path: root/meeting-logs/20120911.txt

<ulm> let's start?						        [21:00]
*** willikins (~rbot@gentoo/bot/Willikins) has quit: Ping timeout: 245 seconds
<Chainsaw> Okay!
<dberkholz> i'm here this time!
<blueness> yes
<dberkholz> was on a 3-week vacation during the last one, forgot all about it
<ulm> anyone logging?
<Chainsaw> You've been missed Donnie. Glad you made it.
<blueness> i will be proxying for fabian
<dberkholz> nothing interesting was happening anyways =)
<Chainsaw> dberkholz: More fireworks expected for this one, yes. EAPI 5.
								        [21:01]
<ulm> betelgeuse, scarabeus, williamh?
<Chainsaw> scarabeus was here earlier.				        [21:03]
<scarabeus> i am here
<Betelgeuse> here
<ulm> does someone have the phone number of williamh?
<Chainsaw> Not posted to council ML, sorry. Don't have it.	        [21:04]
* ulm will try to keep an online summary at
  http://dev.gentoo.org/~ulm/council-20120911.txt		        [21:05]
<dberkholz> cool, thanks ulm
<dberkholz> maybe we should find someone to be a secretary so you don't have
	    to do that?						        [21:06]
<ulm> dberkholz: shouldn't be a problem today
<ulm> I hope so, at least ;)
<ulm> should we proceed?					        [21:07]
<blueness> yes
<ulm> EAPI 5 features						        [21:08]
<ulm> list is here: http://article.gmane.org/gmane.linux.gentoo.project/2140
								        [21:09]
<dberkholz> before we get into the details..
<ulm> dberkholz: yes?
<dberkholz> i'd like to propose that anything we can't agree on by the next
	    meeting should get pushed to 6, and whatever's left on the list,
	    and implemented in portage, is 5.
<dberkholz> in the interests of actually getting this out there
<ulm> dberkholz: o.k., but let's see how it goes		        [21:10]
<dberkholz> just speaking from past experience with some other EAPIs that have
	    dragged on for more than a year =)			        [21:11]
<ulm> I suggest that we vote about the features in the first group en-bloc,
      and do the rest one by one
<ulm> except if someone wants to single out additional things from the first
      group
<ulm> in fact, I'd like to vote on "econf --disable-silent-rules" and
      "doheader" separately ;)					        [21:12]
<blueness> ulm what defines "the first block"?			        [21:13]
<Chainsaw> I like the --disable-silent-rules for EAPI 5. Just not
	   retroactively. That is moving the goal posts and invites arbitrary
	   breakage on remerge.
<scarabeus> blueness:
	    http://www.blesk.cz/clanek/zpravy-udalosti/181196/to-je-ale-zradlo-poslanec-radl-ods-ji-rizek-primo-z-lavice-ve-snemovne.html
<scarabeus> Chainsaw: yeah eapi5+
<ulm> blueness: see list of eapi 5 features, link in agenda
<Chainsaw> ulm: What that list is *sorely* missing is... plain text summaries
	   of the feature.
<Chainsaw> ulm: I have unified diffs of LaTex (do you expect me to render that
	   in my head?) or long bug reports.			        [21:14]
<Chainsaw> ulm: Not ideal.
<blueness> okay: i'd like to vote for user-patches and licence groups
	   separately
<blueness> (or that's in the second group i guess)		        [21:15]
<Betelgeuse> Also for future reference is the motivation for features listed
	     somewhere yet?					        [21:16]
<Chainsaw> Betelgeuse: Vaguely implied in the bug if you're lucky.
<dberkholz> that should probably go into the devmanual		        [21:17]
<ulm_> sorry, I had lost connection
<Chainsaw> Welcome back.					        [21:18]
*** ulm (~ulm@gentoo/developer/ulm) has quit: Ping timeout: 240 seconds
*** ulm (~ulm@gentoo/developer/ulm) has joined channel #gentoo-council
*** ChanServ (ChanServ@services.) has changed mode for #gentoo-council to +o
    ulm
<Chainsaw> ulm: Are you able to scroll back or would you like me to repeat my
	   earlier concern with the list?			        [21:19]
*** spiros (~andyspiro@gentoo/developer/spiros) has joined channel
    #gentoo-council
* ulm is catching up
<Chainsaw> Okay.
