belkinsajrgifford, do you want to test chatb.org/#ubuntu-us-oh tomorrow?01:42
jrgiffordbelkinsa: sure01:42
belkinsaThanks.  Anyone else here is more then welcome to join us also.01:43
Unit193Creeping on jrgifford, sounds fun.01:47
gilbertpaultag: so discission on init-select is not going so well, and i'm not entirely sure why18:06
paultagI have no idea18:07
gilberti don't really understand obstructing anything so simple, non-obtrusive, and solves so many problems18:08
gilbertthere is an undercurrent of deferrence to the TC that i also do not understand18:09
paultagI think it's because everyone is tired of this18:09
paultagand any decision is better than continuing with all of it18:09
paultagbecause it's blocking positive work, and it's holding us all back pretty hard18:10
paultagI'm a strong systemd hardliner, but if upstart was made default, I'd stop fighting and just use it18:10
paultagI'd rather upstart than sysvinit18:10
gilbertthat's the anti-freedom argument that keeps happening.  why suport everyone working on the thing they like?18:10
gilbert*why not*(18:10
paultagbecause then we have to write (and maintain) all sorts of init things in every daemon package18:11
paultagand it means we can't be sure of feature x18:11
paultage.g. we'd make spec files and unit files and initd files and …18:11
gilberti'm sure feature x checks can be written, and when so choose another supported init18:11
gilbertif there are people interested in doing all those things, why get in the way of their interests?18:12
paultagperhaps, but once we switch to {upstart,systemd}, the other becomes hard to support, since we'd not be forced to ship init scripts18:12
paultagI wouldn't18:12
paultagbut there's no one who's interest it is to patch every dameon18:12
paultagplus no one will maintain logind18:13
paultagoutside systemd as pid 118:13
paultagwhich means a lot of things get broken18:13
paultagso the upstart or openrc or whatever folks would need to join the systemd team to maintain logind18:13
paultagwhich is a lot of work and code hacking18:13
paultag(or GNOME breaks)18:13
gilbertor gnome automagicially init-select's systemd18:14
paultagthen why support something else? :)18:14
gilbertit could basically declare that it was incompatible with the other inits if the work didn't get done18:14
gilbertbecause people are interested in those other things, and freedom, i think, is the ultimate ideal18:15
* paultag shrugs18:15
paultagwe'll see where it lands18:15
paultagif it's upstart, systemd gets torn apart and used as parts18:15
paultagif it's systemd, all other initds get torn up18:15
gilbertor different parts of the system choose the init they prefer: xfce and kfreebsd lands get upstart, gnome and possible kde get systemd18:17
gilbertits about choice and freedom18:18
gilbertand i can't understand it any other way, but maybe i'm too much of an optimist18:19
paultagI think it's bascially breaks down to those who think Debian's about freedom, and others who think it's about making a stable OS with predictable configs to ensure stability18:22
paultage.g. some heisenbug when you swap out your libc and also init to foo when running software x results in a segfault or something18:22
gilbertthat's not really a heisenbug as long as reportbug reports your non-default init config18:23
paultagyeah, but which was it? :)18:24
paultagthe eglibc, the init, or software x? :)18:24
paultagerm libc*18:24
paultagboth views are valid18:24
paultagand that's likely where some of that comes from18:24
gilbertthat's when the maintainer says, hey, try swapping these things to the defaults and let me know if the problem still exists18:24
paultagyou see how it'd not be first-class18:25
gilbertyes, but there is tons of second-class stuff already in debian that gets worked on, figured out, and fixed18:25
gilbertlike all of the non-popular architectures18:25
paultagthat's different than a release arch18:26
paultagif we get a bug on mipsel, we'll still have to deal with it or it's rc-buggy18:26
gilbertno, i mean everything except i386/amd6418:26
gilbertbut those really are second-class archs even though they are release archs18:26
gilbertthey are not the thing that the majority of users use18:26
paultagI can see it both ways18:27
gilberti know...18:27
thafreakdid I wander into #systemd-vs-upstart by accident? ;)18:33
paultagnot really18:33
gilbertnot exactly18:33
paultagsince gilbert and I are in alignment on philosiphy18:33
paultagbut it's related18:33
gilbertyou wandered into gilbert wants peaceful coexistence of all inits18:33
paultagwhich is great18:34
gilbertbut perhaps overly idealistic18:34
paultagI like upstart's maintainers more, but I do like systemd more technically18:34
thafreaki don't like learning new things :)18:35
thafreaki'mma fork sysv-init and keep rocking it forever18:35
paultagdo it!18:36
paultagopenrc looks like a step up btw thafreak18:36
paultagit wraps sysvinit18:36
gilbertyea, its already called openrc18:36
gilbertand debian may one day (fingers crossed) get packages for that18:36
paultagzigo really did great work18:36
paultaggilbert: it's in experimental18:36
paultagas of an hour or two ago18:36
gilbertpaultag: oh, awesome!18:37
paultagNEW is a well oiled machine, gilbert18:37
gilbertnice :)18:37
paultagnow that I have a finger again :>18:37
gilbertthat's always useful ;)18:37
Unit193paultag: Yeah, saw that you processed that and really got it down.18:39
paultagand another new member is keeping it low18:40
paultagI can do large scale processes at once, and he's good at incremental processing18:40
paultagso it's a really nice team :>18:40
paultagso I did ~200 over vacation18:40
paultag    paultag: {ACCEPT: 271, PROD: 11, REJECT: 24}18:40
paultag    paultag: {ACCEPT: 1170, PROD: 140, REJECT: 136}18:40
Unit193Need to reject more. ;)18:41
paultagyeah, only ~10%18:41
belkinsajrgifford: ping20:05
belkinsaWoah, hey there, BobJonkman.23:40

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!