18:00:00 #startmeeting FESCO (2012-01-23) 18:00:00 Meeting started Mon Jan 23 18:00:00 2012 UTC. The chair is nirik. Information about MeetBot at http://wiki.debian.org/MeetBot. 18:00:00 Useful Commands: #action #agreed #halp #info #idea #link #topic. 18:00:01 #meetingname fesco 18:00:01 #chair notting nirik mjg59 mmaslano t8m pjones sgallagh mitr limburgher 18:00:01 #topic init process 18:00:01 The meeting name has been set to 'fesco' 18:00:01 Current chairs: limburgher mitr mjg59 mmaslano nirik notting pjones sgallagh t8m 18:00:04 hello. 18:00:10 Afternoon 18:00:11 Hello 18:00:11 * notting is here. have to leave in 55 mins, though 18:00:24 Hello 18:00:30 we have a pile of features, then another pile of features. ;) 18:01:00 hi 18:01:08 nirik: Don't forget about the avalanche of features too 18:01:17 sgallagh: oh, there might be some features too. ;) 18:01:23 ok, lets go ahead and dive in. 18:01:27 yeah. and for some reason when those get filed, the urls for the part that's useful to read aren't links. *really* annoying. 18:01:36 from last week: 18:01:39 #topic sponsor request - sdake 18:01:40 .fesco 724 18:01:41 nirik: #724 (sponsor request - sdake) – FESCo - https://fedorahosted.org/fesco/ticket/724 18:01:45 we tabled this last week. 18:01:51 further thoughts now? 18:02:09 Any new package reviews since last week? 18:02:11 Hi 18:02:59 some more in progress I think... 18:03:00 * nirik looks 18:03:34 2 more finished reviews, 3 more in progress. 18:04:38 Sounds like he's committed to the job 18:04:38 * nirik is +1 now. Not a vast number of reviews, but they all seem good to me, and we need more folks helping in the could space with sponsoring. 18:04:46 * sgallagh agrees. +1 18:04:48 yeah, +1 18:04:49 I am +1 as well 18:04:50 +1 18:05:22 sure, +1 18:05:51 #agreed sponsorship is approved. 18:06:05 #topic F17 Feature: DIET - https://fedoraproject.org/wiki/Features/F17_DIET 18:06:05 .fesco 751 18:06:07 nirik: #751 (F17 Feature: DIET - https://fedoraproject.org/wiki/Features/F17_DIET) – FESCo - https://fedorahosted.org/fesco/ticket/751 18:06:20 +1 18:06:22 +1 18:06:33 sure, +1 18:06:40 +1 18:06:55 +1 18:07:05 Fedora on a diet 18:07:16 moderately concerned about the name overlap with the old pseudo-libc and utils, but sure, +1 18:07:16 +1 18:07:32 notting: does anybody but us remember that? ;) 18:07:36 also: diet + corba? 18:07:39 odd combination 18:07:46 notting, :D 18:07:57 * nirik remembers it. man, must be old. 18:07:59 +1 18:08:04 #agreed Feature is approved 18:08:25 #topic F17 Feature: Rework Live CD - https://fedoraproject.org/wiki/Anaconda/Features/ReworkLiveCD 18:08:26 .fesco 752 18:08:27 nirik: #752 (F17 Feature: Rework Live CD - https://fedoraproject.org/wiki/Anaconda/Features/ReworkLiveCD) – FESCo - https://fedorahosted.org/fesco/ticket/752 18:08:41 so, one concern I have here is making live images in koji... 18:08:52 Note that this requires running virt guests as part of rel-eng processes 18:08:55 hey, just in time :) 18:09:15 yep. ;) 18:09:23 mitr: but it also gets us off of the old hacky way to do it, which has always been pretty awful. 18:09:26 so, yeah, it's nice to be able to do live composes in koji... 18:10:09 pjones: I'm fine with it as a matter of principle, just wondering whether rel-eng will be able to use it 18:10:21 can't koji already do live composes? 18:10:27 yeah, I understand. Guess that's a question for rel-eng. 18:10:40 notting: yes, with the current process. not this new process, AFAIK 18:10:45 As far as the feature goes, I'm fine with it 18:10:59 Implementation details may always prevent something we approve from happening 18:11:07 I am fine with the feature +1 18:11:07 I'm not sure this qualifies as a Feature 18:11:11 This sounds more like a process. 18:11:19 *process change 18:11:37 I like moving it all into the same place, I don't like the requirement on libvirt instances. ;( 18:11:40 I guess I don't understand what this adds to Fedora that is worth putting in release notes, etc. 18:11:45 sgallagh: nevertheless in light of no dramatic overhaul of the feature process having occurred, people are still effectively being encouraged to make this sort of thing a Feature. 18:12:04 pjones: I'm not sure it matches up even with the current flawed feature process 18:12:10 Right, having the code is awesome, I don't know about advertising the code if it isn't used 18:12:32 sgallagh: I'm not sure anything does. It's sortof beside the point. 18:12:52 mitr: If it's not used then it won't reach 100%. Or am I misunderstanding you? 18:12:59 * limburgher is here Sorry I'm late, $DAYJOB 18:13:05 I'm +0 on this. 18:13:19 mjg59: integrating the code into rel-eng processes isn't in scope AFAICS 18:13:21 this would also seem to preclude building live images for certain secondary arches 18:13:35 I guess I am ok with the feature, but would want to -1 it if we don't use it in rel-eng to produce things... because if we still use the old method, but advertise the new to users it could cause doom. 18:13:45 mitr: The fallback is to carry on using python-imgcreate 18:13:59 mitr: So if we carry on doing that then the feature isn't complete 18:14:06 mjg59: oh, right 18:14:26 notting: as in, some arches don't have virt at all? 18:14:33 mitr: yep 18:14:36 bcl? 18:14:51 nirik: I'm leaning more towards -1 for this release and work it in tightly to f18 18:14:52 that's what qemu-sh4 is for ;) 18:14:53 proposal: +1 to feature, but drop if not able to be used by releng. Ask feature owners to work with rel-eng to come up with a solution? 