path: root/meeting-logs/20120814.txt

21:00 <@  grobian> dah, ok, I started roll call
21:00 -!- grobian changed the topic of #gentoo-council to: meeting now | agenda: http://dev.gentoo.org/~grobian/agenda-20120814 | http://www.gentoo.org/proj/en/council/utctolocal.html?time=1900 | http://www.gentoo.org/proj/en/council/
21:01 <@Betelgeus> \o/
21:01 <+scarabeus>   |
21:01 <+scarabeus> damn
21:01 <  WilliamH> I'm here.
21:01 <@ Chainsaw> Present.
21:01 <@  grobian> yeah, only ulm and dberkholz are missing
21:01 <@  grobian> so let's wait a bit
21:01 <@ Chainsaw> ulm was here earlier.
21:01 <@ Chainsaw> Perhaps we can give him a call?
21:02 <@Betelgeus> I can
21:02 <@ Chainsaw> It is appreciated.
21:02 <@      ulm> here
21:02 <@  grobian> good
21:02 <@ Chainsaw> Just dberkholz then.
21:02 <@Betelgeus> then call donny
21:03 <@  grobian> right
21:03 <@  grobian> let's wait 2 more minutes
21:03 <@ Chainsaw> I don't have numbers I'm afraid.
21:03 <@Betelgeus> Went into some weird google voice answering machine
21:03 <@ Chainsaw> Just the ones that replied to the posting this year.
21:04 <@ Chainsaw> Betelgeuse: Well, thanks for trying all the same.
21:04 <@  grobian> yeah, much appreciated
21:05 <@Betelgeus> donnie gets a slacker mark right?
21:05 <@  grobian> yup
21:05 <@  grobian> ok, people, we have only one item
21:05 <@ Chainsaw> EAPI5?
21:05 <@  grobian> if you haven't yet, please load the agenda (see topic)
21:06 <@  grobian> so, there's two things from my point of view
21:06 <@  grobian> the original idea was to vote on items for EAPI5 that are ready to go NOW
21:06 <@ Chainsaw> I am opposed to having EAPIs in flux. So we can vote now on making what's ready EAPI5.
21:06 <@  grobian> however, the extensive list that ulm sent out needs some thorough research, of which I did a bit, included in the agenda
21:06 <@ Chainsaw> Or we can postpone.
21:06 <@  grobian> so, point one is, do we want to discuss any of it this meeting
21:07 <@ Chainsaw> Looking at the list, what's ready is... not a long list.
21:07 <@  grobian> 1 or two entries
21:07 <@  grobian> with my not-so-thorough research
21:07 <@  grobian> point 2 would be to vote on the features to be in or out
21:07 <  WilliamH> My vote would be to postpjone.
21:07 <@  grobian> but honestly, I'd like to defer
21:07 <@      ulm> some of the things marked as not implemented are trivial
21:07 <  WilliamH> postpone
21:07 <@  grobian> ok
21:08 <@  grobian> let's vote for postponing the EAPI5 thing
21:08 <@ Chainsaw> Postpone please.
21:08 <+scarabeus> I agree we should only pick from those that are ready
21:08 <@      ulm> yeah, let's postpone it to September
21:08 <@ Chainsaw> And if the trivial things could then be implemented and marked so that there is a more impressive list...
21:08 <@ Chainsaw> That would be great.
21:08 <@  grobian> scarabeus: Betelgeuse what do you vote for?
21:09 <@  grobian> postpone or handle it now?
21:09 <@Betelgeus> grobian: postpone but we can use the time to talk about things
21:09 <+scarabeus> it can be postproned by all means; but we can talk about it
21:09 <@  grobian> ok
21:09 <@ Chainsaw> *nod* Sounds fair.
21:10 <@  grobian> well, that is fine, although I don't feel much need to discuss it here
21:10 <@  grobian> because it needs some more research  from my side
21:10 <@  grobian> I'd prefer good follow up's on ML
21:10 <  WilliamH> Same here.
21:10 <  Arfrever> You can discuss things already implemented in Portage.
