17:30:01 <nirik> #startmeeting FESCO (2011-03-30)
17:30:01 <zodbot> Meeting started Wed Mar 30 17:30:01 2011 UTC.  The chair is nirik. Information about MeetBot at http://wiki.debian.org/MeetBot.
17:30:01 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
17:30:01 <nirik> #meetingname fesco
17:30:01 <zodbot> The meeting name has been set to 'fesco'
17:30:01 <nirik> #chair mclasen notting nirik SMParrish kylem ajax cwickert mjg59 mmaslano
17:30:01 <nirik> #topic init process
17:30:01 <zodbot> Current chairs: SMParrish ajax cwickert kylem mclasen mjg59 mmaslano nirik notting
17:30:58 * notting is here
17:31:23 * mclasen still in another meeting, but here
17:33:43 * nirik waits a bit for more folks to show up
17:33:58 * jsmith lurks
17:34:29 * gholms|work passes out bagels
17:37:14 <nirik> humm... we may not have quorum today...
17:37:49 <gholms|work> But... I brought bagels!
17:38:03 <gholms|work> :'(
17:38:21 <notting> marcela voted in the tickets, iirc. that might help towards quorum
17:38:41 <nirik> I think we still need one more?
17:39:17 <ajax> sorry i'm late
17:39:23 <nirik> and hey. :) Welcome ajax
17:39:38 <nirik> so, I guess lets go ahead...
17:39:49 <nirik> #topic Meeting Time
17:40:01 <nirik> Do we want to shift the meeting time now that .eu has moved to DST?
17:40:12 <ajax> no opinion
17:40:13 <nirik> I guess we don't have any of our .eu members present to weigh in.
17:40:43 <ajax> unrelated, mjg59's likely to miss today's meeting, he's on pto
17:40:49 <nirik> ok.
17:40:54 <nirik> hey cwickert1
17:41:18 <nirik> cwickert1: we are wondering if we should adjust the meeting time any due to .eu going to DST now. Thoughts?
17:41:38 * nirik prefers to just stick to this time, but I can adjust if others want to move.
17:42:20 <notting> either/or for me
17:42:38 <nirik> I guess I can file a ticket and we can vote in ticket?
17:43:34 <cwickert> nirik: works for be, but ATM I'm busy with dayjob
17:43:48 <nirik> cwickert: ok.
17:44:07 <nirik> #action nirik will file a ticket and collect votes on moving meeting time.
17:44:37 * nirik skips features repo and metrics...
17:45:04 <nirik> #topic #563 suggested policy: all daemons must set RELRO and PIE flags
17:45:04 <nirik> .fesco 563
17:45:06 <zodbot> nirik: #563 (suggested policy: all daemons must set RELRO and PIE flags) - FESCo - Trac - https://fedorahosted.org/fesco/ticket/563
17:45:13 <nirik> I guess this was kylem's...
17:45:18 <nirik> so we should skip it as well.
17:45:29 <nirik> #info will revisit next week.
17:45:32 <nirik> #topic #576 Encourage to be package maintainers to introduce themselves in devel list
17:45:32 <nirik> .fesco 576
17:45:33 <zodbot> nirik: #576 (Encourage to be package maintainers to introduce themselves in devel list) - FESCo - Trac - https://fedorahosted.org/fesco/ticket/576
17:45:51 <nirik> #info mmaslano voted +1 to this in ticket.
17:46:10 <nirik> I'm fine with adding this as a suggestion/etc...
17:46:23 <ajax> i would really like this not to be mandatory
17:46:40 <notting> seems like a reasonable suggestion. why the gpg?
17:46:45 <ajax> i would _prefer_ we not do it at all, at least on the mailing list.  i get enough noise.
17:46:57 <cwickert> +1 from me
17:46:59 <nirik> well, the new packager flow these days is pretty small.
17:47:11 <nirik> Not sure about the sponsor announcing packages...
17:47:30 <notting> while it's noise, it's noise that provides positive feedback and encouragement to new contributors
17:47:33 <cwickert> erm, the +1 was meant for introductions, I agree with mmaslano
17:47:56 <cwickert> how many new maintainers are there a week? 1, 2?
17:47:59 <nirik> so there have been 2 this week
17:48:11 <nirik> previous week had 3
17:48:14 <cwickert> that's not really noise, is it?
17:48:25 <mclasen> sounds like a fine thing to encourage people to do
17:49:06 <ajax> sure, why not
17:49:07 <nirik> so, +1 to me for suggesting/noting this in the new packager flow
17:49:37 <nirik> notting: not sure why gpg there... should we suggest removing that?