18:14:56 hmm, reading... 18:15:23 nirik, that seems like reasonable proposal 18:15:25 * mitr would be +1 to proposal 18:15:25 nirik: I like your proposal, +1 18:15:27 pjones: the horror 18:15:27 Seems kind of late to make that sort of change anyway. Especially since it needs to be finished in time to build the alphas 18:15:36 I'm going with -1 after more thought. 18:15:54 note: you can also use it w/o virt if you are running it from the same release as you are targeting. 18:16:06 Well, I don't think we should -1 out of hand just because we don't know what rel-eng things. We should ask rel-eng instead. 18:16:09 so for secondary arches w/o virt that is an option. 18:16:13 bcl: in a mock/chroot? 18:16:13 bcl: hm, that might simplify things 18:16:32 not in mock. mock and loops don't always work right. 18:16:33 * nirik nods. that might make things doable for the koji plugin. 18:17:06 hum, bummer then. 18:17:13 i am +1 to the idea of the feature, but final completion really does depend on rel-eng 18:17:23 yeah, I'm +1 as well. 18:17:33 +1 on the assumption that if it's not finished then it's not a feature, but that's true of everything isn't it? 18:17:56 mjg59: yes. 18:18:10 well, I was suggesting we add a requirement on it... 18:18:16 that it be used by rel-eng. 18:18:17 * sgallagh remains -1 for F17 18:18:37 nirik: that's already implicitly there though 18:18:43 notting / pjones / mjg59: is that +1 to my proposal? or just the feature as itself? 18:18:56 +1 to the feature 18:18:58 pjones: it is? 18:18:59 just to the feature 18:19:05 I don't see it on the feature page anywhere. 18:19:23 Contingency Plan 18:19:23 Continue to use python-imgcreate 18:19:52 I think it is a separate decision as to whether releng uses it for live media. 18:19:53 but it doesn't have a requirement that we produce our live images with it. 18:19:59 hence the word "implicitly", yes. the contingency plan clearly assumes that rolling it out is part of it. 18:20:09 If we use python-imgcreate then we're on the contingency plan 18:20:17 And if we're on the contingency plan then the feature isn't complete 18:20:22 There's also the 3rd part spins guys who make lots of live stuff. 18:20:50 bcl: so you're saying the feature is really just the availability of it, then? 18:21:01 right. 18:21:06 yes 18:21:16 okay. 18:21:32 I would prefer that it had more testing by spins before being used by releng. 18:21:52 I'd like to avoid the "I downloaded the desktop ks file and make a live with livemedia-creator and it has diffferent bugs from the official media that can be downloaded" 18:22:00 But I think it is a good Feature because it finally gives us a way to make correct images instead of chroot+yum images. 18:22:34 nirik: I think we've already got that 18:22:40 so, we are -1, +4 to feature + rel-eng requirement, and +1 to the feature as is? 18:22:56 nirik: just the fact that you might get updates applied causes that. 18:22:56 pjones: how so? 18:23:48 if you enable them, but you could also change the ks file, etc, etc. I am talking about reproducing the official media... doing one method then asking users to use another is likely to cause confusion and more bugs. I guess it's not the end of the world... 18:24:22 really it should have less. since it uses anaconda :) 18:24:48 * nirik agrees it's the way to go, but just isn't sure we should advertise it before we are using it. 18:25:36 so, where are we then? other votes? thoughts? 18:26:03 nirik: but how do we get the testing bcl wants to have before it goes into rel-eng if the message we're sending is to not use it? ;) 18:26:19 sure there is that... 18:26:23 pjones: Spins, not official releases. 18:26:49 sgallagh: we have official spin releases, no? 18:26:59 proposal: could we add into ticket condition -> speak with releng about it? 18:27:04 * sgallagh shuts up 18:27:29 would folks revote noting if they are ok with the feature as is, or require us to be using it for official live media first? 18:28:18 +1 as is 18:28:46 -1 For F17 18:28:57 +1, but would like to see rel-eng using it or at least a way for them to do so down the road. 18:29:09 i guess i'm -1 as is - concerned about the fact that it requires kickstart changes 18:29:11 I'm still +1 as is as well 18:29:20 +1 and debate with rel-eng 18:29:21 +1 if used by rel-eng (in everything or even in a single small part) 18:29:24 +1 as is, though rel-eng soonish. 18:29:26 I really don't care about the rel-eng requirement although it would be clearly nice to have a word from them if they'll be able to use it 18:29:28 so +1 18:30:18 Alternatively, +1 as is, if the release notes (which are empty currently) explicitly describe the current status 18:30:45 * nirik tries to count votes. 18:31:09 I'm all for explaining the current status in the relnotes 18:31:10 -2, +3 (with releng) and +4 as is I think. 18:31:43 anyone want to change votes? ;) 18:31:56 hm. 18:32:19 Perhaps it would be prudent not to have a multiple-choice here? 18:32:25 if it gets re-defined in the (!releng) case from "here is the new way to build images" to "here is a new way to build images that we're working on...", yes? 18:32:50 sgallagh: open to suggestion. ;) 18:33:43 ok, how about: 18:33:58 proposal: table, ask rel-eng folks for input and revisit next week. ;) 18:34:04 +1 18:34:04 nirik: +1 18:34:07 (when you can't decide, defer!) 