=== asac_ is now known as asac [01:44] jdong: I just tested installing into an Etch chroot from etch-backports and it pulled dependencies from backports like you would want. [01:44] So I think replicating the backports.org approach has potential. === ssweeny is now known as rmonroe === rmonroe is now known as ssweeny [01:46] what is that approach? [01:49] coppro: Pin the backports repo to a lower priority than the main one http://backports.org/dokuwiki/doku.php?id=instructions [01:50] but then you manually have to do that [01:50] for every dependency [01:52] coppro: That's the thing is you don't. [01:52] If it can only satisfy the dependency from the lower priority repo, it'll pull it from there. [01:54] oh, I see [01:54] do you need mutt for that? [01:56] No. That's just an example. [01:57] oh [01:57] so pinning the backports repo definitely seems good then [01:59] Seems to me. [02:13] ScottK: yeah I gave it some thought over the night and think that pinning is probably our best bet [02:13] We mostly need good U/I for it. [02:13] forgive the 2AM grouchy jdong who wanted a more hairy solution that got everything right :) [02:14] maybe at 2AM I should start writing dbus and hal [02:14] oops did I say that aloud? [02:14] yikes, scary netsplit [02:14] Speaking of which, can you fix it so the print screen key fires up Ksnapshot again? That's be really cool. [02:14] jdong: ^^ [02:15] ScottK: I'll look into that when I have some time and dig out a keyboard with a printscreen key :) [02:15] wouldn't be surprised if it got mangled by the big xmodmap file of DOOOOOM [02:15] Dunno. It worked in KDE3, but not KDE4 === cody-somerville_ is now known as cody-somerville === woody86_ is now known as woody86 [02:35] ScottK, jdong, dont use xmodmap to fix the problem... the equivalent to gnome-settings-daemon should just make sure it's listening for the (correct) standard keycode [02:39] * wgrant would like to know what that equivalent is. [02:42] superm1: that's my point :) [02:42] superm1: the point is there's no equivalent AFAICT [02:42] QT4 looks *HORRIBLY* broken [02:42] first off, listening to all key events doesn't even capture the brightness keys XFMonBrightness{Up,Down} [02:42] so kubuntu xmodmaps them into XF86Launch4 and XF86Launch5 [02:43] which are, oddly enough, QT::Key_Launch6 and QT::KEy_Launch7 [02:43] (??) [02:43] Ha ha ha. [02:43] KDE loses. [02:43] and you can guess where this road of hackjobs leads :) [02:43] I'm glad I don't have to care about KDE. g-s-d is bad enough. [02:43] I'm gonna dig into qt4-x11 later this week{end} [02:43] gonna see why on earth this is the case [02:44] fun trivia fact: Converting the hex keysym for XFBrightnessUp using QT4 to a string calls it "Meta+(unicode mystery block)" [02:44] :) [02:45] Oh good. [02:45] and registering that KeyEvent is as good as registering text message alerts for the Dow Jones going up these days :) [02:46] but it's gotten me intrigued enough that I've gotta fix it :) [03:02] nhandler: any luck? [03:02] pangloss: I'll try now. I've been pretty busy today [03:03] nhandler: no rush if you are busy, I have alot of work to do as well so I wont be doing any ubuntu stuff tonight === boomer is now known as Guest20284 [03:11] Tip #5999 of bug reporting: Try to collect your thoughts for at least 5 minutes before adding a comment. This isn't twitter. [03:12] watch, now I jinxed it and there'll be stream-of-consciousness bug reporting in edge.launchpad when I wake up tomorrow. Maybe with bzr somehow integrated. [03:12] Don't you mean identi.ca jdong ;) [03:13] LP needs a "You already opened your mouth within the past hour. Wait before posting again" [03:15] nhandler: No. We're talking LP, not Free software. [03:15] Actually never mind, identi.ca isn't FOSS either. [03:16] ScottK: I thought identi.ca was meant to be FOSS [03:16] * nhandler goes to re-read the info on their site [03:17] nhandler: Depends on if you think AGPL is a Free license. I really don't think it is. [03:17] I think the term is Less Unfree (aka AGPL) [03:17] ack ScottK beat me to the punch. [03:17] jdong: Less Unfree is a good way of putting it. [03:18] or "Free or I'll sue