18:01:35 #startmeeting FESCO (2015-07-01) 18:01:35 Meeting started Wed Jul 1 18:01:35 2015 UTC. The chair is thozza. Information about MeetBot at http://wiki.debian.org/MeetBot. 18:01:35 Useful Commands: #action #agreed #halp #info #idea #link #topic. 18:01:36 #meetingname fesco 18:01:36 #chair ajax dgilmore hguemar jwb nirik paragan rishi thozza sgallagh 18:01:36 #topic init process 18:01:36 The meeting name has been set to 'fesco' 18:01:36 Current chairs: ajax dgilmore hguemar jwb nirik paragan rishi sgallagh thozza 18:01:40 \o/ 18:01:42 .hello sgallagh 18:01:45 hi all 18:01:51 Hi 18:01:58 .hello hguemar 18:02:00 hello everyone 18:02:15 Hi! 18:02:38 .hello jkurik 18:02:47 I count 4 18:03:09 dgilmore wrote he won't be around 18:03:14 sgallagh: sgallagh 'Stephen Gallagher' 18:03:17 sgallagh: sgallagh 'Stephen Gallagher' 18:03:20 number80: hguemar 'Haïkel Guémar' 18:03:23 jkurik_mtg: jkurik 'Jan Kurik' 18:03:41 zodbot: naughty you 18:03:47 I think nirik also on vacation 18:04:01 he is 18:04:17 jwb: with you we are 5 18:04:30 hi 18:04:34 * rishi` is here 18:04:40 quorum reached :) 18:04:50 ok, so let's start, we have a lot of work to do 18:05:02 #topic #1450 F23 System Wide Change: Python 3 as the Default Implementation 18:05:02 .fesco 1450 18:05:03 thozza: #1450 (F23 System Wide Change: Python 3 as the Default Implementation) – FESCo - https://fedorahosted.org/fesco/ticket/1450 18:05:20 So I asked change owners to participate, they are here 18:05:26 I had a discussion with them 18:05:27 <- here 18:05:33 <- here 18:05:57 the bottom line is that the change was technically around for 2 releases, but the first one was aiming for rawhide 18:06:04 so this is second and last time 18:06:09 er 18:06:12 no 18:06:16 Here as well 18:06:28 jwb: no what? 18:06:46 the main argument against this was what "default" means, and several people on FESCo expressed that default should include the tools actually used to build Fedora 18:07:10 agreed 18:07:35 We have defined in scope what does default means. 18:07:46 and FESCo members disagreed with that definition 18:07:59 jwb: not all of them 18:08:13 correct. so call for a vote and see if you have enough that do 18:08:17 jwb: so you would like to have also the fedora infra tool in the scope? 18:08:18 So yet again, we argue over the name of the Change instead of its contents? 18:08:18 * rishi` raises his hand 18:08:22 Lets first discuss this please. 18:08:38 Regarding the issue of Koji using python3 ... 18:08:38 so it's about the name? 18:08:41 jwb: I think we should sort out the issues rather than vote and let it fail 18:09:01 ajax raised it with Adam Miller during a presentation that he did 2 weeks ago. 18:09:04 thozza: +1 18:09:08 Stating that Python3 is default for Fedora as a distribution is important in eyes of community and projects which yet do not fully support Python3. 18:09:13 the issues have been pointed out. they were hand waved away with "not in the defined scope" 18:09:29 all i really have to say here is, if 'python3 is self-hosting' isn't an eventual goal, i think 'python3 is default' isn't an especially meaningful set of goals 18:09:35 rkuska: Lying to get better marketing is not generally valued, OTOH 18:09:39 if those goals are all that's being attempted, fine 18:09:41 jwb: I don't think that fedora infra is a show stopper 18:09:42 This is not lying. 18:09:43 but i'd want better 18:09:46 is more a nice to have thing 18:09:57 thozza, yes. we had this discussion 2 weeks ago. you and i disagree 18:10:03 We will have only python3 shipped by default on live distribution. 18:10:10 it requires time and I'm sure it will be in TODO, but it is impossible to make it for F23 18:10:19 then the change should be deferred 18:10:26 jwb: I think making the Python3 as default is beneficial for making others to actually change their packages and do some work. Also Py2 will go EOL in 2020 18:10:32 thozza: I am saying that we did raise it with Fedora infra' (Adam Miller) 2 weeks ago. 18:10:47 rishi`: what was the outcome? 