18:00:05 #startmeeting FESCO (2015-03-18) 18:00:05 Meeting started Wed Mar 18 18:00:05 2015 UTC. The chair is nirik. Information about MeetBot at http://wiki.debian.org/MeetBot. 18:00:05 Useful Commands: #action #agreed #halp #info #idea #link #topic. 18:00:06 #meetingname fesco 18:00:06 #chair ajax dgilmore jwb mitr nirik paragan rishi thozza sgallagh 18:00:06 #topic init process 18:00:06 The meeting name has been set to 'fesco' 18:00:06 Current chairs: ajax dgilmore jwb mitr nirik paragan rishi sgallagh thozza 18:00:11 Hello 18:00:16 Hi 18:00:37 Hello 18:00:49 i am here-ish 18:01:07 * jreznik will try to be around if needed by FESCo :) 18:01:12 short meeting today hopefully. ;) 18:01:18 hi all 18:01:26 * decause waves 18:01:46 ok, lets go ahead and get started. 18:01:53 hello decause. ;) 18:01:56 #topic #1419 provenpackager request (FAS - nonamedotc) 18:01:56 https://fedorahosted.org/fesco/ticket/1419 18:01:56 .fesco 1419 18:01:58 nirik: #1419 (provenpackager request (FAS - nonamedotc)) – FESCo - https://fedorahosted.org/fesco/ticket/1419 18:01:58 nirik: hi there :) 18:02:10 so, by our policy this didn't get 3 +1s in a week so it goes to meeting.... 18:02:18 sgallagh voted +1 in ticket 18:02:32 * mattdm is lurking 18:03:09 any other votes? I'm +1 (as I was in the ticket) 18:03:39 +1 18:03:53 +1, considering the variety of different kinds of packages 18:04:05 sure, I'm +1 18:04:51 #agreed request is approved (+5, 0, 0) 18:05:05 ok, do we want to discuss the dnf vs yum ticket? or leave it for next week? 18:05:43 I would leave the discussion in the ticket for now 18:05:57 I'm not sure if there is anything to discuss for F22 18:06:13 we already agreed we don't want dnf-yum to be installed by default in F22 18:06:35 ok, we can bring up next week if there's proposals in ticket. 18:06:37 remind me which ticket that is? 18:06:49 https://fedorahosted.org/fesco/ticket/1312 18:06:54 .fesco 1312 18:06:55 nirik: #1312 (F22 System Wide Change: Replace Yum With DNF - http://fedoraproject.org/wiki/Changes/ReplaceYumWithDNF) – FESCo - https://fedorahosted.org/fesco/ticket/1312 18:07:24 ahh, there are new comments I didn't see 18:07:33 yes there are :) 18:07:35 lets touch on it then I suppose... 18:07:40 #topic #1312: F22 System Wide Change: Replace Yum With DNF - http://fedoraproject.org/wiki/Changes/ReplaceYumWithDNF 18:07:41 https://fedorahosted.org/fesco/ticket/1312 18:07:41 .fesco 1312 18:07:42 nirik: #1312 (F22 System Wide Change: Replace Yum With DNF - http://fedoraproject.org/wiki/Changes/ReplaceYumWithDNF) – FESCo - https://fedorahosted.org/fesco/ticket/1312 18:07:48 nirik: If we do go with the dnf-yum-in-/usr/bin/yum rote, we should do so very soon. But that proposal has been only 5 hours ago, so getting all information an everybody involved today is probably impossible. 18:08:02 mitr: yeah. 18:08:28 yes if we need to propose any such changes good to implement it soon 18:08:47 you'd think PackageKit-command-not-found would cover this already 18:08:50 in particular I'd like to hear what the yum folks have to say... 18:09:00 I’d say: Make sure jzeleny, dgilmore, and whoever else we can think of, is Cc:ed, ask for specific feedback on this plan, and aim to get a decision in the ticket before the next meeting. 18:09:13 fwiw I'm in favor of kevin/nirik's "middle ground" suggestion 18:09:49 * paragan also likes nirik's suggestion in ticket 18:10:34 I have already asked dnf developers to CC themselves in that ticket, it was having Ales as CC 18:11:07 if someone could point the packaging-team (who I think 'owns' yum ?) at this also that might be good... 18:11:42 nirik, I already asked them 18:11:50 Hum, /usr/lib/python2.7/site-packages/mockbuild/package_manager.py actually invokes the yum _command_ 18:12:05 paragan: Just edit the Cc:s directly 18:12:08 #info please add stakeholders to ticket and see if we can gather consensus before next meeting 18:12:34 Both rholy and jsilhan are in CC in that ticket 18:12:35 mitr: sure, some things call /usr/bin/yum, some use a yum object... 