<ulm> the first vote would be for inclusion of the 12 items listed in
      http://dev.gentoo.org/~ulm/council-20120911.txt
<ulm> blueness: user patches and license groups are in the second group
								        [21:20]
<blueness> ulm: yes i see that,  i can vote for all of part 1 as a block
<blueness> when we are ready for the vote
<ulm> anyone wants to discuss items (out of these 12) separately?       [21:21]
<ulm> both spec and Portage implementation are ready for all of them
<ulm> please vote on inclusion of these 12 items in EAPI 5	        [21:22]
* ulm votes yes
* blueness votes yes
<dberkholz> no objections					        [21:23]
* Chainsaw approves and votes yes
* scarabeus yep
<Betelgeuse> yes
<Chainsaw> (Next time a ~2 sentence summary per item in the list please, so I
	   don't have to open 12 tabs to find all this out on the day...)
<Betelgeuse> although can someone repeat what the sub slot stuff is for?
<ulm> I count 6 yes						        [21:24]
<ulm> Betelgeuse: most ingenious feature of all of them ;)
<ulm> will allow us to get rid of revdep-rebuild and preserved libs     [21:25]
<Arfrever> ulm: You might paste list of features on IRC so that future readers
	   of log of meeting have no problem with finding what was voted.
	   (council-20120911.txt might not exist in the future.)
<ulm> Arfrever: good point
<Chainsaw> Arfrever: It will. ulm is extremely organised.
<ulm> * Slot operator dependencies
<ulm> * Sub-slots
<ulm> * Profile IUSE injection					        [21:26]
<dberkholz> only took 3 hours to read through all the emails about it =)
<ulm> * At-most-one-of operator for REQUIRED_USE
<ulm> * EBUILD_PHASE_FUNC variable
<ulm> * Mandate GNU find
<ulm> * new* commands can read from standard input
<ulm> * Parsing of the EAPI assignment is mandatory
<ulm> * src_test support for parallel tests
<ulm> * Stable use forcing and masking
<ulm> * Option --host-root for {has,best}_version
<ulm> * usex helper function
<ulm> ^^ all accepted unanimously
<Betelgeuse> ulm: seems I missed part of the diff when reading it previously.
	     It was clearer reading it again.			        [21:27]
<ulm> OK, next one is "doheader helper function"
<ulm> spec and implementation exist
<ulm> however, this was already proposed for EAPI 3 (then called doinclude)
<ulm> and the council had rejected it				        [21:28]
<ulm> link to previous council summary is here:
      http://www.gentoo.org/proj/en/council/meeting-logs/20090423-summary.txt
<Chainsaw> Good name, seems clear what it's for. I see no reason not to have
	   it.							        [21:29]
<dberkholz> i'll be back in a few ... in case the votes come up, i'm against
	    retroactively applying anything and the repository stuff (which
	    may need further discussion)			        [21:30]
<ulm> Chainsaw: it certainly won't harm :)
<ulm> the former argument against it was that to many do* functions would
      confuse new devs
<Chainsaw> ulm: May have been the inclusion of doexample in the same item that
	   did it in?
<ulm> Chainsaw: that's possible					        [21:31]
<ulm> anyway
<ulm> please vote on inclusion of doheader in EAPI 5
* Chainsaw votes Yes
* ulm votes yes
<Chainsaw> Betelgeuse?
* scarabeus ++							        [21:32]
<ulm> Betelgeuse, blueness?
<blueness> grobian had some issues with block 2			        [21:33]
<ulm> blueness: your vote about "doheader"?
<blueness> oh doheader: yes
<Betelgeuse> yes
<Chainsaw> With dberkholz having stepped out, I think that's as many votes as
	   we are going to get.					        [21:34]
<ulm> 5 yes 0 no
<ulm> accepted
<Betelgeuse> but should note that a common use case in the tree is to install
	     headers to a subdirectory
<Betelgeuse> so the usage will probably be marginal
<ulm> Betelgeuse: yeah
<ulm> same as for dolib
<blueness> ulm sorry an emergency just came up, i have to go help my wife
<blueness> can i give you grobians' wishes?
<ulm> blueness: yes, please					        [21:35]
<blueness> i'm so sorry
<blueness> her car broke down
<blueness> this is for part 2 -> http://dpaste.com/799573/
<Chainsaw> blueness: Family always comes first.