18:34:20 nirik, +1 18:34:27 +1 18:34:34 +1 18:34:39 +1 18:34:47 sure, why not. 18:34:55 Who will take that action item? bcl: can you do so? 18:35:17 #agreed will ask rel-eng for input and revisit next week. 18:35:18 err, which? 18:35:45 talk to rel-eng about this and see if there's a way for them to use this to compose offical images. 18:35:45 (ie, talk to dgilmore) 18:35:57 sure. 18:36:07 #action bcl to talk with rel-eng. 18:36:09 thanks 18:36:18 #topic F17 Feature: ABRT Backtrace Deduplication - https://fedoraproject.org/wiki/Features/ABRTBacktraceDeduplication 18:36:18 .fesco 754 18:36:22 nirik: #754 (F17 Feature: ABRT Backtrace Deduplication - https://fedoraproject.org/wiki/Features/ABRTBacktraceDeduplication) – FESCo - https://fedorahosted.org/fesco/ticket/754 18:36:26 +1. less dupes good. 18:36:33 +1 18:36:36 +1 18:36:41 +1 18:36:41 doesn't it already do this? +1 anyway 18:36:44 +1 18:36:45 +1 18:36:51 it does, but this is a new and shiny method I gather. 18:37:10 #agreed feature is approved. 18:37:10 +1 18:37:14 I feel like I should -1 anything that involves a retrace server. 18:37:37 [btw, can I add an agenda item please?] 18:37:39 yeah, sadly thats the path we are on tho. 18:37:50 dmalcolm: sure, can it wait for open floor at the end? 18:37:57 nirik: sure, thanks 18:38:01 #topic F17 Feature: Font Configuration Tool - https://fedoraproject.org/wiki/Features/FontConfigurationTool 18:38:01 .fesco 755 18:38:04 nirik: #755 (F17 Feature: Font Configuration Tool - https://fedoraproject.org/wiki/Features/FontConfigurationTool) – FESCo - https://fedorahosted.org/fesco/ticket/755 18:38:20 pjones: what's the concern? privacy, HW availability, something else? AFAICS this feature does not depend on doing the retrace 18:38:46 +1 to font cfg tools - ideally should be provided by the desktops directly, but having a desktop-independent fontconfig-only tool can't hurt 18:38:50 mitr: no, it just makes use of it. but we should have gone the route of making debuginfo better available on the client end for various reason, privacy included. 18:39:12 +1 to this feature, we could use more good font tools. Hope it's done in time. 18:39:15 +1 to the feature 18:39:18 +1 18:39:19 This is just making the tool available, right - not the default? 18:39:20 +1 18:39:21 sure, +1 18:39:22 mitr: well, it gives users a way to botch theiir own config 18:39:32 mitr: I agree to +1, but I'd like to lobby the devs to make it into an API for easier integration with native desktop environments 18:39:36 yeah, just advertising the tool as I read it. 18:39:41 +1 18:40:01 hey, someone who may have an opinion on this :) 18:40:16 notting: ? 18:40:23 if he's really here. ;) 18:40:37 nim-nim: as a font person, any comments on https://fedoraproject.org/wiki/Features/FontConfigurationTool ? 18:41:33 #agreed feature is approved. 18:41:42 * nirik guesses he's not really around. 18:42:17 #topic F17 Feature: English Typing Booster - https://fedoraproject.org/wiki/Features/english-typing-booster 18:42:18 .fesco 756 18:42:20 nirik: #756 (F17 Feature: English Typing Booster - https://fedoraproject.org/wiki/Features/english-typing-booster) – FESCo - https://fedorahosted.org/fesco/ticket/756 18:42:27 * notting votes -1 to the prior feature, for the record 18:42:57 +1 18:43:03 "Create dictionary tables from wikipedia dumps for English." hm. 18:43:06 I read through this and I couldn't find a clear vision of how this works from upstream 18:43:38 i *think* that should be OK from a licensing standpoint 18:43:45 Why not use existing dicts? 18:44:00 notting: WP is CC-BY, correct? 18:44:07 sgallagh: CC-BY-SA 18:44:09 "Since the suggestions come from a validated word dictionary database, the selected words always give 100% accurate spelling." 18:44:09 They'd need to give credit somewhere. 18:44:13 I am +1 - if this is not on by default which I suppose it isn't 18:44:18 sgallagh: I'm not sure what you mean - AFAICS it's just one more input method that can be manually enable. I didn't find any screenshots/videos either :) 18:44:29 I actually had a conversation about this last week 18:44:33 the experience to set up in how-to-test looks icky 18:44:37 mitr: Right, I missed the word 'ibus' on my first read 18:44:46 Summary: The copyright status of the output of a model trained on a given text is unclear 18:44:47 I wasn't sure if this was supposed to work at the console too 18:44:56 limburgher: I wondered as well, but I'm not interested in micromanaging that 18:45:18 So if the training text is CC-BY-SA, there is an argument that anything you typed with it would have to follow those terms 18:45:42 mitr: Yeah. As long as the licensing is ok. . . 18:45:59 So I'm +1, but would just note that the licensing does need to be checked carefully 18:46:02 mjg59: that's an odd interpretation 18:46:09 I don't see why adding such new package is a feature but it's +1 18:46:22 I don't think wordlist created of a text is a derived work of that text if the text is sufficiently large. 18:46:23 * nirik is +1 I guess. 