18:01:05 #startmeeting FESCO (2012-11-14) 18:01:05 Meeting started Wed Nov 14 18:01:05 2012 UTC. The chair is mmaslano. Information about MeetBot at http://wiki.debian.org/MeetBot. 18:01:05 Useful Commands: #action #agreed #halp #info #idea #link #topic. 18:01:10 Hi 18:01:10 #meetingname fesco 18:01:10 The meeting name has been set to 'fesco' 18:01:16 hello 18:01:19 morning 18:01:19 #chair notting nirik mjg59 mmaslano t8m pjones mitr limburgher jwb 18:01:20 Current chairs: jwb limburgher mitr mjg59 mmaslano nirik notting pjones t8m 18:01:23 #topic init process 18:01:33 morning. 18:01:51 * limburgher here 18:02:02 Hello 18:02:28 * notting is here 18:02:37 * jreznik is lurking 18:02:47 let' start 18:03:10 .fesco #963 18:03:10 mmaslano: Error: '#963' is not a valid integer. 18:03:20 #topic #963 change of names of configuration files 18:03:30 .fesco 963 18:03:30 .fesco 963 18:03:31 pjones: #963 (change of names of configuration files) – FESCo - https://fedorahosted.org/fesco/ticket/963 18:03:34 mmaslano: #963 (change of names of configuration files) – FESCo - https://fedorahosted.org/fesco/ticket/963 18:03:42 mjg59: could you help michal or Kay? 18:03:45 Yeah 18:03:52 Sorry, hadn't really been online much for the past week 18:03:58 I'll follow up on that 18:03:58 sure, thanks 18:04:21 They could just unpack/analyze rawhide themselves I'd expect.. 18:05:50 any proposal? I would be fine with mjg's help 18:06:07 if mjg59 wants to help, I'm completely fine 18:06:11 proposal: defer for now and see if mjg59 can get them access to the info they seek 18:06:13 but it should be done before freeze if there are broken packages 18:06:43 mmaslano: We already froze for Beta 18:06:50 yes 18:07:05 well, packages have been 'broken' inasmuch as there were two configuration files for two releases already. what is the plan if the work mjg59 is helping kay with does not happen? 18:07:20 if there are pending fixes we could pass them thru the freeze as a NTH... or wait until after. 18:07:41 As the breakage is not critical I think that the fixes can be done in updates 18:08:03 so for now I agree with nirik's proposal 18:08:22 notting: We are not currently equipped to handle that question AFAICT. (We do need to be, but that's for another time.) 18:08:57 * nirik shrugs. I don't think I would block on it. 18:08:58 nirik, +1 18:09:09 it would be very nice ot fix before release tho... 18:10:20 Sure, where possible that clearly ought to happen 18:10:39 more votes on nirik's proposal? 18:10:41 +1 nirik 18:10:56 +1 I think 18:11:40 * nirik is for his own proposal too oddly enough. 18:11:46 * mmaslano too 18:11:56 +1 for lack of an alternative 18:12:01 +1 I guess. 18:12:09 sure 18:12:34 #agreed defer for now and see if mjg59 can get them access to the info they seek (+8,-0) 18:12:34 I mean, that's the process anyway if we /don't/ vote on it.... 18:12:55 #topic #971 Freezing for Fedora 18 Beta 18:13:00 .fesco 971 18:13:02 mmaslano: #971 (Freezing for Fedora 18 Beta) – FESCo - https://fedorahosted.org/fesco/ticket/971 18:13:13 I took that to be 'no great objection to freeze' 18:13:19 we're already frozen 18:13:25 so, close this, freeze happened. ;) 18:13:31 great 18:13:40 jreznik, who is "RH's SB manager" 18:13:52 #info Fedora 18 was already frozen 18:14:18 jwb: in as much as 'guy running the meetings', jack 18:14:18 jwb: the person responsible for SB as far as I know 18:14:33 jreznik, i'm looking for a name, not a role definition 18:14:58 * nirik wonders what SB expands to.... oh, secure boot? 