18:10:51 thozza, we disagree on that too. it isn't "default" today and anaconda was ported just fine 18:10:59 He didn't blow the idea way completely, in my understanding. 18:11:12 jwb there are many other projects which wait for python3 becomes more mainstream 18:11:12 i'm not interested in completely rehashing the conversation we had 2 weeks ago and on the list in the interim 18:11:14 jwb: because it was advertised last release as default 18:11:19 thozza: He said something on the lines of "I'll see what I can do and revert back". 18:11:25 it helps to make other to prioritize the work 18:11:44 thozza: It was ajax who did the actual talking. :) 18:11:48 jwb: and this counts towards python3 becoming more mainstream, for example ansible 18:12:06 yes, i've heard all this. i disagree. 18:12:28 Proposal: rename the Change Proposal to: Python 3 is the only guaranteed python interpreter on default installs 18:12:39 sgallagh: +1 18:12:58 * ajax tempted to move the goalposts by adding fedora-packager to the workstation dvd kickstart... 18:13:02 (no not really) 18:13:30 ajax, it's certainly a possibility though 18:13:51 just for a sake of your argument? or whats the point of adding fedora-packager? 18:14:07 it negates sgallagh's proposal 18:14:11 because it requires python2 18:14:24 rkuska: the workstation is the developer experience. i would want that to include the ability to develop _fedora_. which means i need to be able to invoke koji(1). 18:14:28 we work on this change for a long time, and there was not a fedora-packager in the kickstarter 18:14:36 i'm aware 18:14:45 Python3 is already preffered by the packaging guideliness\ 18:14:58 which is excellent 18:15:01 so this is more of a technical and marketing change as I see it 18:15:01 i am, how you say, trolling. 18:15:11 yes, but we need it to be default also for older packages, the guideliness refers only to new packages 18:15:18 ajax: But you'd still have that ability, albeit not on the default install, but is that really that important? 18:15:46 This conversation is troublesome enough without trolling... 18:15:53 by the same logic, I could say, we can add python2 to the kickstart just to kill the goals, right? 18:15:55 mstuchli: define "that important". if we can't build python3 from a fedora toolchain that relies on no other python, then i think we haven't actually finished bootstrapping. 18:16:04 mstuchli: I think it is more important that you can't build Fedora without python2, if we are still talking about "default". 18:16:11 jwb: What change would you like to see to this proposal to allow this work to land in F23? 18:16:12 and i would expect "bootstraps" to be implied by "default" 18:16:20 but, not my feature 18:16:32 Well in that case, just theoretically, would you agree with renaming the change? 18:16:44 Because there is *real value* in making this effort "blessed" and *supported* by FESCo 18:17:03 sgallagh: exactly my point 18:17:09 sgallagh: +1 18:17:09 I agree, but this is not "default" 18:17:12 mstuchli: i wouldn't object to renaming it. for that matter i wouldn't object to _not_ renaming it. i'm primarily concerned with the work being done, and it does look like good work. 18:17:16 sgallagh, i'd prefer the scope included the tools required to build Fedora. if that isn't possible for F23, i'd suggest deferring it until it is possible. 18:17:30 ajax: I see :) 18:17:34 sgallagh, barring any of that, the bare minimum change sould be s/default/perferred 18:17:38 if y'all want to call it default and define default to mean something less rigorous than my personal sense of taste, well, okay 18:17:43 i'm disappointed if so, but 18:17:55 er, preferred 18:17:58 jwb: That's a self-sustaining problem though. If we (FESCo) aren't pushing for it to happen, then it won't happen 18:17:59 jwb: why not to have change of its own for the infra tools? 