17:50:13 <notting> maybe making it an optional part of the optional announcement. although that's trending down the rabbit hole
17:50:30 <notting> +1 to having this as a suggestion, regardless
17:50:42 <nirik> yeah, "you could optionally include your gpg key or other information you feel the community would use to contact you"
17:51:07 <nirik> #agreed proposal is accepted
17:51:42 <nirik> so, how about the sponsors mail when they sponsor someone?
17:51:56 <nirik> I think that could be noise, and seems like it would be duplicating the first bit?
17:51:59 <notting> ... wouldn't that be covered by the new packager in any case?
17:52:08 <nirik> yeah
17:53:02 <nirik> so, that part I am -1 on unless I am missing some reasoning there.
17:53:42 * notting would be -1, for the same reason
17:54:16 <ajax> same, -1
17:55:50 <nirik> more votes? mclasen / cwickert ?
17:57:41 <nirik> I can ask mether if we are missing some rationale on that one.
17:58:09 <nirik> #info will ask for more rationale on the sponsors mailing the list when they sponsor someone.
17:58:23 <nirik> #topic #577 FHS exception for Heimdal
17:58:23 <nirik> .fesco 577
17:58:24 <zodbot> nirik: #577 (FHS exception for Heimdal) - FESCo - Trac - https://fedorahosted.org/fesco/ticket/577
17:58:49 * ktdreyer is here to answer questions if needed
17:59:08 <nirik> #info mmaslano voted -1 in ticket
17:59:14 <nirik> thanks for coming ktdreyer
17:59:38 <nirik> so, the options suggested won't work for some reason? alternatives / libdir ?
18:00:15 <notting> i liked %{_libdir}/heimdall, personally.
18:00:43 <ajax> ktdreyer: do the heimdal libs have the same sonames as the mit kerb libs?
18:01:04 <ktdreyer> ajax: I'm not an expert... I think libkrb5 is the same
18:01:18 <ktdreyer> orion has posted a list of the exact conflicting files in the Review bug
18:01:30 * nirik notes there is still an open merge review on krb5...
18:01:33 * ajax reads
18:01:45 <nirik> so it's not something that should be put forth as a compliant package
18:02:03 <nirik> .bug 225974
18:02:04 <ajax> if they have the same sonames then i'm hesitant to allow heimdal to be packaged at all, at least until rpm figures out how rpaths work
18:02:05 <zodbot> nirik: Bug 225974 Merge Review: krb5 - https://bugzilla.redhat.com/show_bug.cgi?id=225974
18:02:50 <ajax> since, as noted, the library APIs are different
18:03:01 <ajax> (reading the bz)
18:03:21 <ktdreyer> ajax: so what do you recommend?
18:03:42 <tibbs> Could really use nalin's input as well.
18:03:56 <ktdreyer> tibbs: yeah, agreed
18:04:21 <notting> libraries don't appear to conflict (although headers do)
18:04:38 * nirik is -1 to granting any exception at this time. Would suggest trying to work on a technical solution more.
18:05:22 <ajax> man this is ugly
18:06:17 <ajax> well, okay.  if the libs don't have conflicting sonames, then the krb5-config script should be able to point you at whichever devel environment you have installed, and neither ld.so nor rpm shoudl get confused by having both sets of libs installed
18:06:40 <ajax> heimdal-devel and krb5-devel would probably have to conflict and everything else would need alternatives
18:07:19 <notting> or, heimdal-devel could install to a different include path
18:07:35 <ajax> notting: heimdal-devel includes /usr/bin/krb5-config
18:07:49 <ajax> i _guess_ you could alternatives that too, but ick
18:07:56 <ktdreyer> I imagine we would have to rename that one, yeah
18:08:42 <notting> ajax: meh, include a patched one in both
18:08:54 <ajax> ktdreyer: problem si if you do rename it to heimdal-config, you can't rebuild apps against heimdal (to the extent that there's no api problems) without patching the app
18:09:22 <ktdreyer> ajax: what about renaming it to krb5-config.heimdal, and in the .specs that build against Heimdal, set an alias?
18:09:24 <ajax> since they're going to expect krb5-config to give them the header and linker flags
18:09:41 <ktdreyer> alias krb5-config to /usr/bin/krb5-config.heimdal
18:10:16 <ajax> i guess it also depends what we expect to have linking against heimdal
18:10:29 <notting> hm
18:10:46 <notting> you wouldn't really want a common krb5-config, if users need to call *different* APIs depending on which
18:10:57 <ajax> i think just having the two -devels conflict is saner.  it's not like anyone could reasonably want both in the same app.