18:15:06 nirik: yep 18:15:07 jreznik: he wants to know who you were referring to 18:15:21 pjones: notting answered 18:15:39 okay 18:15:46 that is the most obtuse way of answering my question, ever. 18:15:53 whatever. move on 18:16:16 #topic #969 libexecdir guideline conflicts with extant packages 18:16:24 .fesco 969 18:16:25 so, I have some info here. 18:16:26 mmaslano: #969 (libexecdir guideline conflicts with extant packages) – FESCo - https://fedorahosted.org/fesco/ticket/969 18:16:37 FPC is currently discussing a very similar case of this related to java. 18:16:55 it's not been decided either way, but it's very similar to this issue. 18:17:15 nirik: Something different than the multilib exception for java that we voted on recently? 18:17:19 so, I propose we defer this, ask FPC to consider it with java and wait for them to rule on that before we do anything. 18:17:59 mitr: it's related. The question then becomes can non multilib things just use lib (like java and systemd do) 18:18:24 (at least my quick understanding of it. Didn't read the full FPC meeting) 18:18:59 if they want to tackle it first, i'm ok with it. w.r.t. the ticket, i'd be proposing 1) and 2) listed there anyway 18:19:42 see also: http://lists.fedoraproject.org/pipermail/packaging/2012-November/008757.html 18:20:53 I wasn't quite clear from the ticket 18:20:58 So at this point I would suggest 18:21:00 proposal: formally vote to confirm that As a general matter, Fedora policies trump upstream intentions and preferences (although they can be considered). Rationale: the only way to make Fedora a consistent distribution. 18:21:02 And I'd support a deferral of the ticket for a specific resolution re:systemd (unfortunately I haven't had enough time to form an opinion) 18:21:03 Is it /usr/lib, or is it /usr/%(libdir)? 18:21:26 mjg59, /usr/lib 18:21:32 it's %{prefix}/lib/ on all arches. 18:21:35 mjg59: the complaint is about x86_64 packages using /usr/lib instead of /usr/%(libdir) or /usr/libexec 18:21:50 notting: Ok from the ticket it sounded like it was primarily about libexec 18:22:15 * gholms is here 18:22:28 s/can be considered/can and should be considered/ above 18:22:39 gholms, why FPC rejected to allow /usr/lib for non-multilib packages in the systemd case? 18:22:43 mitr: I'm -1 to that as written 18:23:06 mitr: Or, rather, not as written. I agree with the statement, but I don't know that it's the appropriate statement to make in this case 18:23:30 mitr: Where our policies are incompatible with the behaviour of many upstreams, we should reconsider whether our policies are actually helping or harming 18:23:37 well, I agree with it, but it's kinda a given isn't it? 18:23:38 t8m: I wasn't involved in that discussion, but I can try to find out if necessary. 18:23:58 we shouldn't recklessly trump upstream. Only in cases we really must for good reasons. 18:24:04 mjg59: I don't know how that applies to systemd in this case. But we have systemd maintainers thinking otherwise, and I don't want the opposite to become a widely accepted oinion. 18:24:05 t8m: The decision itself is here, along with some justifiction in the comentary: https://fedorahosted.org/fpc/ticket/158 18:25:40 mjg59: https://fedorahosted.org/fpc/ticket/158#comment:6 sums the situation up. 18:25:41 * nirik nods. based on the info at the time... 18:25:50 anyhow, any votes on my proposal? ;) 18:26:17 Ok. What's the benefit of requiring libdir rather than /usr/lib for this case? 18:26:26 nirik: +1 to deferral. 