<blueness> please do not count this against him, it was a pure emergency
<blueness> thank you gentlemen
<ulm> blueness: so, all no for part 2 from grobian		        [21:36]
<scarabeus> thats straight
<scarabeus> :D
<ulm> and for part 3, in fact
<ulm> OK, next one
<ulm> econf --disable-silent-rules
<Chainsaw> Okay for EAPI 5. *Nothing* gets applied retroactively. *EVER*
<ulm> I suggest that we vote for this one in EAPI 5 first
<scarabeus> ^^what Chainsaw  said				        [21:37]
<ulm> then discuss the retroactive stuff
* Chainsaw votes Yes for --disable-silent-rules in EAPI 5
<ulm> please vote for "econf --disable-silent-rules" in EAPI 5
<Betelgeuse> yes
* ulm votes yes
<ulm> 4 yes
<ulm> and that's all votes we get, I fear			        [21:38]
<Chainsaw> Sounds unanimous then.
<ulm> yes, accepted for EAPI 5
<scarabeus> (quite redux from 7 to 4 :-/)			        [21:39]
<Chainsaw> scarabeus: Four is more exciting. With an even number of votes you
	   can deadlock.
<ulm> do we want to apply this retroactively to EAPI 4?
<Chainsaw> ulm: NO.
<Betelgeuse> no
* ulm abstains
<Chainsaw> scarabeus?						        [21:40]
<ulm> scarabeus?
<scarabeus> what i said
<scarabeus> above
<scarabeus> so no
<Chainsaw> Excellent. Reason prevails.
<Chainsaw> Next?
<ulm> rejected for EAPI 4
<ulm> we don't have to vote for EAPIs 0 to 3 then		        [21:41]
<scarabeus> yea
<ulm> next: "user patches"
<Chainsaw> ulm: You could, but I'm just going to say no louder and more often.
<ulm> no working implementation in Portage so far
<Chainsaw> Controversy as well. Push to EAPI 6.			        [21:42]
<ulm> and IMHO the spec is incomplete too
<Chainsaw> It's simply not ready, agreed.
<scarabeus> eapi6 to speed things up, other than that idea is good, it needs
	    to be speced better
* WilliamH thought the time was 3 my time... I'm here late... probably most
  everything is done now. :(
<ulm> WilliamH: welcome
<Chainsaw> WilliamH: Please e-mail a phone number to the council list if at
	   all possible.					        [21:43]
<ulm> some of the controversial stuff still left
<WilliamH> Chainsaw: no problem.
<ulm> so, please vote on "user patches" for EAPI 5
* ulm votes no
* scarabeus nacks						        [21:44]
* Chainsaw votes No (suggests postponing to EAPI6)
* WilliamH votes no
<Chainsaw> Betelgeuse?
<Betelgeuse> later
*** NeddySeagoon (~NeddySeag@gentoo/developer/NeddySeagoon) has joined channel
    #gentoo-council
<Chainsaw> Hey Ned.						        [21:45]
<ulm> Betelgeuse: that is, no for 5 and include it in later EAPI?
<Betelgeuse> ulm: yes
<Chainsaw> ulm: That's how I read it, yeah.
<ulm> rejected for EAPI 5
<NeddySeagoon> hi Chainsaw
<ulm> next: "License groups in ebuilds"				        [21:46]
<ulm>  
<ulm> see my alternative proposal
<scarabeus> i like the alternative that works everywhere
* Chainsaw votes No on license groups and Yes on the ulm approach of GPL2+
<scarabeus> so no for the point
* ulm votes no
<Betelgeuse> ulm: should we ask if there is any item that gets support from
	     someone on the list for EAPI 5?
<Chainsaw> From the blueness list of grobian opinions, that is a "no".  [21:47]
<ulm> Betelgeuse: yeah, that could speed up things
<Chainsaw> I see nothing remaining that I would vote in.
<Chainsaw> I would say gather it up, polish it for EAPI6 until it shines, and
	   resubmit.						        [21:48]
<ulm> no support from me for any of the remaining ones
<scarabeus> actually the only thing i like is EJOBS because I implemented lots
	    of interesting hacks in scons eclass and similar and this would
	    ease stuff in future
<Chainsaw> scarabeus: Needs a rationale, a summary, etc.	        [21:49]
<WilliamH> I don't have an issue with ejobs either.
<Betelgeuse> agreed with Chainsaw, ulm
<Chainsaw> scarabeus: You okay to have that in EAPI6, or do you need it now?