18:18:05 Something needs to upset the status quo 18:18:11 thozza: +1 18:18:13 sgallagh, anaconda didn't happen because of something fesco did 18:18:14 thozza: +1 18:18:22 jwb: I think it did 18:18:24 thozza: +1 18:18:29 sgallagh, the REST OF THE PYTHON WORLD that needs work isn't going to magically fucking happen because we bless a wiki page 18:18:33 at least it helped to prioritize the effort there 18:18:46 we plan to do this, we are not stopping with this change, we plan to port also fedora-packager tools, but this is for now too much work 18:19:04 jwb: I think you might be surprised at how much of the Python world follows the examples of Fedora/RH and Ubuntu 18:19:14 sigh 18:19:34 jwb: I don't think it is fair to ask them to do it in the change a make it a hard requirement while fedora RCM does not seem to be interested 18:19:42 because Py3 is not default 18:19:45 ;) 18:19:52 or I don't know why 18:20:00 more because they're busy with other work 18:20:05 right 18:20:06 which is legitimate 18:20:08 like everyone 18:20:10 I get the feeling that nobody is really interested in pursuing this with the Fedora infra' people. 18:20:12 :/ 18:20:15 i don't think it's fair to portray a reality that doesn't exist 18:20:37 rishi`: We most certainly are, it's just that we've focust on other things so far 18:20:37 And none of that other work will ever get prioritized without someone (i.e. us) pushing for it 18:20:55 jwb: we are not stopping with only with having pytohn3 by default on live cd 18:21:06 rkuska, i'm aware and i think that is absolutely excellent 18:21:22 mstuchli: I already said (twice), that this was brought up with them, and it was not totally blown away. I am interested in knowing what they have to say. 18:21:23 jwb: so the only problem is calling it "default"? 18:21:33 YES 18:21:35 jwb: do you really think that someone will care to spend time porting to python3 if there is no real need? 18:21:47 thozza, YES 18:21:52 jwb: dream on 18:22:01 before continuing a neverending discussion, do we agree that it's ok to ship python3 as the only python interpreter on default install? 18:22:09 I'm not sure I'm really supposed to (or allowed to?) talk in these meetings but it was just brought up that something was raised with Fedora Infra via me two weeks ago and I'm not sure that's entirely accurate or that I'm the best person to even be dicsussing that with since I'm not on the Fedora Infra team 18:22:09 jwb: there are many upstream who dont, take samba for example 18:22:19 number80: s/it's ok/it's mandatory/ ? 18:22:35 guess what. you're repeating yourselves. 18:22:38 sgallagh: if we agree, it'll be mandatory :) 18:22:59 number80: By which I mean: make a proposal, let's vote 18:23:04 ok, 20 minutes passed 18:23:11 i'm honestly fine with the feature as worded, and calling any future instances of requiring python2 to do distro work as a bug 18:23:18 Somebody make a proposal we can vote on. 18:23:37 proposal: let's make python3 the only python interpreter on default installs 18:23:41 ajax: I like that interpretation 18:23:52 number80: +1 18:23:57 +1 18:24:12 number80: what does this mean? 18:24:21 it means the feature as written, i sincerely hope 18:24:23 does it approve the change as is? 18:24:47 ajax: yes, but I understand jwb opinion (which is valid) as the proposal is not about making python3 default 18:25:05 I don't want to waste time on wording issue, we can rename the proposal to fit reality 18:25:06 +1 then 18:25:13 it is, but I think it is silly to expect it will happen over one release 18:25:19 number80: +1 then 18:25:30 +4 18:25:35 number80, +1 18:25:35 anyone 18:25:37 thozza: sadly, yes and we need incentive 18:25:49 +1 18:25:51 (to push people porting apps to python3) 18:26:14 #agreed let's make python3 the only python interpreter on default installs (+5, 0, -0) 18:26:23 \o/ 18:26:34 dear feature owners, please rename accordingly your proposal 18:26:34 -1 18:26:34 Supplementary proposal: As ajax suggested: keep the "default" terminology and treat python2-only work as a bug in the implementation 18:26:43 please amend the vote record 18:26:44 * thozza hopes he got the keyword correctly 18:26:50 sgallagh: +! 