20:10 <stgraber> #startmeeting Ubuntu Technical Board meeting
20:10 <meetingology> Meeting started Mon Jun 24 20:10:09 2013 UTC.  The chair is stgraber. Information about MeetBot at http://wiki.ubuntu.com/meetingology.
20:10 <meetingology> 
20:10 <meetingology> Available commands: #accept #accepted #action #agree #agreed #chair #commands #endmeeting #endvote #halp #help #idea #info #link #lurk #meetingname #meetingtopic #nick #progress #rejected #replay #restrictlogs #save #startmeeting #subtopic #topic #unchair #undo #unlurk #vote #voters #votesrequired
20:10 <stgraber> #topic Action review
20:10 <stgraber> none listed on the wiki page
20:10 <stgraber> #topic ColinWatson (puppeting for RickSpencer): "rolling" alias for development series
20:11 <kees> everything last week got taken care of before I updated the agenda, so we're clean there
20:11 <mdz> hi
20:11 <kees> heya mdz
20:11 <mdz> sorry I'm late
20:11 <kees> no worries, we just got started
20:11 <cjwatson> Ah, good
20:12 <stgraber> hey mdz. No worries. I'll mark your as chair for the next meeting if that's alright (you were supposed to chair this one)
20:12 <stgraber> *you
20:13 <mdz> eek, sorry about that
20:13 <cjwatson> So this (topic) is a leftover from UDS, which we assigned to Rick to come back to us with a name for approval: the name is to be used as an alias for whatever the current development series is, both for the publisher (so that people can choose to just stick with the alias rather than remembering to change every six months) and for uploads (so that people targeting "whatever's current" can just keep uploading to the ...
20:13 <cjwatson> ... current alias without having to remember to change processes every six months)
20:13 <cjwatson> After a WHOLE bunch of back-and-forth, Rick has come back with "rolling"
20:14 <cjwatson> Hands up if you're surprised :)
20:14 <cjwatson> He's on holiday this week, so I said I'd puppet
20:14 <kees> for uploading, this means the release in the changelog?
20:14 <cjwatson> From my point of view, as long as the name doesn't clash with anything else significant, I consider it a bikeshed and am prepared to go with basically anything reasonable
20:14 <cjwatson> Right
20:14 <cjwatson> (Or the sftp target, for those who use that)
20:15 <kees> well sftp already has the "ubuntu" target.
20:15 <cjwatson> It's actually a bit more than that :)
20:15 <cjwatson> It's just not well known
20:15 <cjwatson> (Anyway, that's a fine detail)
20:16 <kees> for changelog, i'm less happy... this means there isn't really a sourceful hint as to the target release. hm
20:16 <cjwatson> Hasn't hurt Debian
20:16 <kees> but I guess it doesn't matter in practice.
20:16 <cjwatson> I personally wasn't going to use it for regular Ubuntu uploads
20:16 <cjwatson> But it would be available
20:16 <kees> things going into -proposed will still need a release name dash proposed...
20:16 <cjwatson> Not true
20:17 <cjwatson> Magic aliasing :)
20:17 <kees> well, okay, not NEED
20:17 <cjwatson> I haven't uploaded anything explicitly to -proposed for about six months
20:17 <kees> ok, I'm over my mental disruption about chnagelog now :)
20:17 <cjwatson> Direct uploads to the release pocket are redirected to -proposed automatically, so there's been no need
20:18 <cjwatson> I haven't started the implementation yet, but it wouldn't surprise me if the rolling -> <current dev> aliasing would happen most naturally after -proposed is stripped off, anyway
20:18 <kees> yeah, and nothing should use the changelog for that.. it should use actual pkg and version details
20:18 <cjwatson> Right, especially given widespread copies in modern Ubuntu
20:19 <kees> (thinking about the security workflow, for example)
20:19 <cjwatson> Yeah, I would be surprised if security didn't want to continue just using the true suite names; they don't have an awful lot of "oh, just use whatever's current" mindset
20:19 <kees> okay, I'm cool with what a rolling release would mean for the technical bits.
20:20 <kees> now, name. "rolling" is fine, but it'd be nice to have something snappier
20:20 <cjwatson> My only personal caveat about the name is that it sort of socially engineers us into coming closer to an *actual* rolling release, but arguably, we're gradually heading that way anyway and as long as we don't do it precipitously (which I don't think is going to happen) ...
20:21 <cjwatson> So I'm cool with it
20:21 <kees> what were the other suggested names, and why were they not used?