21:10 <@  grobian> so I suggest to continue the agenda, and then you lot can chat what you want
21:10 <@  grobian> Arfrever: I'm going to ignore your suggestion here
21:10 <@  grobian> sorry
21:11 -!- ulm_ [~ulm@gentoo/developer/ulm] has quit [Ping timeout: 244 seconds]
21:11 <@  grobian> So, there are no open bugs with council involvement
21:11 <@  grobian> done
21:11 <@  grobian> open floor
21:11 <+scarabeus> I was asked by johu for one thing
21:11 <@  grobian> anyone (everyone) who wants to raise some issue
21:11 <@  grobian> Arfrever: if you want, go ahead here
21:11 *** Chainsaw activates the microphone
21:12 *** grobian listens
21:12 <+scarabeus> if we could change "dev should use latest eapi when bumping" to "dev must use latest eapi when bumping if not forbidden by eclasses"
21:12 <@  grobian> deafening silence
21:12 <@      ulm> patrick has asked me to bring up deprecation of EAPI 1
21:12 <@ Chainsaw> scarabeus: I have ebuilds where that's unworkable.
21:12 <@  grobian> scarabeus: rationale?
21:12 <@ Chainsaw> scarabeus: Example: net-misc/bird
21:12 <@      ulm> what's the general opinion about it?
21:12 <@ Chainsaw> scarabeus: Reason: awkward build sequence.
21:12 <@ Chainsaw> ulm: Rationale?
21:12 <@  grobian> ulm: again? wasn't that horse beaten to death already?
21:12 <@ Chainsaw> ulm: Again, I have ebuilds where it simply can't be done.
21:12 <@Betelgeus> Chainsaw: new phases could always be no-ops?
21:13 <@ Chainsaw> Betelgeuse: That sounds incredibly awkward to me.
21:13 <@ Chainsaw> Betelgeuse: I will, where feasible, bump EAPI on my ebuilds.
21:13 <@      ulm> he gave no rationale
21:13 <@ Chainsaw> Betelgeuse: Where it isn't, I don't see adding empty phases as a particularly acceptable workaround.
21:13 <@      ulm> only a plan
21:14 <+scarabeus> problem is that quite few devs just cp without checking on it
21:14 <@      ulm> <bonsaikitten> my plan would be: deprecate eapi1 "now", add repoman warning, make repoman warning fatal in 3 months (not many ebuilds anyway)
21:14 <@      ulm> <bonsaikitten> then plan to deprecate eapi2 in, say, 6 months
21:14 -!- kallamej [~kallamej@gentoo/developer/kallamej] has quit [Quit: leaving]
21:14 *** Chainsaw wonders where this obsession to remove backward compatibility comes from
21:14 <+scarabeus> -    if( !pResultBitmapEx )
21:14 <@ Chainsaw> Upgrade paths are going to *suck*.
21:14 <+scarabeus> http://qa-reports.gentoo.org/output/eapi_usage.txt
21:14 <+scarabeus> this
21:15 <+scarabeus> it is 550 pkgs
21:15 <+scarabeus> and you can see eapi3 is more candidate for killing than eapi2
21:15 <@ Chainsaw> scarabeus: Yes, net-misc/bird is one of the EAPI1s listed there.
21:15 <  Arfrever> Let's discuss deprecation of EAPI="0". Any opinions?
21:15 <@      ulm> was about 800 few months ago
21:15 <@      ulm> Arfrever: upgrade path
21:15 <@  grobian> all EAPI=2 can be sedded to EAPI=3
21:16 <+scarabeus> they are mostly killed with stabilisations
21:16 <  Arfrever> ulm: All sys-apps/portage ebuilds already use EAPI="2" or EAPI="3".
21:16 <@  grobian> Arfrever:you can't, so not really a fruitful suggestion , is it?
21:16 <+scarabeus> I am all for kiling it but it must forbid adding new ebuilds of such eapi, not just stop commit if it is in the folder
21:16 <@ Chainsaw> Arfrever: Again, obsession with the eradication of backward compatibility. Why?
21:16 <  Arfrever> Chainsaw: Simplification of eclasses.
21:17 <@      ulm> Arfrever: not true
21:17 <  Arfrever> ulm: What is not true?
21:17 <@      ulm> <Arfrever> ulm: All sys-apps/portage ebuilds already use EAPI="2" or EAPI="3".
21:18 <@  grobian> scarabeus: I think it would help if you'd flesh this proposal out a bit more, it doesn't sound too bad to me to have a policy to only add new stuff of the latest EAPI, for as long as EAPIs don't have orthogonal features
21:18 <@      ulm> portage-2.1.6.7_p1.ebuild is still EAPI 0
21:18 <  Arfrever> ulm: There is no such ebuild in gentoo-x86.
