[CentOS-devel] Log from CentOS SCLo SIG sync-up meeting on #centos-devel (2015-08-05)

Honza Horak hhorak at redhat.com
Wed Aug 5 16:25:22 UTC 2015

(05:06:27 PM) hhorak: #startmeeting SCLo SIG sync-up (2015-08-05)
(05:06:27 PM) centbot: Meeting started Wed Aug  5 15:06:26 2015 UTC. 
The chair is hhorak. Information about MeetBot at 
(05:06:27 PM) centbot: Useful Commands: #action #agreed #help #info 
#idea #link #topic.
(05:06:31 PM) hhorak: #meetingname sclo-sig-syncup
(05:06:31 PM) centbot: The meeting name has been set to 'sclo-sig-syncup'
(05:07:10 PM) The account has disconnected and you are no longer in this 
chat. You will automatically rejoin the chat when the account reconnects.
(05:08:03 PM) The topic for #centos-devel is: CentOS Developers and 
Contributors room | If you have questions regarding CentOS please see 
(05:08:03 PM) Topic for #centos-devel set by 
z00dax!~kbsingh at chakra.karan.org at 09:26:30 AM on 12/11/2012
(05:08:08 PM) The account has disconnected and you are no longer in this 
chat. You will automatically rejoin the chat when the account reconnects.
(05:08:45 PM) The topic for #centos-devel is: CentOS Developers and 
Contributors room | If you have questions regarding CentOS please see 
(05:08:45 PM) Topic for #centos-devel set by 
z00dax!~kbsingh at chakra.karan.org at 09:26:30 AM on 12/11/2012
(05:09:08 PM) hhorak1: hm, I lost connection shortly after starting the 
meeting, so might miss if anybody already waived :)
(05:10:00 PM) ***langdon lurks
(05:10:06 PM) DV [~veillard at 2001:41d0:1:d478::1] entered the room.
(05:10:19 PM) ***lalatenduM _o/
(05:10:23 PM) alphacc: hhorak: o/
(05:10:30 PM) jstribny: hi
(05:10:41 PM) bstinson: -_o_-
(05:10:53 PM) hhorak1: langdon: great you're here :) I've added the 
prefix thing on the agenda... :) so let's start with the whole vagrant 
(05:10:59 PM) hhorak left the room (quit: Ping timeout: 252 seconds).
(05:11:04 PM) You are now known as hhorak
(05:11:10 PM) jstribny: yes please
(05:11:16 PM) lalatenduM: :)
(05:11:39 PM) langdon: hhorak, you already know what I am gonna say :)
(05:11:48 PM) jstribny: so we need to build rh-ruby22 and rh-ror41 and 
then sclo-vagrant1
(05:12:06 PM) langdon: jstribny, we should already have the ruby and 
ror, no?
(05:12:12 PM) jstribny: where?
(05:12:15 PM) ***langdon means "in cent scl sig"
(05:12:39 PM) hhorak: So, we want vagrant in centos and since we're 
gonna build rh-ruby22 and rh-ror41 collections there, it makes sense 
having vagrant as SCL as well in centos (comparing to normal RPM)
(05:12:59 PM) hhorak: langdon: we have it only in copr, same as other scls
lalatenduM Lalufu langdon larsks lars_kurth
(05:13:20 PM) langdon: hhorak, ohh.. sorry.. i thought they had come 
over already.. i get my dates confused some times :)
(05:13:26 PM) hhorak: langdon: but the plan is to build all the 
collections and keep their source in centos, so other centos projects 
can use it
(05:13:34 PM) langdon: hhorak, +1
(05:13:44 PM) langdon: with links from scl.o would be nice too
(05:14:17 PM) hhorak: langdon: right, that's also on the todo list, just 
a bit more down :)
(05:14:17 PM) lalatenduM: hhorak: +1
(05:15:20 PM) jstribny: is there some semi-automatic rebuild of scls 
(05:15:26 PM) jstribny: i mean like the copr one
(05:15:42 PM) dougsland [~dougsland at tchelinux/moderator/dougsland] 
entered the room.
(05:16:54 PM) hhorak: jstribny: yes, we have already the packages from 
scl.org imporeted into cbs.koji.org, so we shouldn't have issues with 
bootstrapping.. and should be able to do the builds almost automatically
(05:17:01 PM) hhorak: (hopefully fully automatically)
(05:17:41 PM) alaird [~alan at] entered the room.
(05:17:46 PM) hhorak: so, since vagrant is not the rh collection, it 
will be the first collection as upstream only -- that means it will not 
use rh- prefix.. but which prefix then? have we already discuss this?