18:26:51 #undo 18:26:51 Removing item from minutes: AGREED by thozza at 18:26:14 : let's make python3 the only python interpreter on default installs (+5, 0, -0) 18:27:02 #agreed let's make python3 the only python interpreter on default installs (+5, 0, -1) 18:27:04 sgallagh, -1 18:27:12 -1 18:27:26 sgallagh: +1 18:27:46 +1 to my supplementary proposal for the record. 18:27:46 sgallagh: -1, that sounds too much like a marketing gimmick 18:28:10 there is already not enough of us to reject nor approve the additional proposal 18:28:18 with given votes 18:28:31 if you don't have enough votes for approval, it is rejected 18:28:45 it doesn't stay in some weird limbo state 18:28:52 jwb: but is there some keyword for it? 18:29:25 proposal: Python 3 as the only Python interpreter on default installs as a step towards Python 3 as default 18:29:35 #info Supplementary proposal to treat python2-only work as a bug in the implementation had insufficient votes (+2, 0, -3) 18:29:46 #undo 18:29:46 Removing item from minutes: INFO by sgallagh at 18:29:35 : Supplementary proposal to treat python2-only work as a bug in the implementation had insufficient votes (+2, 0, -3) 18:29:51 #info Supplementary proposal to treat python2-only work as a bug in the implementation had insufficient votes (+3, 0, -3) 18:29:54 (miscounted) 18:29:56 sgallagh: thanks I was about to submit it 18:30:04 yeah, that works 18:30:17 sgallagh: the previous count was right 18:30:23 no? 18:30:33 oh right, mstuchli doesn't have a vote. Mea culpa 18:30:37 #undo 18:30:37 Removing item from minutes: INFO by sgallagh at 18:29:51 : Supplementary proposal to treat python2-only work as a bug in the implementation had insufficient votes (+3, 0, -3) 18:30:40 #info Supplementary proposal to treat python2-only work as a bug in the implementation had insufficient votes (+2, 0, -3) 18:30:44 * mstuchli is sorry! 18:30:57 let's move on 18:31:00 #topic #1445 F23 Self Contained Changes 18:31:01 .fesco 1445 18:31:02 thozza: #1445 (F23 Self Contained Changes) – FESCo - https://fedorahosted.org/fesco/ticket/1445 18:31:08 do you want to go 1 by 1? 18:31:17 I think it makes sense with the ammount 18:31:19 yes 18:31:19 thozza: Policy is to approve in bulk 18:31:36 Unless there are reservations, then address only those separately 18:31:46 *nods* 18:31:56 * ajax scans the change pages... 18:31:57 I think the netizen spin is the hard one ;) 18:31:58 the io.js thing for instance 18:32:11 netizen and io.js are the big ones. 18:32:20 right 18:32:31 i'm here for the netizen spin 18:32:31 (the node/io.js merge makes it less interesting) 18:32:35 sgallagh: io.js is not there AFAIK 18:32:43 thozza, yes it is 18:32:50 "io.js Technology Preview" 18:32:53 jwb: was not 18:32:58 edited 30 hours ago 18:33:02 * number80 has the trac opened 18:33:08 i'm staring at it right now 18:33:09 it's there 18:33:15 yes 18:33:48 do you want to discuss some particular change more? 18:34:18 thozza: So quick vote to approve the rest? 18:34:29 I think except netizen and io.js changes all others looks good to be approved 18:34:30 Anyone else have any problems with any others? 18:34:37 the same 18:35:02 Proposal: All self-contained changes are approved (except netizen spin and io.js which are discussed separately) 18:35:02 sgallagh: even from mailing list the "others" looks "non-conflict" 18:35:02 those two are my only concern, yeah 18:35:12 thozza: +1 18:35:15 jkurik: ack 18:35:20 +1 18:35:21 thozza, +1 18:35:21 thozza: +1 18:35:28 I'm +1 18:35:36 thozza: I hope the ibus change doesn't break GNOME's input sources stuff. 18:35:57 rishi`: it's a different proposal 18:35:58 rishi`: link? i don't know what "input sources" means here 18:35:59 rishi`, it will not, its just default input method change 18:36:12 ajax: keyboard layouts and input methods 18:36:39 rishi`: can you elaborate on some example scenario? 18:36:42 how it can break? 18:36:48 ibus is ticket #1458, we're discussing #1445 18:36:51 paragan: Well, that is what was said when the local resolver change was passed. Then we saw that networking broke. 18:36:57 we can add a note to cooperate with GNOME folks ;) 18:37:21 rishi`: nothing broke 18:37:28 only my hope 18:37:32 rishi`, I don't think it will break 18:37:38 for some particular desktop 18:37:40 rishi`: it seems like it'd be a functional change if you're used to ibus-chewing, but otherwise... 