=== safaci2000 is now known as AlienFreak [01:51] anyone remember the name of that tool that would calculate how much free software was being used on your system? [01:51] * ianorlin doesn't [01:56] aha, vrms [02:00] oh I heard about that [02:01] :) [02:03] * ianorlin just installed it but it did not notice a proprietary braodcom driver [02:03] didn't notice chrome either [02:03] or hipchat, I think it only checks a few repos [02:29] ah [03:49] hm to be comprehensive it could check binaries in PATH and report 'known free', 'known nonfree', and 'unknown status' [04:20] yeah [04:28] might even have to include library paths, then would need awareness of the packaging directory layout of python/ruby/perl/etc [05:27] I am very happy because I found out why Django was incredibly broken on my shared hosting. It was because I had Django 0.96 hiding along with a PYTHONPATH variable. This has been remedied. [05:32] I am also happy because I just discovered why Django wasn't running, so I actually beat ASO support to a fix. This is remarkable because the support team are ninjas. [05:33] Of course, I'm getting Python import errors again, so it's back to throwing keyboards. [05:40] haha [05:41] progress! [05:54] And now it works. :D [06:00] woo [06:01] i realized the pythin script i wrote only did half of what i wanted/needed [06:01] and just learned about "enumerate" [06:06] arrith: just wait until you learn about list comprehensions. :) [06:06] It is tea time. [06:07] nhaines: oh somehow i already stumbled on that [06:07] that stuff is cool [06:08] not sure how to use it beyond the simple like [i for i in thing] or [num**2 for num in thing] [06:13] arrith: it's a start. :) [07:11] LOL. virtual Richard M. Stallman that's awesome. [07:12] ie. vrms [07:14] it doesn't actually agree with him, though :s [07:14] well.. stallman barely agrees with himself... [07:15] I think if it was accurate it'd probably mark 90% of my system as non-free software [07:22] eh, AlienFreak i'd challenge you to find a contradiction in what RMS has says [07:23] does get kind of tricky in some minutiae, say like selling proprietary licenses, or LGPL, but it's all pretty consistent [07:24] he's a unique character... I would simply argue that most people's definitions of free software doesn't match his [07:25] either ways.. it's too late and i'm too sleepy to argue RMS freedom rules [07:26] AlienFreak: haha, well when you're not, i'm all for checking my own consistency [07:27] but yeah, he does use different definitions for some words and tries to get others to use them. that might not be the best use of his time [07:27] * AlienFreak nods. [07:27] gnu/linux particularly [07:27] don't get me wrong.. I appreciate his prescence. He defines the extreme left view of the free/proprietary debate [07:28] I just think there's room to use F/OSS without going all the way to the extreme left. Last time I met him, he was arguing that java was evil. (unrelated to the memory usage pattern on the machine ) [07:28] in a room full of CS students whose curriculum is 70% java [07:30] AlienFreak: actually now it's gnu+linux [07:30] and no one knows that except for people that follow the latest on RMS heh [07:30] i thought it was GNU/systemd+linux [07:30] is it? [07:30] RMS is probably the most consistent person I've ever met. [07:30] AlienFreak: well he says "gnu plus linux, or gnu slash linux" [07:31] didn't he finish his kernel now. Does he call his os gnu + hurd? [07:31] rww: soon to be, thanks to those traitors in the debian high council :( [07:31] * rww rolls eyes [07:31] AlienFreak: he goes out in the jungle for years with only a mips machine and the emacs src, and makes progress on hurd [07:32] rww: debian was the chosen one! *rest of star wars quote* [07:32] and it chose well :3 [07:32] no one left to save us now [07:32] The transistion to systemd is a real shame, but hardly the end of the world. If they work well with others, anyway. [07:32] rww: pssh people had fedora, they had arch [07:32] leave me my debian. even pass that "no encouraging one init system" thing, that was decent [07:33] nhaines: that so hasn't happened yet. nor are there really hints that it will [07:33] arrith: you are aware that non-Ubuntu-created init daemons are *normally* used by more than one distro, yes? [07:34] arrith: and systemd is better in a technical sense than sysvinit, and I tend to find that the anti-systemd ranters are rather less personable than Lennart [07:34] rww: i'm not necessarily advocating for upstart [07:34] that gentoo one didn't seem that bad. personally i'm just for having something *besides* systemd, so people can take it all more piecemeal [07:34] Is your expectation that every other init system in the Debian archives will magically stop