18:11:26 <ktdreyer> I guess that sounds reasonable to me
18:11:46 <ktdreyer> so make -devel conflict, and use alternatives for everything else?
18:12:14 <nirik> which will also require changes in krb5... have you heard anything from it's maintainer?
18:12:18 <ajax> that's what i'm leaning towards for the moment, but there might be something subtle i haven't hit yet.
18:12:45 <notting> other than 'not have two implementations installed', why would the user care which lib their kerberized telnet (ed note: HA) uses?
18:13:04 <notting> i.e., do we really need two versions of all of those to alternatives-around with?
18:13:10 <ajax> also: for everything else that needs it, which (iiuc) oughtn't include the libs
18:13:30 <ajax> just the apps really
18:13:51 <nirik> notting: I suppose one or the other might have a bug a user wishes to avoid by using the other one?
18:14:00 <nirik> are they otherwise all compatible?
18:14:06 <gholms|work> Their admin line protocols are different.
18:14:13 <gholms|work> (kadmin vs. hadmin)
18:14:17 <nirik> ok. Shows what I know of them. ;)
18:14:32 <notting> gholms|work: those would be different binaries, yes?
18:14:44 <notting> nirik: bad reason. fix the bug! :)
18:14:46 <gholms|work> Should be.
18:14:58 <gholms|work> Not sure about the libs, though.
18:16:27 <nirik> so, any votes on the exception? or do we need more info?
18:16:33 <ktdreyer> /usr/sbin/kadmin is in both MIT and Heimdal
18:16:42 <ktdreyer> I don't think Heimdal calls it "hadmin"
18:16:44 <gholms|work> Ugh, they both call it kadmin?
18:16:46 <ktdreyer> yeah
18:16:51 <gholms|work> Must be an openbsd thing.
18:17:08 <gholms|work> (That's the only place I've used heimdal)
18:17:36 <ajax> -1 to the exception for now, but please do bring it back up if there's a need
18:17:57 <ajax> also, someone kick the mit krb5 package to stop using /usr/kerberos already
18:18:11 <nirik> ditto what ajax said here. ;)
18:18:19 <ktdreyer> heh, gl chaging EL5/6 :)
18:18:22 <notting> ajax: that would imply replacing the system telnet/rsh/etc.
18:18:39 <ajax> notting: alright.
18:19:27 <nirik> who uses those anymore anyhow? ;)
18:20:08 * nirik sees we are at -3 currently.
18:20:22 <notting> -1 without more reason, at least.
18:20:35 <nirik> cwickert / mclasen ?
18:20:45 <ajax> i mean, to the extent that the "system" telnet and the krb5 telnet have diverged, that's a problem
18:21:07 <ajax> but that's already a problem
18:22:40 <ktdreyer> so, we'll try to use alternatives where we can, move conflicting files in to subdirs where we can, and if necessary, make the -devels conflict
18:23:19 <ktdreyer> sound ok?
18:23:38 <ajax> wfm
18:24:00 <ajax> since, effectively, not getting a vote quorum here means there's no exception granted
18:24:05 <ktdreyer> by the way, I gather that the lack of an official template for FHS exceptions is not a bug, but a feature ;)
18:24:06 <nirik> yeah, guess so.
18:24:36 <nirik> #info -4 votes on this. Exception is not granted at this time.
18:25:03 <nirik> ok, anything further on this?
18:25:11 <nirik> thanks for coming ktdreyer...
18:25:41 <ktdreyer> sure, thanks for your time.
18:26:24 <nirik> #topic Open Floor
18:26:35 <nirik> I have one item for open floor...
18:26:48 <ajax> bring it
18:27:01 <nirik> I will be traveling next week... will be in Boston on wed. I may or may not be able to run the meeting. :) If someone else would like to run it that would be great.
18:27:34 * mclasen belatedly back
18:27:35 <ajax> i should be around, if no one else is
18:28:06 <nirik> I may also not be able to send out the agenda tuesday...
18:28:25 <nirik> anyone have any other open floor items?
18:28:52 * notting should be around too
18:29:12 <ajax> nothing from me
18:30:04 <nirik> ok, lets call it a short meeting then... :)
18:30:09 <nirik> Thanks for coming everyone!
18:30:13 <nirik> #endmeeting