you" :) [03:18] Personally I'm less likely to base work I'm doing on AGPL than a license that allows modification, but doesn't require redistribution. [03:27] I hope LP doesn't go AGPL, but sabdfl suggests it will be... [03:30] I think AGPL makes a lot of sense for service code [03:31] lifeless: Yes, but many feel it is less than Free. [03:31] it sure sounds like LP is going full-swing in the opposite direction in terms of closedness [03:31] GPL isn't free [03:31] It does make sense. [03:31] GPL is specifically unfree to ensure access to the code transitively across all users [03:33] gpl is free [03:33] coppro: puhlease. You can't make proprietary forks of GPL code. So there are things you are not free to do. [03:34] coppro: BSD is more 'free' from this angle, and public domain is maximally free [03:34] but BSD and public domain don't enforce a ecosystem [03:34] so they are less useful IMO [03:34] AGPL might be less bad if it were less ambiguous. [03:35] they are free by the FSF's definition of free [03:35] feel free to take it up with them [03:35] It doesn't define what degree of interaction with the user is required. [03:35] coppro: the prelude of the GPL states that they deny certain rights in the greater good [03:36] coppro: so please don't invoke greater authority that backs up my point! [03:37] lifeless: Do I need to distribute AGPL source to those users to whom I serve 403s? [03:38] wgrant: I think so [03:38] try #gnu though [03:38] wgrant: thats a really interesting point [03:38] wgrant: spirit-wise I would say no, resoundingly. [03:38] lifeless: But to the letter, I think I'd have to drop all packets from users to whom I didn't want to give source. [03:38] because permission-denied is for someone thats not a use [03:38] *user* [03:39] Even if they couldn't use more than the access-control bit of the webapp. [03:39] I'm just checking [03:39] but only if the 403 is generated by the webapp I think [03:39] if the 403 isn't served by the app, then you don't [03:39] Of course. [03:41] your webapp can't generate 403 messages without giving away source? :) [03:43] "all users interacting with it remotely " [03:43] if you deny a hacker access, I think they are clearly not a user [03:43] AGPL v3 section 13 [03:43] well now there's a vague/grey definition of "user" [03:43] I interacted with your authorisation machinery. [03:43] Give me your code. [03:43] they are interacting with it [03:44] depends on the definition of user [03:44] wgrant: can you not write authorization code that passes the user on to the GPL code? [03:44] I got past your ACL once. give me your code. [03:44] coppro: isn't that linking a GPL app with an AGPL app? [03:44] jdong: you don't need to link [03:45] coppro: oh come on now you are playing on the definition of link to a strict technical definition [03:45] anyhow, I think its fine to deny someone access via an ACL and not offer them source; they haven't become a user at that point [03:45] user isn't defined in the GPL though [03:45] coppro: this is the SAME problem with, saying, popen()'ing to ffmpeg from a proprietary application and saying "I'm not linking to libavcodec" [03:45] While I agree that something like AGPL makes a lot of sense for LP, I don't think the license is good. [03:45] wgrant: +1 [03:45] I like the idea but the implementation looks troublesome [03:46] I support the spirit, but I don't know if I want to adopt the license for a substantial work or not [03:46] lifeless: What if I release my authorisation machinery under the AGPL? [03:46] I guess that's not exactly directly interacting with somebody over the network. [03:46] jdong: I would say it's different; the code doesn't even need to interact with each other [03:46] do an HTTP redirect or something [03:47] coppro: your authorization mechanism doens't need to interact at all with the site it's authorizing? [03:47] wgrant: the phrasing is 'user interacting' not 'machine interacting' or 'human interacting' [03:47] what's your username? [03:47] coppro: I can only do that once my hypothetical authorisation machinery is executed and tells the user to go away. [03:47] lifeless: Is 'user' defined anywhere? [03:47] jdong: well, what I'm saying is, is the authorization machinery related to the site [03:47] wgrant: no [03:47] Sigh. [03:47] wgrant: I quoted the entire relevant text [03:47] coppro: what authorization machinery ISN'T related to the site? [03:47] e.g. does the application need to provide auth services, or can they be managed separately [03:48] coppro: the application needs to at least get the user's id or profile from the authorization service. [03:48] Say I grab a copy of LP, and restrict access to some team, and disable registration. [03:48] coppro: and that would be interacting with it [03:48] you could used a shared database [03:48] which isn't linking [03:48] coppro: that makes no sense. [03:48] wgrant: arguing that a http request which is 403'd with no other content (than the error page) shown is a 'user interaction' would strongly depend on the rest of the site [03:48] why not? [03:48] coppro: user "foo" authenticates now goes into the site [03:48] coppro: how do I know who he is? [03:48] have the auth code check the DB from the site [03:49] cookie [03:49] wgrant: if there is any anonymous access, then the 403 discussion is irrelevant, they can get the code from the 'please log in' page [03:49] coppro: how can you set a cookie? [03:49] coppro: now you're abusing cookies like a communications pipe [03:49] of course [03:49] that's really nothing more than circumventing a license. [03:49] yeah, I guess [03:49] and I thought v3 of the GPL was supposed to be designed to be resistant against these silly technicality tricks? [03:49] wgrant: otherwise, if all interactions until logged in get a 403 challenge, I would argue they are not interacting with it [03:49] wait, wgrant, are you deciding to put a project under AGPL, or trying to use an existing one [03:49] wgrant: they are trying to *start* interacting [03:50] lifeless: They are using my authorisation machinery in order to get their 403. [03:50] btw, I thought we wanted folk to have a valid real name in ubuntu channels [03:50] lifeless: Only launchpad-beta-testers ever required that, and even that was dropped months ago. [03:50] wgrant: but they aren't a user at that point; [03:51] lifeless: How do you know? The license doesn't say. [03:51] wgrant: I'm arguing a point of view [03:51] wgrant: are we discussing applying AGPL to LP? [03:51] or what? [03:51] coppro: That's where this started, but I suspect that this is the failings of the AGPL in general. [03:51] wgrant: well, you can always write an exception in [03:52] a point of view is good, the spirit and ideology is good, but I'm not convinced any of these would be helpful if I had to , say, deal with a legal battle regarding the AGPL and my code. [03:52] coppro: then it's no longer AGPL compatible. [03:52] Greetings all. I filed a bug on window-picker-applet by Canonical, and the bug was fixed [03:52] The i386 Canonical build failed however [03:52] Canonical? [03:52] What does this have to do with Canonical? [03:52] I was wondering what I'd have to do to contribute an i386 build of this applet [03:52] they made it [03:52] I would seek clarification vis-a-vis authenticated services from the fsf; as far as LP goes, the site has anonymous access, so the entire 403 discussion is irrelevant [03:52] ethana2: no, Ubuntu developers made it [03:53] it started as part of ubuntu-netbook [03:53] ethana2: and we do not take contributed binary builds [03:53] ethana2: can you link to the failed build, bug, etc? [03:53] interesting [03:53] ah, i'll try [03:53] lifeless: Canonical's Launchpad instance does. My hypothetical one does not. [03:53] https://edge.launchpad.net/~netbook-remix-team/+archive shows it near the bottom [03:53] the red 'x' with 'i386' next to it in blue [03:53] jdong: my general rule when writing a modification would be to allow the exception to be removed, leaving it compatible. But you're right that it doesn't solve the overall problem [03:54] ok, https://edge.launchpad.net/~netbook-remix-team/+archive/+build/750906 [03:54] https://edge.launchpad.net/%7Enetbook-remix-team/+archive/+build/750906/+files/buildlog_ubuntu-hardy-i386.window-picker-applet_0.4.5_FAILEDTOBUILD.txt.gz [03:54] checking whether build environment is sane... configure: error: newly created file is older than distributed files! [03:54] me is back. [03:54] err, either timestamps are funky on the file you uploaded, or time is funky on the server. [03:54] Hi ScottK. [03:54] I would have to guess the former [03:55] I didn't upload anything [03:55] whoever uploaded it, then [03:55] Hmm, ok [03:55] Since AGPL is not DFSG Free, it's non-free. [03:55] considering it built on all other arches.... [03:55] ethana2: have you tried asking for them to do a rebuild? [03:55] ScottK: DFSG Free? [03:55] ethana2: whoever has admin acceess to that PPA [03:55] How do I go about.. [03:55] ok [03:55] or me [03:55] coppro: Debian Free Software Guidelines. [03:56] so I file a bug asking for a rebuild of 4.5 for i386 fixing timestamps in some file [03:56] ah [03:56] ..right? [03:56] ethana2: no, i just cued ti for a rebuild. [03:56] ethana2: No, that's not right at all. You find a team member, or ask Hobbsee. [03:56] ethana2: well yeah, contact anyone on the netbook-remix-team [03:56] ethana2: they have a button that magically attempts a rebuild [03:56] ..but unless the time stamps are fixed, it'll just fail again, right? [03:57] Hobbsee: do packages often fail to build once, but then just work the second time? [03:57] The Xen instance might have had a bogus timestamp. [03:57] Er, bogus clock. [03:57] Hobbsee: just curious, so buildd-admins apply to PPAs too? [03:57] jdong: oh yes. [03:57] Given that the others succeeded. [03:57] ethana2: not *often* but sometimes [03:57] wgrant: could be timing [03:57] ethana2: depends what the build failure is [03:57] ScottK: how is it not DFSG free? [03:57] interesting [03:57] wgrant: i.e. clock off by 10 minutes on uploader; build queue time [03:57] lifeless: It requires redistribution. Fails the desert island test. [03:57] jdong: Hmm, indeed. [03:57] ScottK: no it doesn't [03:57] ScottK: Nobody is accessing it if I'm on a desert island. [03:58] wgrant: (of course considering how these new tickless kernel things keep time like a clock with a dead battery....) [03:58] ScottK: it requires distribution to folk that use it only [03:58] holy, [03:58] Hobbsee: What? [03:58] fascinating, thanks, Hobbsee [03:58] https://edge.launchpad.net/+builds/samarium I've never seen how this was done before [03:58] oh, there we go, w-p-a has been taken. [03:59] Hobbsee: Lots of i386 builds? [03:59] wgrant: somehow the buildds have managed to get up to a queue of 858 on i386! [03:59] langpacks. [03:59] Lots of langpacks. [03:59] ah, so it is langpacks. I thought there weren't that many [03:59] Could be a couple of releases worth, I suppose. [03:59] that's true [04:00] Wait, but langpack exports aren't meant to be running until 2.1.11, are they? [04:00] i'm used to seeing them at 0, or a very small number [04:00] oh hey, cody-somerville_ :) [04:01] .....or not.... [04:01] ScottK: (which is to say it has exactly the same requirements as regular GPL, modified to the more general 'use' made of web services) [04:01] On the License part of debian/copyright, is it OK to point to the license instead of pasting it? [04:01] lifeless: Not at all. It forces redistridution. [04:02] ScottK: ok, lets get details here. Do you mean section 13? [04:02] * ScottK has to go look it up. [04:02] do so [04:02] http://www.fsf.org/licensing/licenses/agpl-3.0.html [04:03] remember that the desert island test is (paraphrased) 'can I use this software legally on a desert island with no network to the world' [04:03] AGPL passes that test [04:03] Yes. [04:03] I disagree. [04:03] how does it fail - give me a use case [04:03] You can't if you've modified it. [04:03] I'm on an island with my laptop [04:04] I have a webapp - 'helloweb' which is GPL [04:04] on my laptop [04:04] I change the text it shows, run the app, point ff at 127.0.0.1:8080 [04:05] how am I breaking the AGPL at this point [04:05] I have to agree with lifeless here. [04:05] Nobody else is accessing it, so nobody else has the right to get to the code [04:05] I guess. [04:05] there is a 'get code' link on that page, if I click that it gives me the live webapp code [04:05] because thats what the thing is meant to do [04:06] even if there are other people on the island [04:06] if they are able to demand access to the code, they must be in network range of me [04:06] its not *at all* like the brain dead licences that say 'must send changes to authors' [04:06] You argue it better than mako did when I had this discussion with him. [04:07] * ScottK ponders. [04:07] also note that AGPL makes no difference between edited and unedited code- the get source link has to work for the original author as well as downstream sites [04:08] True. [04:08] the original author of GPL or AGPL code with no modifications isn't bound by them always [04:09] coppro: Actually he is bound by AGPL, just not only by AGPL. [04:09] coppro: once he distributes it kicks in [04:09] In any version distributed under the AGPL, the AGPL must be followed. [04:09] He can do whatever he wants with it without the link, but can't distribute it as AGPL. [04:09] lifeless: OK. I think it's DFSG in roughly the same way that Firefox's Trademark license is DFSG free. [04:10] ScottK: hah!. I think its much better than that actually, but unhappy agreement >> disagreement ;) [04:10] for now though, -> code to write [04:10] Forcing distribution as it does is allowed, but one really ought not to do it. [04:10] ScottK: its the moral equivalent of forcing distribution for binaries like 'cat' [04:11] Anyone have an idea how long it takes new builds to propagate to launchpad librarian and PPAs? [04:11] ethana2: They're in librarian seconds after they finish, and published in the PPA after no more than 20 minutse. [04:11] oh, ok, sweet [04:11] so half an hour and then check-update [04:12] * ethana2 hugs .bashrc [04:12] lifeless: It restricts my ability to make use of my own work. There's stuff I write and use that I'm unwilling to distribute because it's a special purpose hack. [04:12] ScottK: so don't permit network access for other people [04:12] ScottK: and voila, you don't have to distribute [04:12] Which kind of defeats the point. [04:12] ScottK: its entirely contingent on other people using the software [04:13] ScottK: which is *precisely* the point of free software [04:13] Interact with/Use [04:13] hmm... reading the DFSG, I don't see any obvious conflicts with the AGPL, but I'm not an expert at this [04:13] which is a fundamental difference. [04:14] coppro: Upon reflection I think it's OK. [04:14] Personally I find it very distasteful and would be very unlikely to make any use of AGPL'ed works in a project of mine. [04:14] I can see where it'd be attractive for LP. === cody-somerville_ is now known as cody-somerville [04:15] I have been running apt-cacher for years, a week or so ago (few days before ibex release) I started getting invalid signature messages. I figured maybe my cache had gotten messed up, and installed apt-proxy on a 2nd box. that wokred fine until today, when I got "Invalid signature" - [04:15] did something in repo land change that may be causing this? todays apt-get error: http://dpaste.com/88601/ [04:15] and where is the right place to post about this? [04:15] cp333? [04:15] CarlFK: I've seen at least two other people with the problem recently. [04:15] perhaps the mirror is dying [04:15] cp333 is my local box [04:16] oh [04:16] hmm [04:16] I don't get that issue [04:16] * StevenK calls his local mirror 'silver' [04:16] wgrant: using a cache, or hitting a real mirror? [04:16] hi ho ? [04:16] Because mirrors used to be made with it [04:16] CarlFK: Using apt-cacher. [04:17] StevenK: Nice, nice. [04:17] wgrant: thanks. I'll stop trying to debug my box :) [04:17] CarlFK: No, keep going, I don't think anybody has worked out why it happens yet. [04:17] ScottK: well, an example of where I'd expect the AGPL to be useful is something like phpBB. [04:17] wgrant: well... um... my plan was to nuke the cache and start over :) [04:18] you want modifiers to have to release source if they let people use it [04:18] wgrant: any sugestions on what to do to debug ? [04:18] CarlFK: No idea. [04:19] welp... just re-ran the command, worked fine. [04:19] CarlFK: Nuke the release files [04:19] apt-cacher is ... thpecial [04:20] None of apt-{proxy,cache{,-ng}} seem to work reliably. [04:20] s/cache/cacher/ [04:21] And mirroring solutions like debmirror and apt-mirror have their own problems [04:21] why did apt-get update say: Hit http://cp600 sid Release.gpg [04:21] debmirror is dumb as a post, and will happily delete 30GB at a pinch [04:22] wait... I may have a sid in my sources.list... [04:22] CarlFK: Because sid appears in your sources.list? [04:22] StevenK: I was mildly displeased when Hardy's debmirror ate my entire mirror due to a bug. [04:22] yep: www.debian-multimedia.org sid main [04:22] wgrant: \o/ [04:22] I had 4 releases. That would have taken like 4 months of off-peak to fix. [04:23] So now I use apt-cacher, and hit it when it breaks. [04:23] wgrant: I managed to avoid that issue. I dumped debmirror, and moved to apt-mirror, since I was sick of the 100-odd line wrapper I had around debmirror [04:24] StevenK: apt-mirror takes something not unlike a sources.list, doesn't it? [04:25] wgrant: I can paste my list if you wish [04:25] coppro: I'd find AGPL useful if I'd written some web service thingy and wanted to make sure no one added stuff to it that I couldn't get access to. [04:25] wgrant: But it's roughly like it [04:25] ScottK: Which is precisely what LP needs. [04:25] But such a license has large issues in practise. [04:25] wgrant: needs/wants, but sure. [04:26] ScottK: Oops, true. [04:26] As a practical matter I think it just discourages other people from caring about extending your software. [04:27] Which might well be considered a feature in this case. [04:27] It wouldn't surprise me. [04:30] After Intrepid, all 'Recommends' get installed, correct? [04:30] yes [04:30] Thank you. [04:32] oh, blast, that build had another bug of the same magnitude that makes it unusable also [04:32] * ethana2 files [04:33] ethana2: Do not file bugs against PPA packages. [04:33] oh [04:33] ok [04:33] Netbook Remix has a bug tracker [04:33] There's a netbook-remix project in LP [04:33] i'm on netbook remix's thing [04:33] code.launchpad.net/~netbook-remix-team/wind.. [04:34] ...so I can file a bug there, right? [04:34] no, that's a team. [04:34] that'snot the project. [04:34] you can't file bugs against people or teams [04:34] Pity [04:34] i know! [04:34] hehe [04:34] ethana2: https://bugs.edge.launchpad.net/netbook-remix [04:34] Grump here wonders if that's OT for MOTU anyway. [04:34] Grump/Grumpy [04:34] It is [04:35] Grumpy and can't type. [04:35] Is the latter the reason for the former? [04:35] Nope. [04:37] bookmarked bug submission page, refreshing X session to verify that this is indeed a bug [04:39] bug 293431 [04:39] Launchpad bug 293431 in transmission "transmission needs to decimal information for large torrents" [Undecided,Invalid] https://launchpad.net/bugs/293431 [04:39] you've got to be joking me. [04:39] THE WORLD WOULD BE A BETTER PLACE IF YOU STOPPED STARING AT THE PROGRESS BAR [04:39] It needs to what now? [04:39] StevenK: the progress bar for torrents needs to have more sig figs [04:39] "When a user (i.e. me) is downloading large torrents (between 4 GIB - 8 [04:39] GIB) the percentage increase between each percentage takes a long time. [04:39] (say from 1 to 2%) . Somewhere down the line I hope it can be made more [04:39] verbose so it shows instead of 1% to something like 1.05 % and so on & [04:39] so forth, so the user knows how much the torrent is moving along." [04:39] Um. It des [04:39] ack crappy paste [04:39] jdong: BAD JDONG [04:40] jdong: It does [04:40] that's a firefox bug [04:40] jdong: It shows 3 significant figures [04:40] StevenK: that's what I thought. [04:40] jdong: I'll kick off a large torrent, and screenshot, and nail the bug shut [04:41] oh, that was shirish. [04:41] Well, it did the last time I ran Transmission, which was Hardy [04:41] that explains all. I thought it looked a bit strange. [04:41] StevenK: it does here too [04:42] what the hell? [04:42] he submitted a bug 1 hour ago [04:42] https://bugs.launchpad.net/ubuntu/+bug/1 (Won't display info) [04:42] and makred it incorrect? [04:42] oh shut up stupid ubottu [04:43] Now now, jdong [04:43] !jdong [04:43]