18:46:39 mjg59: but I think what you're saying is that no amount of checking will actually tell us where we stand? 18:46:43 I'm -1 while there are legal considerations. Send it to legal@ first 18:46:52 * notting agrees with sgallagh. -1 18:46:59 pjones: well, switching to /usr/share/dict/words solves that problem 18:47:02 If this only looks at the context of the single word, I suspect it's fine 18:47:17 If it starts producing predictions based on previous words then things get a little more confusing 18:47:40 mjg59: https://fedorahosted.org/english-typing-booster/browser/english-typing-booster/tables/english-dict.txt (3M download) - single words AFICS 18:47:42 For instance, if choosing the best prediction each time results in it tending to generate the original work... 18:47:55 * nirik notes his +1 was on the assumption that legal was ok 18:48:24 nirik: I think it should be vetted before we give any tacit approvals 18:48:41 mjg59: haha 18:48:42 sgallagh: 18:48:52 Ok - single words so I guess it's probably ok 18:48:58 Although the words list does claim to be GPLv3 18:49:02 I really do not think there is any legal trouble with wordlists 18:49:11 mjg59: well, that's obviously wrong 18:49:13 so +1 18:49:19 so, I see +4, -2 ? 18:49:36 +1 to the feature, and unrelatedly I think it should have some legal checking 18:49:39 mjg59: and i'm not sure if it's even valid 18:49:45 I'm +1 to the feature. 18:49:55 notting: Yeah, quite 18:49:56 +1 to mjg59 18:50:09 +1 if it's legit. 18:50:15 but yeah, it's clear we need to do some legal homework on this one 18:50:35 pjones: s/we/feature owners/? 18:50:35 * notting remains -1 until legal homework done 18:51:04 *shrug* i'm willing to undertake that *with* the feature owners 18:51:36 ok, I think thats +6, -2.. 18:51:42 notting: please do? :) 18:51:48 mitr: notting, obviously. 18:52:15 #agreed feature is approved. 18:52:34 do we want to leave some kind of tracking in place for the legal aspect? 18:53:01 nirik: What do you propose? File a BZ? 18:53:06 notting: I'm not sure it's the cleanest approach possible but they're trying to make fontconfig work in in the real world 18:53:26 sgallagh: perhaps we leave the ticket open and revisit it in our meetings ? 18:54:04 worksforme 18:54:19 ok, so thats the end of the features that were on our initial agenda... 18:54:26 but then we got a pile more. 18:54:38 do we want to look at them today? defer to next week? or look at a subset? 18:55:10 notting: as opposed to the previous wishful thinking (someday someone will create a perfect font set with no overlaps that fontconfig can handle automatically without config) 18:55:10 I think we should look at them today 18:55:23 * nirik is fine to look at them today. 18:55:26 I think we should hit as many as we can in the next, say, hour 18:55:32 we can always defer on ones we want more info on 18:55:36 yes 18:56:07 Today, please - the work won't go away if we defer 18:56:37 Perhaps there is an uncontroversial set that we could block vote on? 18:56:37 ok. i have to head out, i can vote in tickets later 18:56:49 notting: have fun. 18:57:16 notting: fontconfig worst weakness is that it only manipulates complete fonts, and since they all overlap in non-obvious way it's not possible to create a single order everyone likes as keithp thought 18:57:37 pjones / mjg59 / limburgher: shall we try and go on? or you guys want to postpone? 18:58:07 May as well go on 18:58:28 notting: and no one has stepped up to do the clean thing (enhance fontconfig so it can work on font subsets that can be defined in non-overlappy way) 18:58:33 Go on. 18:58:40 we're still at quorum, so may as well make an attempt with an eye on deferring quickly if consensus isn't clearly reachable. 18:58:50 ok, onward. 18:58:51 #topic #758 F17 Feature : Eclipse Juno - https://fedoraproject.org/wiki/Features/EclipseJuno 18:58:52 .fesco 758 18:58:53 nirik: #758 (F17 Feature : Eclipse Juno - https://fedoraproject.org/wiki/Features/EclipseJuno) – FESCo - https://fedorahosted.org/fesco/ticket/758 18:59:17 +1 18:59:24 sure, +1 18:59:26 notting: as far as I understand the i18n feature is to let users reorder fontconfig prefs any way they like, since a perfect distro order is not possible 18:59:34 +1 18:59:40 Is this filing early? 18:59:56 It's claiming that the upstream will be released in June 19:00:02 this seems like their timeline is f18 not f17 19:00:04 That seems rather too late for F17 19:00:06 +1 19:00:08 sgallagh: they want to ship a prerelease 19:00:09 -1 19:00:22 -1 that's what rawhide is for 19:00:26 "Upstream has asked that we clearly communicate that we will be shipping a pre-release in Fedora (due to timing) and will be updating to their final version as soon as it is released. " 19:00:38 I do not like shipping prereleases in the final GA releases of Fedora 19:00:42 +1 - if this is agreed with upstream, why not 19:00:59 sgallagh: Maintainers get to make judgement calls on whether they think something will be sufficiently stable 19:00:59 t8m: We ship a prerelease of GNOME practically the entire development cycle 19:00:59 yeah, I don't much like sticking a prerelease in f17 and then updating to a final (plausibly quite different) version in an update. 19:01:00 -1 19:01:06 t8m: There is definitely precendent for that strategy, though. 19:01:11 mitr, I don't like it either 19:01:16 It works for Firefox. 