working now that systemd is going to be default? [07:34] have it as some spec of different tools that perform different functions. you get the bonus of the *bsds possibly being able to make use of it [07:35] rww: welll... bitrot is a thing [07:35] if they only test systemd [07:35] arrith: they didn't only test sysvinit when it was default. systemd worked quite happily on Debian back then too. [07:35] arrith: and if people care about the other init systems, they will presumably continue this tradition [07:36] if nobody cares enough to do the work, it's not really fair to expect it to be done *shrug* [07:37] rww: but that's the point of why defaults are so important. chances are if something is going to be written/maintained, it will be with various expectations. now the expectation is all the systemd junk will be there. [07:37] This is only really a problem for daemon package maintainers, though. [07:37] arrith: feel free to help with the efforts to provide comparable services to e.g. logind. I believe slangasek would be the person to chat with about that. [07:38] it's fine to require it, like gnome has been. but a slow, pernicious decay of packages depending on systemd in one way or another [07:38] rww: it takes way more than one person. almost the resources of... a distro [07:38] and if people care enough to do the work, they will presumably do it [07:38] people can boycott gnome, you can't boycott the debian repos on debian [07:39] rww: well i would have hoped people would care enough to keep sane defaults [07:39] shockingly, people disagree with you on what is a sane default [07:39] that is one thing i'm curious about, polls of debian user preferences. i remember some poll but i forget if they were debian members or devs [07:39] rww: everyone makes mistakes [07:39] learning from them is what's important [07:40] personally I think that sticking with a mess of bash scripts for initialization when there's a superior option out there that makes use of Linux features (just like FreeBSD's init daemon makes use of its, OS X ditto) is not really sane [07:40] Debian user preferences are worthless. [07:40] It's Debian developer preferences that matter. (They're the ones doing the work.) [07:40] systemd works nicely, the cgroup setup is nice, journalctl has some really useful features, and in general dismissing it all as "junk" and complaining that other people aren't doing the work you want them to do is a bit naff [07:41] rww: my main concerns are simply not being modularized and not being platform agnostic. i suppose both could be fixed in time, but people seem to want to undo the faults of systemd rather than try to do a cleaner implementation. [07:41] just seems like a less than ideal way to go about it all [07:42] arrith: "modularized" means what, specifically? And I'm perfectly fine with Linux being like every other sane OS and actually using the features its kernel provides to make sysadmins lives easier. [07:44] rww: actual separate projects. creating a spec then implementing it rather than effectively being another linux-kernel-level (in size and scope) "do everything". [07:45] rww: sure, i'm all for making things easier. but there is something to be said for the usefulness of efforts like kFreeBSD and how systemd really makes them impossible. [07:45] systemd dependence* [07:45] What's useful about kFreeBSD? [07:45] There's something to be said for not holding back a Linux distribution because a FreeBSD based offshoot of it that's used by like 10 people would have to make sure its alternate init system still works. [07:46] nhaines: lack of a monoculture, alternate features, alternate architecture, alternate performance characteristics, etc etc [07:46] rww: consider kFreeBSD 10% of my concern, to be clear [07:46] That sounds way too vague to be actionable. [07:46] but still notable [07:47] nhaines: the debian project thought it worthwhile enough to give it, i forget their term, but almost up to an official release. just marked 'beta' or something. [07:48] I'm not certain Linux can be described as a "monoculture". [07:48] nhaines: 0days on the kernel are a thing [07:48] The most ridiculous part of the whole systemd-in-debian thing for me is all the energy going into complaining about logind (which is generally the concern for some reason) depending on systemd and all the lack of energy going into writing an alternative to logind that implements its interface. [07:50] rww: well i have heard that various parts of systemd can be disabled, i'm really not sure what percentage of all of systemd is going to be enabled by default [07:50] so you end up with situations like Ubuntu using an old version of logind from pre-cgroup, and you have to wonder why it is that apparently everyone competent