<scarabeus> Chainsaw: it is not needed now, but it should not be forgotten
<Chainsaw> scarabeus: Oh sure, but there are more like that.	        [21:50]
<ulm> so I see nobody speaking up for any of the remaining items
* WilliamH agrees with scarabeus  on this... as long as we don't forget about
  it for later...
<ulm> I guess proponents won't forget
<ulm> if they do, it wasn't important ;)			        [21:51]
<WilliamH> heh
<Chainsaw> Quite :)
<Arfrever> In case of 423245, I suggest separate voting on each subfeature.
<ulm> so looks like we're done with EAPI 5
* WilliamH had a couple of questions about previous eapi 5 issues that I
  missed since I was late...
<WilliamH> Should that wait until open floor?			        [21:52]
<Chainsaw> WilliamH: You can ask us now, if you think it will change anything?
<Betelgeuse> Also there's time until tagging to change minds
<WilliamH> I don't know if it will, but I'll throw them out there... about
	   "parsing of the eapi being mandatory..."
<ulm> Betelgeuse: right						        [21:53]
<WilliamH> Did the community ever actually come to a concensus on that?
<ulm> WilliamH: Portage already does it since a few weeks
<Betelgeuse> WilliamH: there will never be a concensus
<ulm> and that was the original plan
<WilliamH> ulm: Oh ok, I'm fine with it then.
<WilliamH> For "stable use forcing and masking." Does the current proposal for
								        [21:54]
<Betelgeuse> ulm: I think we are done with items to vote on?
<Betelgeuse> ulm: I will need to head our for lunch soon
<WilliamH>  combining e.g. package.use.* and use.* affect anything?
<WilliamH> I don't have a link but that possibility was discussed on -dev.
								        [21:55]
<ulm> WilliamH: that's not on the list so far
<ulm> and I guess that it will need further discussion
<WilliamH> ulm: ok						        [21:56]
<Betelgeuse> off
<Betelgeuse> thanks people
<Chainsaw> ttyl Betelgeuse
<WilliamH> about the "doheader" functions...
<WilliamH> Shouldn't we also have newheader if we have doheader?
<ulm> WilliamH: we do
<Chainsaw> WilliamH: That is listed in the bug.
<Chainsaw> WilliamH: So yes, that will be included :)
<WilliamH> Oh ok, I didn't see it in the bug.
<ulm> Chainsaw: and in PMS and in Portage :)
* Chainsaw repeats the need for a clear summary in the list
*** willikins (~rbot@gentoo/bot/Willikins) has joined channel #gentoo-council
								        [21:57]
<Chainsaw> 2 sentences per item is all that's needed. You shouldn't have to go
	   digging like this.
<ulm> Chainsaw: will do, but give me a few days
<WilliamH> Ok, that's all I had. :-)
<ulm> OK
<ulm> next topic then
<ulm> Open bugs with council involvement
<Chainsaw> ulm: Sure, I don't mean "right this minute". I mean "in future".
<ulm> I see only bug 383467
<Chainsaw> Not the vote counts still?				        [21:58]
<Chainsaw> Let me address the elephant in the room.
<ulm> Chainsaw: what?
<Arfrever> ulm: You seemed to be not against e.g. *.svg. Each filename
	   extension needs separate voting.
<Chainsaw> Can anyone who isn't jmbsvicetto action this?
*** spiros (~andyspiro@gentoo/developer/spiros) has left channel
    #gentoo-council: "Konversation terminated!"
<ulm> Arfrever: we're done with topic 2				        [21:59]
<ulm> and no council member has spoken up for dohtml extensions
<scarabeus> Arfrever: that thing is overly complex and even if ulm acked it it
	    get enough nacks overall
<Arfrever> ulm: But council failed to properly discuss remaining features.
<Chainsaw> Arfrever: Council said "no to all" to the block that included
	   dohtml.
<scarabeus> so who is member of elections team, could someone of them fix the
	    bug?
<Chainsaw> Arfrever: Try in EAPI6.				        [22:00]
<ulm> who volunteers to ping jmbsvicetto about #383467?		        [22:01]
<scarabeus> ulm: technically he is staying at my place on the gentoo miniconf,
	    so i can try to beat him to it :P
<scarabeus> but it will be done on the meeting after next one :P
<Chainsaw> scarabeus: Lock him in a room until it's done?