19:01:27 (GNOME, KDE< and firefox have all done it) 19:01:34 As long as the expectation is that changes will be minimal, I'm fine with this 19:01:41 abadger1999, I know 19:01:47 Generally their final releases are going to happen before Fedora's GA 19:01:51 yup 19:01:54 we're always screwed 19:02:00 either early (this case) or late (4 months) 19:02:07 We're talking about at least a month of additional work here. 19:02:13 I pushed for early this time 19:02:16 Not including slippage 19:02:20 yeah, some upstreams just don't align with our release cycle. 19:02:23 honestly, 99% of upstream is done by April 19:02:30 they have serious ramp-downs 19:02:33 and they don't slip 19:02:34 nirik: Of course, but they should wait for the next bus 19:02:34 guaranteed 19:02:36 overholt_: so we're really talking about very minimal changes? 19:02:39 pjones, yup 19:03:04 I count +5 anyway... 19:03:06 The schedule says API freeze on March 16, i.e. before our Beta 19:03:15 yup 19:03:19 okay, I can be convinced then. 19:03:21 Personally, I rely on Eclipse rather a lot. I'm leery of having a prerelease in F17 19:03:32 no point in holding something up on process formalities if the real changes are already done. 19:03:45 pjones: If we vote it down, it gets backed out 19:04:01 sgallagh: I'm familiar with how voting works, yes. 19:04:09 sgallagh: Er. 19:04:12 I don't want to do an update post-release 19:04:15 sgallagh: No, if we vote it down, it's not a feature. 19:04:19 * nirik counts votes. 19:04:30 sgallagh: We could separately have a discussion on whether to overrule the maintainer 19:04:32 nirik: given overholt_'s input I'm +1 as well. 19:05:03 Obviously I've been outvoted, but I'm -1 (I think this will bite us) 19:05:18 sgallagh, I'd like to work with you personally as an active user to ensure it doesn't 19:05:22 mjg59: I assumed he meant backed out as a feature, not from the distro, but on re-read I see what you're saying. 19:05:30 either way, it seems to be approved, n'est pas? 19:05:37 I think thats +6, -2 ? 19:05:52 #agreed feature is approved 19:06:02 thanks 19:06:18 #topic #759 F17 Feature: Opa - https://fedoraproject.org/wiki/Features/Opa09 19:06:18 .fesco 759 19:06:20 nirik: #759 (F17 Feature: Opa - https://fedoraproject.org/wiki/Features/Opa09) – FESCo - https://fedorahosted.org/fesco/ticket/759 19:06:24 overholt_: Well, I'll be testing the new release, absolutely :) 19:06:32 +1 19:06:39 But I'm wary of it being stable enough for Fedora in general 19:06:49 +1 19:06:55 +1 19:07:01 sgallagh: that was a joke, right? 19:07:02 +1 19:07:03 +1 Looks straightforward to me 19:07:14 (Do things built with Opa end up under AGPLv3 as well?) 19:07:23 +1 19:07:26 +1 19:07:32 I sure would think not. 19:07:33 pjones: You're such a cynic :) 19:08:30 #agreed feature is approved 19:08:34 nirik: Seems a slightly odd license to choose in that case, but I don't think it makes any practical difference 19:08:57 yeah 19:09:18 mjg59: Better ask that question of spot if you're concerned/want it documented in Fedora. 19:10:06 #topic #760 F17 Feature: Internationalization support for additional Indian languages - http://fedoraproject.org/wiki/Features/Additional_Indic_Langs 19:10:07 .fesco 760 19:10:08 nirik: #760 (F17 Feature: Internationalization support for additional Indian languages - http://fedoraproject.org/wiki/Features/Additional_Indic_Langs) – FESCo - https://fedorahosted.org/fesco/ticket/760 19:10:15 Sure +1 19:10:16 +1 19:10:17 might be good to ask that on the package review (which is still open) 19:10:25 +1 19:10:25 +1 19:10:29 +1 19:10:29 22 official languages is absurd, even for a country the size of India. +1. 19:10:41 +1 19:10:49 #agreed feature is approved 19:10:50 +1 19:11:12 #topic #761 F17 Feature: DNSSEC on workstations - https://fedoraproject.org/wiki/Features/DNSSEC_on_workstations 19:11:12 .fesco 761 19:11:14 nirik: #761 (F17 Feature: DNSSEC on workstations - https://fedoraproject.org/wiki/Features/DNSSEC_on_workstations) – FESCo - https://fedorahosted.org/fesco/ticket/761 19:11:15 pjones, you mean something like European Union? 19:11:41 t8m: not a country. 19:11:52 t8m: as they seem hellbent on proving right now. 19:12:12 dnssec please 19:12:24 +1 to dnssec. 19:12:30 +1 dnssec everywhere! 19:12:39 +1 19:12:42 +1 19:12:47 Mildly +1 - like the idea, not absolutely certain about UI 19:12:47 +1 19:13:03 +1 19:13:27 +1 same concerns as mitr 19:13:38 #agreed feature is approved. 19:13:42 #topic #762 F17 Feature: Eclipse Juno - https://fedoraproject.org/wiki/Features/EclipseJuno 19:13:43 .fesco 762 19:13:44 nirik: #762 (F17 Feature: Eclipse Juno - https://fedoraproject.org/wiki/Features/EclipseJuno) – FESCo - https://fedorahosted.org/fesco/ticket/762 19:14:00 nirik: we talked about that 19:14:06 oops. 19:14:07 yeah, sorry. 19:14:08 #undo 19:14:08 Removing item from minutes: 19:14:28 #topic #763 F17 Feature: Ext4 support beyond 16T - https://fedoraproject.org/wiki/Features/F17Ext4Above16T 19:14:28 .fesco 763 19:14:30 nirik: #763 (F17 Feature: Ext4 support beyond 16T - https://fedoraproject.org/wiki/Features/F17Ext4Above16T) – FESCo - https://fedorahosted.org/fesco/ticket/763 19:14:31 it had different number 19:14:43 +1 19:14:46 +1 19:14:46 +1 to letting sandeen do his job. 19:14:47 +1 19:14:48 +1 19:14:49 +1 19:14:52 +1 19:15:03 +1 19:15:08 Any chance he can send us some 32T drives to test with? :) 19:15:19 #agreed feature is approved. 