to write a seat manager is only working on systemd [07:51] arrith: logind requires systemd. systemd requires journald. that's about it in terms of hard requirements [07:51] if it's just logind, that's a lot easier to reimplement than all of systemd, but i doubt it will be [07:51] arrith: GNOME doesn't depend on any of the above, but functions better with logind (apparently, I don't use it so...) [07:51] so go re-implement what logind does with some other non-systemd method, and you're all set [07:52] just for the record, and i don't mean to equate them at all, but leonart pottering tends to be a bit heavy handed in his software planning. effectively the reverse of modularization and following a spec. i'll say i'm not intimately familiar with the specifics of systemd internals but i'm wary [07:52] rww: but logind steals your software independence. [07:52] rww: is that all debian has approved of in terms of systemd use? [07:52] since yeah, the compilation flags for systemd will actually dictate a lot of systemd's effect [07:53] arrith: as I understand it, Debian is going to use systemd as the default init daemon, logind on environments that use it, and I'd expect journald piped to rsyslog but I haven't seen much on that bit yet [07:53] hmm [07:53] plus systemd-udev, of course [07:54] if it's really just init, and they clean other parts up before they use them, or don't use them at all, i would be okay with that [07:54] hrm [07:54] what "other parts" are you referring to? [07:54] well that binary log format i'd love to avoid is one thing [07:54] no reason a json or something couldn't do all it can [07:55] the one that journald uses by default, which you can trivially switch off and have stuff go to rsyslog like usual [07:55] including on Fedora and Arch and what have you [07:55] rww: the defaults for that do matter, imo [07:55] yeah actually, i wonder what arch and fedora have for either. i'm guessing they leave it on. [07:55] and they're a matter of opinion, and Debian's processes disagree with yours, so... [07:55] fedora at least for sure [07:55] as far as I know they use the default journald setup [07:56] well, the council at least. might not necessarily reflect usrs [07:56] it's a tech committee, and Debian users are free to become developers so that their opinion actually matters. that's how things work in Debian. [07:56] rww: +1 [07:56] rww: well it would be quite something of say 80% of debian users were against systemd say [07:57] that would really not be serving the community [07:57] if 80% of Debian users were against systemd and there weren't enough Debian Developers against it to pass a general resolution (majority) against it, that would be quite something [07:57] The Debian community is the developers. [07:57] they're not there to serve the community, they're there to make a good distro [07:58] heh they can have a great distro, that nobody uses. i'm looking at nixos [07:58] okay, have fun with that. I don't expect 80% of Debian users will follow you. [07:58] rww: you haven't seen polls? [07:58] i swear there was one [07:59] a self-selected internet poll on a subject with a bunch of arguing is rather unlikely to be statistically significant [07:59] but, we'll see. If Debian drops 80% of its user count with jessie, I'll come back and tip my hat to you. [08:00] iirc was on some debian mailing list and was kind of official-ish [08:00] haha [08:00] hell, ditto if it drops 10% [08:00] well for all i know the poll was pro systemd [08:00] eh [08:00] and again, if systemd is such a huge problem, why is there no General Resolution overriding the tech committee? [08:00] well ofc that won't happen, you'll just see sysvinit numbers in popcon stay high [08:00] the tech-ctte result specifically allowed for a majority vote GR to override them, but nobody's even proposed one [08:01] rww: it might be that they just want to use 'the good parts' and either slowly add the rest, or ideally make them better before they add them [08:01] i thought ian tried? [08:02] http://ostatic.com/blog/debian-developers-get-user-input-on-systemd [08:02] no, he seconded a GR from someone else that kept systemd as the default and said that package maintainers need to make sure their packages work on all Debian-packaged init systems or some unworkable nonsense like that [08:02] http://people.debian.org/~stapelberg//2013/05/27/systemd-survey-results.html [08:02] Title: [Michael Stapelbergs Debian blog] [08:02] eh [08:02] that's unfortunate if that was the case. just keeping legacy support and being open to systemd alternatives would have been good for me. [08:02] so 15% plus 14% undecided [08:03] Top concerns [08:03] complexity is top, second is not portable to non-linux systems [08:03] arrith: they are open to