21:18 <  Arfrever> ulm: OK. grep failed to find it.
21:19 -!- kallamej [~kallamej@gentoo/developer/kallamej] has joined #gentoo-council
21:20 <  Arfrever> ulm: Another rationale: Upgrade path is required only for 1 year. EAPI="2" was added long time ago.
21:20 <@  grobian> scarabeus: and a ML discussion would be a good thing too
21:21 <+scarabeus> you are right
21:21 <+scarabeus> johu: start a chat on -project :-)
21:21 <+scarabeus> delegation at its best
21:22 <@Betelgeus> no -project?
21:22 <@Betelgeus> on -dev?
21:22 <+scarabeus> ah right
21:22 <  Arfrever> Another item for discussion: What happens when profiles/eapi uses EAPI >=2. I noticed that nothing happens in overlays which use new EAPI.
21:23 <@  grobian> Arfrever: that sounds like something that should be discussed with people into the subject first
21:23 <+scarabeus> grobian: about that sed, I would ask diego to run it through tinderbox to see nothing really exploded, ie python eclass change behaviour with each eapi and such tiny funny things
21:23 <+scarabeus> also I am all hands for blocking additions of new eapi1 ebuilds
21:23 <@  grobian> scarabeus: I don't see the use of removing EAPIs
21:24 <@ Chainsaw> scarabeus: Great. Will you be porting my EAPI=1 net-misc/bird then?
21:25 <@ Chainsaw> scarabeus: And yes, that is a challenge. And no, QA warnings are not acceptable.
21:25 <@  grobian> I can imagine python eclass not supporting EAPI 0,1,2
21:26 <@  grobian> but that doesn't mean those eapi's should be banned
21:26 <@  grobian> an eclass can require its consumers to be using an up-to-date eapi
21:26 <@  grobian> it's one herd anyway
21:26 <@ Chainsaw> They were EAPI=4 holdouts for a long time. Seems fine.
21:28 <@  grobian> ok, can we conclude the open floor with this?
21:28 <@ Chainsaw> grobian: Just waiting for scarabeus to see the EAPI=1 net-misc/bird conundrum.
21:28 <@  grobian> Chainsaw: does it matter for the summary?
21:28 <@  grobian> I guess it does
21:28 <+scarabeus> Chainsaw: i will ask nic guys to fix build system, simple
21:29 <@ Chainsaw> scarabeus: Perhaps you speak their language, that may help. I did try.
21:29 <  Arfrever> Any opinions about handling of icons (*.ico) in dohtml?
21:29 <@ Chainsaw> scarabeus: (Just like we asked for a Quagga-style telnet multiplexer)
21:30 <@  grobian> Chainsaw: I did my best to summarise that one
21:30 <@ Chainsaw> grobian: Okay, thanks :)
21:31 <@  grobian> good, then I'd like to close the open floor
21:31 <@ Chainsaw> grobian: Please proceed.
21:31 <@  grobian> next meeting 11 september
21:31 <@  grobian> most likely I won't make it
21:31 <@  grobian> so ulm will be your chairmaster
21:31 <@      ulm> grobian: your summary says 14 september
21:32 <@  grobian> sorry, 14 september
21:32 <@  grobian> yeah
21:32 <@  grobian> will fic
21:32 <@  grobian> fixed
21:32 <@  grobian> 11 ot is
21:32 <+scarabeus> ack
21:32 <@  grobian> ok, everybody happy?
21:32 <@      ulm> 11 is a tuesday, right
21:32 <@  grobian> I'll send out the editted agenda for review
21:32 <@  grobian> ulm: yes
21:33 -!- ivan\ [~ivan@unaffiliated/ivan/x-000001] has joined #gentoo-council
21:33 <@      ulm> grobian: thank you for chairing
21:34 <@  grobian> ok, let's close the meeting then
21:34 -!- Chainsaw [~chainsaw@gentoo/developer/chainsaw] has left #gentoo-council ["Thanks grobian."]
21:35 -!- grobian changed the topic of #gentoo-council to: next meeting: 2012-09-11 19:00 UTC | http://www.gentoo.org/proj/en/council/utctolocal.html?time=1900 | http://www.gentoo.org/proj/en/council/
21:35 <@  grobian> thanks all for your delightful input ;)