19:15:19 you don't want that. 19:15:23 I would totally help with that. :) 19:15:23 ha 19:15:35 #topic #764 F17 Feature: Haskell Platform 2011.4 - https://fedoraproject.org/wiki/Features/HaskellPlatform2011.4 19:15:38 .fesco 764 19:15:40 nirik: #764 (F17 Feature: Haskell Platform 2011.4 - https://fedoraproject.org/wiki/Features/HaskellPlatform2011.4) – FESCo - https://fedorahosted.org/fesco/ticket/764 19:15:45 +1 I guess... 19:15:46 +1 19:15:57 +1 19:16:04 sure, why not. +1 19:16:07 +1 19:16:08 +1 Not loving the lack of a contingency plan, but. . . 19:16:16 +1 19:16:30 #agreed feature is approved. 19:16:31 limburgher: Do enough people use Haskell for us to care about contingencies? 19:16:32 +1 19:16:41 #topic #765 F17 Feature: New Features of IPA v3 - https://fedoraproject.org/wiki/Features/IPAv3NewFeatures 19:16:41 .fesco 765 19:16:43 nirik: #765 (F17 Feature: New Features of IPA v3 - https://fedoraproject.org/wiki/Features/IPAv3NewFeatures) – FESCo - https://fedorahosted.org/fesco/ticket/765 19:16:43 (Not trying to be flippant, just realistic) 19:16:49 sgallagh: I work with a guy. :) 19:16:57 Haskell updates aren't provably correct? 19:17:14 +1 19:17:21 ha. 19:17:25 sure, +1 19:17:27 +1 19:17:30 sgallagh: Understood. :) 19:17:31 +1 19:17:31 +1 19:17:37 +1 19:17:47 +1 for IPA, full of hops and... oh wait, wrong IPA. 19:17:49 +1 19:17:58 #agreed feature is approved. 19:18:04 #topic #766 F17 Feature: Java 7 - https://fedoraproject.org/wiki/Features/Java7 19:18:05 .fesco 766 19:18:06 nirik: #766 (F17 Feature: Java 7 - https://fedoraproject.org/wiki/Features/Java7) – FESCo - https://fedorahosted.org/fesco/ticket/766 19:18:07 nirik: Come to some of the presentations I give. I occasionally provide FreeIPA-labeled beer :) 19:18:32 sgallagh: oooh. good to know. ;) 19:18:44 +1 19:18:46 +1 on java. I assume most of this already got rebuilt in the mass rebuild? 19:18:50 +1 to Java 7. Last I heard, they had worked out the majority of the issues from F16 19:18:53 +1 19:18:55 +1 19:18:58 sgallagh, although the ssh host keys support must be upstream first! 19:19:16 +1 19:19:24 +1 to java 7 19:19:27 +1 19:19:32 #agreed feature is approved. 19:19:39 #topic #767 F17 Feature - New mkdumprd for kdump - https://fedoraproject.org/wiki/Features/NewMkdumprd 19:19:39 .fesco 767 19:19:41 t8m: The Fedora maintainer has agreed to carry it in Fedora 19:19:42 nirik: #767 (F17 Feature - New mkdumprd for kdump - https://fedoraproject.org/wiki/Features/NewMkdumprd) – FESCo - https://fedorahosted.org/fesco/ticket/767 19:19:59 +1 better late than never. 19:20:20 +1 19:20:21 +1 19:20:22 +1 19:20:25 +1 19:20:26 +1 here. 19:20:28 +1 19:20:28 +1 19:20:31 #agreed feature is approved. 19:20:37 #topic #768 F17 Feature: Virtualization Sandbox - https://fedoraproject.org/wiki/Features/VirtSandbox 19:20:38 .fesco 768 19:20:39 nirik: #768 (F17 Feature: Virtualization Sandbox - https://fedoraproject.org/wiki/Features/VirtSandbox) – FESCo - https://fedorahosted.org/fesco/ticket/768 19:20:44 +1 19:20:52 +1 19:20:58 +1 WANT 19:21:03 sure, +1 19:21:03 +1 19:21:05 +1 19:21:49 +1 19:21:56 #agreed feature is approved. 19:22:29 #topic #769 F17 Feature: numad (Non-Uniform Memory Alignment Daemon) - https://fedoraproject.org/wiki/Features/numad 19:22:30 .fesco 769 19:22:32 nirik: #769 (F17 Feature: numad (Non-Uniform Memory Alignment Daemon) - https://fedoraproject.org/wiki/Features/numad) – FESCo - https://fedorahosted.org/fesco/ticket/769 19:22:43 not sure how many people will care about this one, but sure, +1 for those that do 19:22:51 +1 19:22:56 it's too bad they couldn't find a way to name it "nomad". +1. 19:22:58 +1 19:23:05 Is this intended to run by default? 19:23:19 No, service off by default 19:23:25 +1 19:23:29 Excuse my ignorance, but what systems are NUMA-enabled? 19:23:39 Is this a standard feature of Core processors or something? 19:23:54 Multi-socket systems 19:24:01 sgallagh: well, amd64 is technically numa on multi-socket... 19:24:02 * rbergeron spreads the plague of spot 19:24:15 bgray: No impact on multi-core, single-socket then? 19:24:32 bgray: Or cost if enabled and not helpful? 19:24:49 No, unless new AMD , which has two NUMA nodes per socket 19:25:07 +1 19:25:09 It won't do anything unless there is >1 NUMA node 19:25:12 +1 19:25:19 +1 19:25:25 #agreed feature is approved. 19:25:28 #topic #771 F17 Feature: OpenStack Quantum - https://fedoraproject.org/wiki/Features/OpenStack_Quantum 19:25:29 .fesco 771 19:25:29 nirik: #771 (F17 Feature: OpenStack Quantum - https://fedoraproject.org/wiki/Features/OpenStack_Quantum) – FESCo - https://fedorahosted.org/fesco/ticket/771 19:25:54 +1 19:26:01 +1 19:26:08 +1 19:26:10 I'm not a huge fan of the contingency plan 19:26:14 +1 here 19:26:16 "If it's not ready, we'll ship it anyway" 19:26:30 +1 19:26:41 +1 anyway 19:26:47 yeah, thats not ideal for sure. 19:26:48 sgallagh: that's just one plugin among several 19:26:52 +1 19:27:35 #agreed feature is approved. 