systemd alternatives. upstart is in the archive. go file bugs and write patches for your packages that don't support it [08:03] ditto sysvinit [08:03] not ditto openrc, but that would be nice to see [08:04] arrith: I thought we were talking about people who didn't like systemd, not people who had concerns about it? [08:05] rww: i meant "open" there as in taking future investigations into if people are liking systemd, and if something else came out that would be more suitable. right now its "hope everyone likes systemd" [08:05] rww: there's a difference? [08:05] erm, yes? [08:05] i'm sure if systemd changed drastically i would love it [08:05] or are you of the opinion that things you like can have no issues? [08:05] so i'd say my dislike is all concerns [08:05] heh, i prefer them to have as few as possible [08:06] and I expect that Debian Developers tend to have fewer concerns about systemd than about sysvinit [08:07] i would be curious what, if anything, is brought up about systemd during the next board election [08:07] since that might be, or not [08:08] If you're referring to a tech-ctte election, they're not elected. If you're referring to Debian Project Leader, go ask about it on debian-vote@, since the nomination period started already. [08:09] (per Debian Constitution 6.2, new tech-ctte members are generally recommended by existing ones to the DPL who then decides whether or not to approve them, which is similar to all the other stuff DPL delegates in Debian) [08:10] anyways, it's bedtime for me, I have to be up for work in 6h [08:10] rww: yay work! [08:17] had to put away dishes [08:17] aww [08:18] that is interesting, i really should look into that. i heard of the DPL elections, and i guess i assumed the CTTE were also [08:19] one thing i thought of is overall linus and his lieutenants have done a fine job stewarding linux, and a (if not the most) critical part of linux distros [08:19] the systemd team, and systemd, is all very new. it hasn't exactly proven itself anywhere besides just being the only thing that supports various kernel features [08:20] then you have redhat taking gnome in some odd directions, then they promote systemd as this "let systemd solve everything for you" [08:21] as some somewhat-ignorant user it feels forced and is quite off-putting. [08:21] heh coincidentally brought into debian by an unelected board :P [08:22] i do wonder how much of the existing board a new DPL could dissolve, if the DPL really wanted to [08:26] https://www.debian.org/devel/constitution [08:26] Title: [Debian Constitution] [08:27] might have to drop some DoI [08:28] When in the Course of human events, it becomes necessary for one people to dissolve the Debian Constitution [08:28] Debian Constitution 6.2.5, I'm going to say no. [08:29] wow [08:29] that is some supreme court stuff [08:30] 4.1.4 about overriding [08:31] Yes, that is because the developers are the ones doing the actual work. [08:31] well i mean more, unelected [08:32] devs, those doing the actual work, might want to be able to participate in the selection of the technical committee [08:34] They do in that they elect the Debian Project Leader. [08:38] hm yeah that is interesting though, i wonder why they're not directly elected. there must be some reasoning behind that. [08:40] Sounds like a huge, unnecessary time sink. [08:41] well could do some kind of hybrid system. unelected except say once a year. refreshing period. [08:41] also, only unnecessary until it isn't [08:42] the systemd fiasco clearly shows they've gone mad with power and are unfit to lead [08:43] If anyone really believed that, they could hold a General Resolution. [08:44] who is silencing them is the question! [08:44] how deep does this go [08:47] The only thing I care about is that Ubuntu doesn't break on phones and tablets. [08:56] oh i really don't think it will. ubuntu phone is the top priority atm [08:56] besides i guess ubuntu server [08:58] i actually wonder what distros are actively rejecting systemd, and might be sources for usable alternatives at one point [08:59] possibly gentoo? i know they like that openrc a lot, but i wonder if they're refusing to 'package' systemd [08:59] i doubt they're refusing to 'package' gnome3 [09:01] Ubuntu's main priority is cloud infrastructure. :) [09:01] openstack or that landscape? [09:02] Openstack and Juju. [09:13] ohh [09:14] juju is really going to have to pull off some magick to take on this docker craze [09:14] juju was doing pretyy well, but was hardly known outside of people directly inside ubuntu or following it very closely, then docker exploded [16:43] almost... there... [18:58] arrith: both GNOME 3 and systemd work fine on Gentoo. I've personally used systemd on it, have not tried GNOME 3 on it.