18:12:49 and mock has been ported. :) 18:13:05 it's things like mash, pungi, etc... that still need work. 18:13:10 i'm confused why this is just now coming up 18:13:29 jwb: it did before 18:13:43 jwb: It's coming up again because people started installing the Alpha 18:13:46 jwb: someone filed a bug on the cloud image... 18:13:55 we discussed this ticket like 3 times. how did anyone fail to raise these concerns in a manner that made us notice and expect dnf-yum to be the default? 18:14:05 sorry i am late 18:14:39 jwb: also, the change page still says dnf-yum will be installed, I asked them to change that in the change bug... or clarify... 18:14:46 but I guess I should just edit the wiki. 18:14:53 jwb: IIRC the concerns have been raised but without a proposal we could either go with no dnf-yum or delay the dnf switch. 18:15:10 jwb: Now we have a proposal, which, if it were practical, would IMHO be so much better that reopening this is well worth it. 18:15:12 nirik: I can poke jsilhan tomorrow, he sits a few meters away from me 18:15:28 cool 18:15:55 I'm fine with no dnf-yum, but it seems others dont like that. 18:16:12 I think it makes sense to have it there 18:16:21 the amount of waffling and back and forth is disappointing 18:16:37 so let's figure it out and stop doing that 18:16:57 well, I don't think I have been waffling... I just tried to come up with something that would make more people happy. 18:17:16 jwb: the only way to stop the discussion for good is to say the dnf is ready and they need to follow the original plan and rename to yum 18:17:23 nirik, none of this information is new at all. we said we'd switch to DNF. we knew it wasn't 1:1 18:17:41 * rishi drops in & reads back log 18:17:51 jwb: exactly 18:17:56 jwb: yeah. Agreed. 18:17:58 now we're worried about it not being called yum. it wasn't called yum the first 3 times we talked about it 18:18:08 there is so much investement in Fedora, RHEL, CentOS etc that changing the package installer/updater's name is never going to satisfy a large number of people 18:18:16 jwb: I think there's some confusion tho about dnf-yum... 18:18:19 I don't think I'm waffling. And the original proposal said that it *might* be named yum. 18:18:52 mattdm: the original plan years ago was for the re-write to be named back to yum when ready. 18:19:02 Unfortunately I got to run home today. If we are going to vote on the dnf vs. yum issue, then you can assume that I am with the majority (whatever that is). 18:19:07 the current approved change still says that dnf-yum will be used. 18:19:12 the only way to not have a huge docs, mindshare etc cost is to follow that through 18:19:25 "package 'dnf-yum' is installed by default. It obsoletes Yum and provides its own /usr/bin/yum, a short script that redirects to /usr/bin/dnf with an appropriate warning message that DNF is the preferred package manager now. Notice that upgrading F21 to F22 will not cause the compat package to be installed so will not disturb any upgrading users." 18:19:33 jwb: Upstream doesn't really want it to be called yum, but several FESCo members do. It is not that surprising that DNF upstream is not spending huge amount of time on making /usr/bin/yum a possibility, nor that FESCo members take time to figure alternatives out. 18:19:59 or s/called/available as/ 18:20:04 dgilmore: some of us remember up2date... 18:20:11 ajax: :) 18:20:21 anyhow, I am +1 to no dnf-yum, I am +1 to a dnf-yum that doesn't conflict with or obsolete yum, and -1 to one that does. ;) 18:20:25 ajax: it had a huge cost and there was less investment in it 18:21:44 So, assuming we are not deciding this now, who all needs to be on Cc:? 18:21:44 ok, so what do we want to do today? 18:22:28 nirik: nothing I guess :) 18:22:53 mitr: In that case, I would lean towards upstream's judgement and never install dnf-yum. 18:23:06 "never" == "even post-F22" 18:23:51 rishi: we can never say never 18:23:55 rishi: I think I see their viewpoint but ultimately I disagree; the breakage is just not worth it (both alone, and as a precedent/reinforcing the way we do migrations) 18:24:23 mitr: You disagree with what? Got confused. 