18:26:58 nirik: That's for deferring until fpc decides upon the java stuff, right? 18:27:03 mjg59: consistency I guess? 18:27:12 mjg59: e.g. ability to run an automated test for misplaced 64-bit binaries 18:27:19 yeah, I'm +1 to deferral until the java ruling comes down. 18:27:19 gholms: yes, and ask them to think about this too at the same time. 18:27:25 +1 to deferral 18:27:27 * notting is +1 for deferral 18:27:39 +1 18:27:49 +1 18:27:49 I'm +1 to deferral, but the rationale presented for the policy so far seems weak 18:28:13 DEFER 18:29:11 #agreed defer the decision (until Java rulling comes down) (+9,-0) 18:29:13 Are there any cases where you could simultaneously have a 32-bit version of an executable binary in /lib and a 64-bit version in /lib64 18:29:16 ? 18:29:37 mjg59: yes 18:29:43 mjg59: helpers executed by a library, yes 18:29:46 sure 18:30:01 * nirik looks. systemd is multiarch? 18:30:14 ah, but uses -libs. ok. 18:30:24 nirik: not the things in /lib/systemd, no 18:30:36 So the objection to using /usr/lib is just that a package might accidentally drop a 64-bit binary there despite supporting simultaneous 32 and 64-bit installation? 18:30:59 And forcing the use of libdir means that we can automatically check for that? 18:32:38 kind of, yes. if you're putting 64-bit binaries in /usr/lib, you need to 'know your package is never intended to be installed as multilib and will not accidentally be installed that way' 18:32:42 Ok 18:32:46 which is hard to clearly check against the guideline 18:33:04 also lsb says: http://www.linuxbase.org/betaspecs/fhs/fhs.html#usrlibexec 18:33:18 you just killed a cat. 18:33:21 nirik: Yeah but the LSB is failing to describe reality in this case 18:33:24 yeah, true. 18:33:35 anyhow, I'll stop now. 18:33:48 we have an agreed and are wasting time. ;) 18:33:50 next topic? 18:33:53 notting: It seems like we could solve that easily enough by just permitting packages to ship a list of "I really meant to do this" 18:34:06 (or just drop 32bit. ;) 18:34:09 mmaslano, please 18:34:14 #topic #970 F19 Feature: RPM 4.11 http://fedoraproject.org/wiki/Features/RPM4.11 18:34:17 nirik, or multilib 18:34:17 It's unclear to me that this policy buys us anything other than divergence from upstreams 18:34:23 .fesco 970 18:34:23 Anyway 18:34:25 mmaslano: #970 (F19 Feature: RPM 4.11 - http://fedoraproject.org/wiki/Features/RPM4.11) – FESCo - https://fedorahosted.org/fesco/ticket/970 18:34:26 +1 to rpm 18:34:30 +1! 18:34:35 +1 make it so. 18:34:36 +1 18:34:36 +1 18:34:37 +1 18:34:40 big +1 18:34:41 I'd love more f19 features now. ;) 18:34:45 We'll have the 411! 18:34:52 sure 18:34:53 +1 18:34:55 +1 sure 18:35:15 #agreed rpm was accepted as f19 feature (+9,-0) 18:35:26 that's it 18:35:29 #topic Next week's chair 18:35:52 i will be out next week 18:36:01 I can do it I suppose... 18:36:19 #info nirik will be the chairman next week 18:36:30 #topic Open Floor 18:36:44 It's the day before thanksgiving next week? 18:36:48 Yup. 18:36:52 yep 18:36:56 Not that it matters in my case 18:37:08 I expect many folks will be gone thursday/friday. 18:37:14 I will. 18:37:16 yeah 18:37:17 yeah, I may or may not be here. will comment in tickets in any case 18:37:18 I hope we can get to a beta before next wed. 18:37:33 From your IRC client to $_DEITY's ears. 18:38:13 I'll close meeting in 5 minutes! 18:39:07 just want to thank those who threw their name in the hat for the FESCo elections 18:39:54 And Ankur for wrangling. 18:40:06 * gholms concurs 18:40:23 * nirik nods. 18:44:15 #endmeeting