( for me ). >> >>> Server admins can understand it >>> easily >>> >> We will see , I don't think so. >> > > Meh. Anyway, what does everyone else think? > Na wait :) (here some up_cookie for you so :P) all the other need RTFM|WIKI|ML first we already did :-DD > Thanks, > Alex > > crazy _______________________________________________ Frugalware-devel mailing list [email protected] http://frugalware.org/mailman/listinfo/frugalware-devel