20:22 * kees suggests "icarus"
20:22 <cjwatson> Honestly, I don't have a full list, we delegated the naming to Rick and this is what he came back with
20:22 <kees> or "babel"
20:22 <cjwatson> "ubuntu" was suggested; I said "argh, category conflict and hideous confusion"
20:22 <kees> haha, yeah, no
20:22 <cjwatson> So I'd be -1 on that one, but mentioning it for the record :)
20:23 <kees> i don't like "rolling" for similar reasons "rolling release confusion"
20:23 <cjwatson> If you want a full "why weren't other things used" then we'll have to wait for Rick to get back from holidays
20:24 <mdz> how about "current"?
20:24 <cjwatson> That and "next" were my preferred names, but I ran out of energy for bikeshedding and decided I wasn't too bothered.  I think "current" has the problem that people ask "well, isn't the LTS current?"
20:25 <kees> "current" would be sufficiently different from "stable" and "devel"
20:25 <cjwatson> Which I'd acknowledge - in fact it's similar to part of the problem with "ubuntu"
20:25 <kees> i like "next" a lot
20:25 <stgraber> http://paste.ubuntu.com/5796585/
20:25 <cjwatson> "next" appeals to git people but not sure it's broad enough, and I think there was some concern about confusion with unity next
20:25 <stgraber> that's the list of names we came up during the vUDS session ^
20:26 <mdz> "next" WFM also
20:27 <kees> i don't like "tip" "head" or "trunk" because they're too tied to vcs imho. i don't have a vcs association with "next" as strongly.
20:27 <kees> i think "devel" is too overloaded already
20:28 <kees> (oh and "master" is too vcs-y for me too)
20:29 <stgraber> I personally like "next" and dislike "rolling" for the same reason kees mentioned earlier.
20:29 <kees> it sounds like "next" is best from that list? using a declarative instead of a proper noun also avoids release name confusion
20:29 <cjwatson> next and rolling are the same part of speech :-)
20:29 <cjwatson> Well, ish
20:30 <kees> i meant "rolling" being confused with "rolling release"
20:30 * slangasek gets summoned
20:30 <cjwatson> kees: I guess my question is, how inaccurate would that really be
20:30 <kees> and I prefer "next" because it isn't an animal
20:31 <cjwatson> kees: Oh, certainly we should stay away from anything that could fit into the category this is aliasing
20:31 <kees> cjwatson: inaccurate? well, it's not a release, and to switch from "next" to "rolling" for an _actual_ rolling release would take non-zero work
20:31 <slangasek> so I'm not sure I can offer any more concrete justifications for one name or another, but it would be nice if this could be decided since it's been carried over several times now (through no fault of the TB)
20:32 <cjwatson> We (collectively) nacked the notion of moving straight to a rolling release in one step; but I think we are kind of heading in that direction in a variety of ways
20:32 <cjwatson> So I suppose it's kind of an aspirational name
20:32 <slangasek> if there's not a consensus for "rolling", should this maybe go to the next meeting, when rickspencer3 can be available?
20:32 <cjwatson> I agree that "rolling" and "release" (the latter in the traditional Ubuntu sense anyway) are kind of oxymoronic in combination
20:32 <kees> right, so I'd like to discourage a rolling release by not naming this alias "rolling"
20:33 <cjwatson> I can start the work either way if the TB doesn't object to the fundamental concept, and slot in the name at the end, so it doesn't necessarily delay us a lot to wait
20:33 <kees> slangasek: how about we go with "next" and if it has to be renamed, do that at a separate TB meeting. that way you're unblocked
20:33 <cjwatson> (The name would likely be a database entry anyway)
20:34 <kees> +1: alias for whatever is currently devel
20:34 <kees> +1: naming it "next"
20:34 <slangasek> cjwatson: does that unblock you?
20:34 <cjwatson> I'd hoped not to have to go round again, but it does look like "next" is the rough consensus here
20:34 <kees> what happens when there is no devel release? (right after stable cuts?)
20:35 <cjwatson> The interval there is minutes
20:35 <kees> heh, ok
20:35 <cjwatson> Well, maybe an hour or two at most
20:35 <cjwatson> I will try to make it not actually explode LP in that interval :)
20:35 <kees> well, i guess it appears, but can be frozen. nm
20:36 <cjwatson> slangasek: For now, yes
20:36 <kees> having this alias will make several of my tools much happier :)
20:38 <cjwatson> Do we need to explicitly vote here, or do we have clear consensus?
20:38 <stgraber> sorry, was re-watching some of the vUDS discussion. So sounds like we all agree on having the alias (I thought we did that already at an earlier meeting but it's good nobody changed their mind) and that we tend to prefer "next"
20:39 <stgraber> Do we want to vote on the name and have this set in stone or do we feel like we should give Rick a chance to convince us to use "rolling" instead?