18:25:59 ok, lets continue in ticket I guess and see what comes up next week? 18:26:15 yeah 18:26:17 ack 18:26:30 #info continue in ticket with proposals, etc. 18:26:33 #topic Next weeks chair 18:26:36 who wants it? 18:27:20 I can next week 18:27:27 thanks paragan 18:27:32 #info paragan to chair next week 18:27:36 #topic Open Floor 18:27:40 anything for open floor? 18:28:13 i'd like to revisit the hardening change for F23 18:28:30 * decause raises hand for next in queue 18:28:37 i should really have been aware that it implied -z now, which is a rather drastic change 18:28:40 ajax: ok, you want to discuss that now? or next week? 18:29:05 nirik: just raising awareness really. i'd like to get a concrete set of changes by next week 18:29:10 which i'm happy to work on 18:29:25 just want to see if that raises any red flags for anyone 18:29:39 * nirik is ok with adjusting things if we have good reason to do so. 18:29:57 this isn't exactly apples-to-apples, but mockbuild of xorg-x11-proto-devel takes significantly longer against f23 than f22 18:30:02 ajax: Is there a writeup anywhere of the cases that are broken by this? ISTM that this would break mostly wrong/cyclical/unusual dependency setups (which is not to say that we would be justified in ignoring them) 18:30:30 since the build process for that package is almost entirely the time spent executing ./configure 18:30:40 ajax: that might also be gcc changes? one of the hardening things mentioned was gcc would be slower... 18:30:45 ah, nevermind then 18:30:47 it implies that we're doing a _lot_ more work on every exec, which is consistent with -z now forcing all the bindings up front 18:31:18 so what i'd like is basically _just_ defaulting executables to PIE 18:31:36 it wouldn't give you "full relro" as checksec.sh has it, but 18:31:55 mitr: not as such, but i can come up with a couple. 18:32:07 ajax: would we be better off moving some flags from the hardened build into the standard cflags 18:32:09 I should also look into at least the guile issue 18:32:16 and still having the hardened option? 18:32:35 I am reminded of: http://harald.hoyer.xyz/2015/03/05/libtool-getting-rid-of-180000-sed-forks/ 18:33:11 dgilmore: basically yeah 18:33:22 ajax: I would be good with doing that 18:33:31 anyway. i'll write it up in the ticket. 18:33:49 hopefully with patch in hand. 18:33:58 * ajax passes the mic 18:34:06 decause: you had something? 18:34:25 nirik: yep 18:35:44 fire when ready. ;) 18:35:50 There is this new thing called FOSCo that is going to be spinning up over the next couple of weeks 18:36:13 in anticipation for it, we've got some infrastructure set up that I wanted to get onto your collective radars 18:36:33 there is now #fedora-outreach on freenode 18:36:47 and we just got our trac instance up at http://fedorahosted.org/fosco 18:37:18 these are the places where we'll be gathering input about (firstly) the structure of FOSCo, and then the goals/strategies/actions for the group 18:37:24 #info FOSCo is starting up over the next few weeks. IRC: #fedora-outreach Trac/ticketing: https://fedorahosted.org/fosco 18:37:49 cool. ;) 18:37:53 I'm still finding my way around, but thanks to threebean and lmacken I think I'm starting to tread water 18:38:08 Anything in particular you need from FESCo at this point? or just letting us know about things... 18:39:23 nirik: mostly I'll be spending my time in the next couple of weeks listening and reading, and I would like to schedule some kind of call with FESCo folks in the near future 18:39:54 * decause has a lot of history to grok 18:40:13 decause: ok. We are a pretty widely spread group to manage a call, but you can find most all of us on irc at various times of the day. :) 18:40:20 I'll be posting something more detailed to the mail lists in the near future, but wanted to drop by today also 18:40:33 excellent. :) Looking forward to it. 18:40:34 s/call/irc-meeting 18:40:45 we're here every week :) 18:40:48 :) 18:41:08 ok, anything else decause ? or anyone else for open floor? 18:41:31 nirik: nope :) 18:42:09 ok, will close out the meeting in a minute if nothing else. 18:43:01 Thanks for coming everyone! 18:43:04 #endmeeting