18:37:42 number80: I thought we were passing everything except io.js and netizen. 18:37:55 rishi`: it is already approved 18:37:56 ;) 18:37:57 rishi`: nope, #1445 is a bundle of proposal 18:38:07 thozza: I really don't want to start arguing over the definition of "broke". 18:38:12 OK, so... Netizen? 18:38:20 rishi`: we did for default ;) 18:38:38 #agreed All self-contained changes are approved (except netizen spin and io.js which are discussed separately) (+5, 0, -1) 18:38:59 so the netizen 18:39:04 number80: Eh? 18:39:12 cleong: I'm not sure I'm in favor of creating a Netizen spin that is only a set of packages. That would be better solved by tools like Software and the KDE equivalent (whose name escapes me) 18:39:27 rishi`: https://fedorahosted.org/fesco/ticket/1445#comment:8 18:39:32 I think what was the issue before didn't change.... like that it is bunch of preinstalled packages, but no additional integration 18:39:40 at least from what I read 18:39:42 It would be more interesting as a complete, configured live system 18:39:43 sgallagh: Apper 18:40:12 didn't see comment:8, addressed the detailed description 18:40:13 number80: The first item there is ibus-libzhuyin 18:40:22 Or better still, as a set of tools to convert a Workstation or KDE system into a Netizen system 18:40:33 number80: ok 18:40:34 rishi`: oups, we have another ibus tickt 18:40:40 (sorry) 18:40:49 sgallagh: maybe an additional option in the installation menu 18:40:50 the packages are the first arrangement or organization 18:41:22 thozza: That's a big effort and one better suited to something like the Workstation WG to push for 18:41:41 cleong: To me, that's not justification for a spin (with all the effort that entails) 18:41:52 cleong: are you in workstation WG? 18:42:08 no i am not in workstation wg 18:42:18 sgallagh: then nobody will push for it in Workstation WG 18:42:26 the benefits and description sets the justification imo 18:42:38 cleong: you don't need to be formally part of a WG to participate in their sessions 18:42:57 ok thanks for that 18:42:58 thozza: As number80 says; cleong can bring it up with them 18:43:22 cleong: are you ok with such outcome? 18:43:25 cleong: I'm not denying that a Netizen configuration has merit. I'd like to see it. 18:43:26 the target market are specific to public charities, gov, and other ngo's 18:43:41 I don't think a Spin is the correct solution without configuration 18:44:27 And I'd *much* rather see it as part of the Workstation or KDE SIG efforts, because they have the knowledge and resources to do it right 18:44:31 cleong: my last question... ? 18:44:36 i plan is to present the spin at conferences and demoing how to configure from the defaults 18:44:49 what type of outcome? 18:45:22 I mean that we don't agree on the spin, but rather on option to install the whole suite as part of installation 18:45:38 having own option in anaconda 18:45:52 for Workstation product e.g. 18:46:02 ahh ok i would be fine with that 18:46:15 great 18:46:17 Note: we're not asserting that they will do this. 18:46:27 You would still have to pitch it to those groups 18:46:35 as long as i can resubmit for a spin in the future 18:46:47 ok 18:46:55 Proposal: Netizen is not approved as spin. We approve the option to have netizen as optional suite in anaconda. Please work with Workstation WG. 18:47:05 thozza: +1 18:47:12 cleong: You can always resubmit it, but I think a better solution would be for this to become an official piece of the Workstation offering 18:47:13 I'm +1 18:47:31 thozza, +1 18:47:34 cleong: this is encouragement to keep refining that effort withing the Workstation WG 18:47:34 thozza: +1 18:47:37 ok i'll join and research the workstation wg 18:47:59 thank you for the advisement and time 18:48:03 jwb: are you voting? 