19:27:40 #topic #772 F17 Feature: OpenStack using Qpid - https://fedoraproject.org/wiki/Features/OpenStack_using_Qpid 19:27:41 .fesco 772 19:27:42 nirik: #772 (F17 Feature: OpenStack using Qpid - https://fedoraproject.org/wiki/Features/OpenStack_using_Qpid) – FESCo - https://fedorahosted.org/fesco/ticket/772 19:27:51 +1 19:27:55 +1 19:28:05 +1 19:28:11 +1 19:28:13 * nirik wonders if some of the openstack features couldn't be folded into one 'better openstack' one... but I guess they might be running at different speeds, etc. 19:28:27 +1 19:28:32 +1 19:28:43 +1 anyhow. 19:29:03 #agreed feature is approved. 19:29:23 #topic #773 F17 Feature: Ruby 1.9.3 - https://fedoraproject.org/wiki/Features/Ruby_1.9.3 19:29:23 .fesco 773 19:29:24 nirik: #773 (F17 Feature: Ruby 1.9.3 - https://fedoraproject.org/wiki/Features/Ruby_1.9.3) – FESCo - https://fedorahosted.org/fesco/ticket/773 19:29:27 +1 19:29:28 +1 19:29:39 +1 19:29:41 +1 19:29:41 +1 19:29:49 +1 19:29:49 +1 19:30:13 #agreed feature is approved. 19:30:42 #topic #774 F17 Feature: SELinux Deny Ptrace - https://fedoraproject.org/wiki/Features/SELinuxDenyPtrace 19:30:42 .fesco 774 19:30:44 nirik: #774 (F17 Feature: SELinux Deny Ptrace - https://fedoraproject.org/wiki/Features/SELinuxDenyPtrace) – FESCo - https://fedorahosted.org/fesco/ticket/774 19:31:05 +1 19:31:15 +1 19:31:19 +1 19:31:29 +1. I'd almost want it on by default, but I'm sure that would cause pain and yelling. 19:31:30 +1 , and /me would eventually like to be able to enable this by default without impacting normal operation 19:31:32 +1 19:31:38 +1 19:31:53 #agreed feature is approved. 19:32:17 though my real question is: how does this interact with abrt? does abrt wind up using ptrace at all? 19:32:17 #topic #775 F17 Feature: PrivateTmp services - https://fedoraproject.org/wiki/Features/ServicesPrivateTmp 19:32:17 .fesco 775 19:32:18 nirik: I imagine setroubleshoot would make flipping the switch painless, but the number of programs that currently require ptrace for normal operation seems pretty large 19:32:19 nirik: #775 (F17 Feature: PrivateTmp services - https://fedoraproject.org/wiki/Features/ServicesPrivateTmp) – FESCo - https://fedorahosted.org/fesco/ticket/775 19:32:28 Worth mentioning that it also breaks strace? 19:32:30 pjones: I think it gets a core file on stdin 19:33:04 Well, that was fun. 19:33:31 +1 to ServicesPvtTmp 19:33:36 Is there a general concern WRT kerberos? 19:33:38 +1 as well 19:34:00 +1 here as well 19:34:05 mitr: We're hoping to move kerberos over to using /var/run/user 19:34:08 +1 19:34:22 And updating gssd and friends to look there by default 19:34:37 sgallagh: Fair enough, so it's being handled... 19:34:41 +1 to feature 19:34:43 +1 19:35:13 ok +1 19:35:15 #agreed feature is approved. 19:35:22 #topic #753 F17 Feature: Wallaby - https://fedoraproject.org/wiki/Features/Wallaby 19:35:26 .fesco 753 19:35:27 nirik: #753 (F17 Feature: Wallaby - https://fedoraproject.org/wiki/Features/Wallaby) – FESCo - https://fedorahosted.org/fesco/ticket/753 19:35:58 +1 plus disclaimer about the feature process 19:36:03 why not +1 19:36:05 +1 19:36:12 sgallagh: specifically? 19:36:16 +1 19:36:30 +1 - sounds vaguely like puppet, but google doesn't know about any connection 19:36:38 mitr: Eh, just a heavily-targeted feature interesting to a very small subset of users 19:36:40 +1 19:37:01 sgallagh: Ah, ok. 19:37:14 #agreed feature is approved. 19:37:25 +1 why not 19:37:55 ok, thats finally the end, however... we had some more in ready for wrangler... 19:38:02 https://fedoraproject.org/wiki/Category:FeatureReadyForWrangler 19:38:11 rbergeron: any of those ready? or those for next week? 19:38:28 (Note for mjg59 and nirik Re: Opa licensing -- The OPA upstream considers anything written in OPA to be licensed under the AGPL unless you buy a separate license from them: http://blog.opalang.org/2011/08/opa-license-contributions.html ) 19:38:35 Let's say next week? 19:38:39 abadger1999: really? wow. 19:38:39 * mull pinged rbergeron about Euca, but she just got home 19:38:40 abadger1999: Ok that makes sense 19:38:52 abadger1999: Oof. In that case, should we be shipping it? 19:38:58 nirik: Only network-zones has any comment by rbergeron on the talk page 19:39:05 sgallagh: It's not closed source.... 19:39:15 I... guess? 19:39:21 * nirik is ok with doing any features that are ready by tomorrow next week. 19:39:25 sgallagh: does that differ from the old mysql dsituation? 19:39:31 abadger1999: wow. that's bordering on "maybe we shouldn't be shipping this" 19:39:42 nirik, I think we can do them next week 19:39:42 #topic Opa redux 19:39:50 pjones: Eh, it's along the lines of anything that has a GPL runtime library 19:39:52 do we want to revisit that feature/package? 19:40:04 Yeah, we've managed 25 features today, so we probably can handle 6 next week :) all of them are fairly new 19:40:23 mjg59: hence the "maybe". 19:40:53 I guess it's up to people using that to comply with the license. 19:41:00 As this is really open source I don't think we should block the opa package from fedora however it's questionable if the advertising-by-feature is justified here 19:41:01 * abadger1999 thinks opa-agpl shouldn't affect the "feature" or inclusion... will just affect adoption and future license auditing. 19:41:29 AGPLv3 is a valid license for fedora packages as well. 19:41:51 Eh, let's run with it. 19:41:59 so, I'm fine with leaving it as is, but perhaps note in the feature page/release notes about the licensing? 