<scarabeus> Chainsaw: naah, i will tell him the wifi pw only if he agrees to
	    do it right away					        [22:02]
<scarabeus> advanced chinese torture
<ulm> so, we're 10 minutes behind schedule
<ulm> only :)							        [22:03]
<Arfrever> I suggest to vote on DEPENDENCIES in EAPI="5", which was not voted
	   on by council yet.
<Chainsaw> Arfrever: Council said "no to all" to the block that included
	   DEPENDENCIES reform.
<Chainsaw> Arfrever: Try in EAPI6.
<ulm> Arfrever: it wasn't even on the list
<Arfrever> Chainsaw: That block did not include DEPENDENCIES.
<Chainsaw> Arfrever: I am not going to vote "No" all over again just to please
	   you. It was denied. Try in EAPI 6.			        [22:04]
<ulm> do you all agree so far with the summary in
      http://dev.gentoo.org/~ulm/council-20120911.txt ?
<ulm> before we go to open floor
<Chainsaw> ulm: Summary approved with no comments.		        [22:05]
<dberkholz> looks good
<Arfrever> ulm: Some descriptions are truncated.
<ulm> of course, I'll scan the log for additional comments that should be
      included							        [22:06]
<scarabeus> ulm: ++
<ulm> Arfrever: only one of them, actually
<Arfrever> ulm: 2
<ulm> should be corrected now
* WilliamH agrees with the summary
<ulm> Arfrever: right, two					        [22:07]
<ulm> I'll complete them for the final summary			        [22:08]
<ulm> next topic
<ulm> Open floor
* Chainsaw switches on the microphone
<Arfrever> ulm: Add {,package.}use.stable.{force,mask} in descriptions of 2
	   items.
<ulm> Arfrever: yes ;)						        [22:09]
<ulm> anything for open floor?
<Arfrever> I suggest that council change schedule of meetings so that meeting
	   is in each week.					        [22:10]
<Chainsaw> I think that grossly overestimates the availability of council
	   members.						        [22:11]
<WilliamH> Yeah I don't think weekly meetings would work well for the same
	   reason.
<dberkholz> given that there hasn't exactly been a glut of topics, i don't see
	    the demand
<scarabeus> we would never join together so often
<scarabeus> and the councils in past that did it didnt have much fun doing it
	    either						        [22:12]
<ulm> we can do an extra meeting between regular ones if there's need
<ulm> although there wasn't any need during this term, so far
<WilliamH> Right, there is nothing that stops us from having extra meetings if
	   we need them.					        [22:13]
<ulm> anything else for open floor?
<ulm> seems not							        [22:14]
<Arfrever> I would like that each council member write rationale for his
	   voting on each rejected feature.
<Chainsaw> What is the word limit on this essay?
<willikins> ulm: https://bugs.gentoo.org/383467 "Council webpage lacks results
	    for 2010 and 2011 elections"; Website www.gentoo.org, Projects;
	    CONF; hwoarang:jmbsvicetto
<scarabeus> ulm: what a fast bot :D				        [22:15]
<ulm> yeah :)
<Arfrever> Because it seems that they failed to understand some features
	   before voting.
<Chainsaw> Arfrever: Then I would suggest that you do not vote for this
	   council again next term.
<ulm> Chainsaw: +1
<Arfrever> Chainsaw: Several sentences of rationale per feature would suffice.
<scarabeus> Chainsaw: he is not developer :-)
<ulm> though he isn't allowed to vote ;)			        [22:16]
<Chainsaw> I wonder why.
<ulm> I think it's time to close this meeting
<Chainsaw> Yes, that would be good.
<ulm> thank you all for participating
<scarabeus> bye							        [22:17]
<Chainsaw> Thanks for hosting ulm.
<Chainsaw> See you later.
<chithead> [21:36:50] <Chainsaw> Okay for EAPI 5. *Nothing* gets applied
	   retroactively. *EVER* << not totally correct, old-style virtuals
	   were removed retroactively from old EAPIs
<WilliamH> My appologies for being late.
<Chainsaw> chithead: During my tenure?
<chithead> no
<Chainsaw> chithead: Precisely.
<Arfrever> chithead: Also retroactive changes are allowed in my EAPIs :) .
<ulm> next meeting will be October 9th				        [22:18]
<Chainsaw> ulm: I'll be there.
* WilliamH will be there