18:48:06 +1 18:48:15 +1 18:48:37 +1 18:48:50 #agreed Netizen is not approved as spin. We approve the option to have netizen as optional suite in anaconda. Please work with Workstation WG. (+7, 0, -0) 18:48:58 so the io.js now 18:49:06 will do 18:49:13 cleong: thanks 18:49:23 the io.js thing will be at best a short-lived effort 18:49:28 cleong: we really appreciate your effort 18:50:01 number80: right, if someone wants to do it, let them do it 18:50:09 as long as it does not break things for others 18:50:40 I'm not sure if it needs a feature, but of course, anyone is free to submit it 18:50:56 I wonder what other distributions are doing about this. 18:51:13 Proposal: io.js is approved. 18:51:51 Counter-proposal: If it's merging back anyway, maybe just build the nodejs package from the io.js sources? 18:51:53 It sounds like LO vs Apache OO from the outside. 18:52:02 seems somewhat superfluous 18:52:05 sgallagh: that's what will happen upstream 18:52:39 sgallagh: Yes, that sounds like a more sane idea to me. Although, I have very little clue about the node.js world. 18:53:03 /me wishes he had thought of this more than 30s ago and had mentioned it on the list... 18:53:08 Could they build it in COPR? 18:53:23 the more i look at this the more this looks like a simple alternatives setup 18:53:28 thozza: The main advantage is that it runs on more architectures than current node 18:53:38 So there's value to the secondary arches in having it in Fedora proper 18:53:40 sgallagh: right, that is good 18:53:49 if there's someone driving the io.js part then great, they get the bugs when it breaks, and if it's too broken we stop trusting their acumen in the future 18:54:00 thozza: +1 approve 18:54:10 I am not sure if it requires packaging guidelines update 18:54:11 ajax: The io.js submitter is the node.js primary maintainer 18:54:19 even better 18:54:25 the domain expert is on it 18:54:37 /me nods 18:54:46 module package names start with iojs-* 18:55:13 I'm wary of the iojs-* solution 18:55:22 so any more votes/proposals? 18:55:25 Because (as noted above) this is merging back into the next version of node.js 18:55:40 It seems foolish to bring in a whole bunch of parallel packages that are going to merge back 18:55:54 I think its better to have it as a copr repo only 18:56:13 I'd rather it just replace node.js entirely, if it's the 100% replacement claimed 18:56:14 paragan: secondary arches again 18:56:29 How about we vote for sgallagh 's idea? 18:56:34 Or at least alternatives and it can host the existing nodejs-* packages 18:56:55 rishi`: I'd rather not vote on that without the maintainer involved, honestly. 18:56:59 sgallagh: the same owner made the node.js 0.12 proposal, so I think he pretty much has a gist of the path to move toward the merged back io.js 18:57:02 Can we defer a week and I'll raise it on th elist? 18:57:05 Proposal: Let the owner consider replacing node.js completely and revisit next week. 18:57:15 thozza: +1 18:57:17 sgallagh: That is true. 18:57:31 thozza: +1 18:57:36 thozza, +1 18:57:41 thozza: +1 18:57:55 I'm +1 ;) 18:58:07 jwb: ajax? 18:58:08 (note this delay also the node.js 0.12 proposal vote) 18:58:33 number80: it is not in the agenda 18:59:18 sure, defer, why not 18:59:19 +1 18:59:28 #agreed io.js change: Let the owner consider replacing node.js completely and revisit next week. (+6, 0, -0) 18:59:43 yeah, +1 18:59:50 #undo 18:59:50 Removing item from minutes: AGREED by thozza at 18:59:28 : io.js change: Let the owner consider replacing node.js completely and revisit next week. (+6, 0, -0) 19:00:01 #agreed io.js change: Let the owner consider replacing node.js completely and revisit next week. (+7, 0, -0) 19:00:05 thozza: it was marked as a meeting item this afternoon (if it isn't anymore, my apologies) 19:00:19 number80: too late ;) 19:00:21 #topic #1451 F23 System Wide Change: SELinux policy store migration 19:00:21 .fesco 1451 19:00:22 thozza: #1451 (F23 System Wide Change: SELinux policy store migration) – FESCo - https://fedorahosted.org/fesco/ticket/1451 19:01:13 mgrepl: is around to answer questions 19:01:22 sure, together with plautrba 19:01:41 generally, we need to stay in touch with the upstream, the change also incorporates some performance enhancements 19:01:58