19:42:08 nirik: +1 19:42:26 Perhaps mandate that the RPM description should include that caveat as well? 19:43:36 * mitr really doesn't know whether to advertise the company or not 19:43:42 or perhaps FPC could look at any additional documentation needs for AGPL packages? 19:44:04 anything more on this? or shall we move on? 19:44:11 In the long run, I agree with mjg59 -- it's no different than gpl on runtime libraries. 19:44:14 It's not the AGPL itself that's the problem. It's that the output of the package is considered a derivative work. 19:44:38 That's a little unusual for what's effectively a compiler, but it's not without precedent 19:44:44 even that isn't a problem, it's more than people may not expect it and be ready to meet the licensing needs. 19:45:22 In the shorter term, people might be confused that gcc doesn't demand that the programs compiled against it be gpl licensed whereas opa does make that demand. 19:45:46 * nirik nods. 19:46:01 so, document and move on? ask fpc to visit? 19:46:14 * abadger1999 is for document and move on 19:46:27 I'd assume that people who want to ignore licenses will ignore this just as well as anything else, and people who are meticulous will notice (BTW some of the LGPL requirements are quite surprising for a binary-only distributor) 19:46:56 #info OPA should document the output licensing 19:47:00 * nirik is fine with that. 19:47:12 I had one more thing before the lovely open floor... 19:47:12 +1 document and move on 19:47:14 mitr: yes, but there's some expectation that compilers have exemptions, which is sortof a shortcut for not /needing/ to be meticulous about their output. 19:47:43 nevertheless, it does appear to comply with our rules, it's just messy. 19:47:47 pjones: *shrug* I was thinking about the mandatory runtime 19:47:50 +1 document and move on 19:48:01 #topic EOL package branches default acls 19:48:11 so, I'd like to ask fesco for any opinion on: 19:48:16 https://fedorahosted.org/fedora-infrastructure/ticket/3094 19:48:30 basically do we care what acls the git repos keep for eol releases? 19:48:43 if we nuke them all and default to deny it saves info that pkgdb needs to keep 19:49:00 worksforme. (wow, this meeting is still going) 19:49:03 Seems sane to me. We're not realistically allowing changes to be made. 19:49:14 +1 to not caring. EOL is EOL. 19:49:15 welcome back notting. ;) 19:49:55 * nirik is fine with defaulting to deny and saving space/time/pkgdb cycles. 19:49:58 +1 19:50:07 I thought EOL are blocked anyway,so why should allow changes +1 19:50:29 #agreed fesco is fine with default acls for eol releases. 19:50:33 +1 19:50:43 #topic Open Floor 19:50:47 +1 to blocking access 19:50:49 dmalcolm: you had something? 19:50:54 hold on 19:51:03 holding. 19:51:05 I'm sorry for not being as on top of feature process as I should be; I'm dealing with a Python security vulnerability right now. 19:51:05 I suspect this is more a question for rbergeron 19:51:06 Questions re: http://fedoraproject.org/wiki/Features/StaticAnalysisOfCPythonExtensions 19:51:06 In particular, see the "Current Status": 19:51:06 ... but I wonder about archiving the data in case we needed to look up past history 19:51:19 F16 shipped with the static checker, but with only some minor tests enabled, that I didn't feel were particularly exciting. 19:51:19 Given this, I tried to revoke this for F16, but somehow I think it still got listed as a F16 feature. 19:51:19 The "really compelling" part of it described on that feature page *is* now working, and I've been using it to find memory leaks in C Python extension modules (and other bugs). 19:51:51 I want to do an F17 feature of this with this new code 19:52:05 QUESTION: Do I start a new feature page, or reuse the old one? 19:52:34 I'd do a new one, based on and referencing the old. 19:52:47 "StaticAnalysisOfPythonRefCounts" perhaps 19:52:57 Yeah, it seems this was in f16 release notes 19:53:02 that scopes it better also 19:53:07 A clone of the page seems cleanest. 19:53:20 yeah, I'd do a new one. 19:53:34 mitr: and then hack it into shape? (before the deadline, of course) 19:53:44 That way people don't have to look at the wiki edits to see history. 19:53:49 yeah, I'd also say do a new one. 19:53:55 ok, thanks. 19:53:56 will do 19:54:04 thanks. that was my question 19:54:17 dmalcolm: deadline to submit is tomorrow tho... ;( So, get it in and we can talk about it next week? 19:54:21 * dmalcolm goes back to security vuln 19:54:30 nirik: it will be a long day :) 19:54:33 we can also consider exceptions if necessary 19:54:40 although obviously not encouraged 19:55:06 yeah. 19:55:09 anything else for open floor? 19:55:10 will try to get it in by deadline 19:55:21 I'd like to bring up the topic of Closed/UPSTREAM (at the risk of starting a flame war) 19:55:33 #topic Cloed/Upstream bugs 19:55:52 Given that we're almost at two hours, I'd like to suggest doing this next week 19:56:02 * nirik is fine with that as well. 19:56:04 great idea 19:56:15 I'm okay with that 19:56:17 Yes. 19:56:45 please 19:56:49 #action defer to next week. 19:56:56 #topic Open floor again 19:57:10 will close out in a minute if nothing more comes up... 19:57:26 next chair? 19:57:30 ah yes. 19:57:36 who wants the hot seat next week? 19:57:53 I can 19:58:04 #action mmaslano will chair next week 19:58:06 thanks! 19:58:14 thanks for coming everyone. 19:58:21 #endmeeting