20:39 <kees> i didn't see if soren had thoughts
20:39 <kees> i would like it actively not be "rolling" :P
20:40 <mdz> "rolling" seems misleading but apparently there are some political considerations here
20:40 <soren> kees: I didn't have a strong enough opinion to speak up :)
20:40 <kees> soren: heh :)
20:40 <stgraber> same here, I plan on -1 any vote to get "rolling" as that name, doesn't mean there can't be enough +1 to balance that out :)
20:41 <cjwatson> Well, at least for now, that would clearly require more votes than are present, so let's assume we wouldn't
20:41 <cjwatson> It sounds like if Rick wants to press for rolling then he needs to argue that case directly rather than by inadequate proxy :)
20:41 <kees> let's go with "next" and if it needs additional bikeshedding it can happen later
20:42 <cjwatson> I'll suggest that either he make the next meeting or he argue by e-mail
20:42 <kees> regardless, it sounds like you're unblocked and the dev work to support "next" can start
20:43 <stgraber> it's just going to be a pain to change after this goes live, so I'd suggest we let cjwatson do the actual implementation but just wait until after the next TB meeting to flip the switch
20:43 * kees nods
20:43 <stgraber> so Rick (or anyone else for that matter) will have until then to argue for another name or we'll go with "next"
20:44 <kees> +1
20:44 <cjwatson> This certainly encourages me to put the name in the DB rather than hardcoding it, which is probably a good idea
20:44 <mdz> ok, 30 minutes on this topic. time to move on? :-)
20:44 <cjwatson> Not that I was massively inclined towards the latter, but still
20:44 <cjwatson> Sure
20:45 <cjwatson> Thanks
20:45 <stgraber> alright, moving on
20:45 <stgraber> #topic Scan the mailing list archive for anything we missed (standing item)
20:45 <kees> openss
20:45 <kees> l
20:46 <stgraber> right and a Micro release exception for Xen
20:46 <cjwatson> Oh, god.  I haven't really processed kees' reply on openssl yet.  Sorry ...
20:46 <Daviey> Discussion on OpenSSL as a system library seems to be wedged.
20:46 <kees> what's the update history on xen? do we have good testing for it?
20:47 <cjwatson> I don't think I'm persuaded by kees' position but I still need to reply properly
20:47 <kees> cjwatson: yeah, i figure openssl should continue a bit longer on the list, but we will need a vote at some point. slangasek promised a rebuttal too
20:47 <cjwatson> Daviey: Is mongodb upstream's promise to add an exception still proceeding?
20:48 <Daviey> cjwatson: yes, but MUCH slower than we hoped.
20:48 <Daviey> cjwatson: (but this is also related to the squid issue aswell.)
20:48 <kees> i didn't suspect you would be, since i can see the key elements where we have a separate conclusion on the same details :0
20:48 <kees> er, :P
20:49 <kees> i will follow up on the list with mre questions for xen.
20:50 <kees> btw, we have several standing "provisional" mres still. bdmurray pointed this out to me. do we want to set a specific time to review them?
20:50 <Daviey> I'm not comfortable speaking on behalf of the kernel team regarding xen, but my feeling as a server rep is that it isn't that good.  That said, I am supportive of this - providing a good level of QA is performed.
20:50 <stgraber> right, so we need to continue both the openssl discussion and the Xen MRE discussion on the mailing-list. I meant to ask some questions wrt testing to Stefan but apparently got distracted and forgot about it...
20:50 * kees nods
20:50 <stgraber> kees: we can add that to the agenda of our next meeting
20:51 <kees> agreeds
20:51 <kees> and I'll add an official vote for the openssl thing, just to have it done.
20:51 <stgraber> ok
20:51 <cjwatson> Daviey: Hm, I thought it was a lot better from precise on
20:52 <cjwatson> Is that not the case?
20:52 <Daviey> cjwatson: Oh, it is MUCH better from precise.  But I do not know how good our regression testing is.
20:52 <kees> me too, but I haven't actually paid close attention
20:52 <cjwatson> Fair enough
20:52 <kees> cool
20:52 <stgraber> #topic Check up on community bugs (standing item)
20:52 <stgraber> count is still 0, moving on
20:52 <kees> empty!
20:52 <stgraber> #topic Select a chair for the next meeting
20:53 <stgraber> that'll be mdz
20:53 <mdz> ack
20:53 <stgraber> #topic AOB
20:53 <stgraber> anything?
20:53 <kees> nothing from me.
20:53 <mdz> nope
20:54 <cjwatson> Not I
20:55 <stgraber> ok then, thanks everyone!
20:55 <kees> thanks stgraber!
20:55 <stgraber> #endmeeting