(05:17:56 PM) hhorak: sclo- seems like one option...
(05:18:32 PM) jstribny: hhorak, i dont really like prefixes at all to be 
(05:19:14 PM) hhorak: jstribny: yeah, you're not alone, that's what we 
have langdon, who can nicely explain why they are no evil
(05:19:15 PM) hhorak:  :)
(05:19:38 PM) hhorak: #link 
(05:19:49 PM) hhorak: (just if anybody wonders which prefixes we talk about)
(05:20:41 PM) langdon: sorry.. was talking in my other call
(05:21:08 PM) hhorak: from my PoV we simply need them to distinguish 
vagrant SCL that includes all plugins available in the world (e.g. build 
in centos upstream) and vagrant with only some necessary plugins 
available in RH once...
(05:21:54 PM) nshaikh [nshaikh at nat/redhat/x-gojnfqdoasqifhss] entered 
the room.
(05:22:03 PM) langdon: so.. in short .. the "rh-" prefix indicates that 
"red hat peopl decided that these are the right things to put in to a 
vagrant collection" .. vs "cox- = company x decided these are the right 
things to put in the vagrant collection"
(05:22:08 PM) hhorak: This issue is more relevant in python collections 
where also, modules are either available or not... so the set of 
plugins/modules kind of defines what the software collections provides..
(05:22:59 PM) jstribny: i see, but then we should stick with rh-
(05:23:01 PM) langdon: it does not (necessarily) imply that rht supports 
it, endorses it, etc..
(05:23:16 PM) hhorak: another reason is to avoid conflicts we already 
saw (mysql55 package and mysql55 SCL -- both rpms, but every contains 
something different)
(05:23:31 PM) langdon: it could as easily be jstribny- (except you need 
a shorter name and a llanal registration ;) )
(05:23:36 PM) masta left the room (quit: Ping timeout: 244 seconds).
(05:24:29 PM) lars_kurth left the room (quit: Remote host closed the 
(05:24:53 PM) lars_kurth [~lars_kurt at 97e55056.skybroadband.com] entered 
the room.
(05:25:08 PM) hhorak: langdon: wait, do we need register sclo-? that's 
just part of the name for the collection...
(05:25:51 PM) jstribny: langdon, hhorak well i dont think vagrant1 would 
conflict with anything in centos (as opposed to other possible scls in 
(05:26:10 PM) alphacc: to not confuse (more) people I'll stick to 
rh/sclo at this time :)
(05:26:18 PM) langdon: hhorak, technically, we are supposed to.. that 
was the rule.. (and it would be just sclo, but i figure you meant 
that).. but... i get shaky on sclo- prefix because it implies that it is 
the canonical scl for that scl.. which is not necessarily a good implication
(05:27:04 PM) ***lalatenduM is already confused :-)
(05:27:07 PM) jstribny: i am personally inclined to use rh-vagrant1
(05:27:17 PM) langdon: jstribny, the point is not if it conflicts now.. 
it is if it conflicts in the future or if someone is doing a "secret" 
one now that they want to push to public later
(05:27:46 PM) ***lalatenduM prefers centos inplace of rh. not sure if it 
makes sense
(05:28:37 PM) jstribny: langdon, i know but then again if someone do 
unprefixed scl.....why that someone should not be us? :))
(05:28:54 PM) hhorak: jstribny: I'm afraid that rh-vagrant1 would imply 
we need to ship/support all the plugins in rhscl (once vagrant is 
becoming part of that, if at all) that we have in centos
(05:29:13 PM) langdon: jstribny, no un-prefixed scls.. is the recommendation
(05:29:32 PM) langdon: hhorak, i disagree but understand your point..
(05:29:45 PM) langdon: half-kidding, epel-vagrant1 ?
(05:29:59 PM) hhorak: langdon: now even I am confused :)
(05:30:00 PM) langdon: or rh-epel-vagrant1
(05:30:37 PM) jstribny: hhorak, sclo should be always used for upstream?
(05:30:38 PM) hhorak: I think we need some generic prefix usable also 
for collections that will never ever have anything common with rh
(05:30:49 PM) langdon: i don't think an rh- prefix should ever indicate 
that it will be in rhscl.. just that rht employees were the leads on its 
definition.. but I understand why that may not be obvious to everyone
(05:30:51 PM) hhorak: jstribny: that was how I meant it
(05:31:05 PM) jstribny: then i believe this would be upsteam
(05:31:12 PM) langdon: hhorak, that is why i like something like epel- ...
(05:31:23 PM) ***RemiFedora uses "more" (e.g. more-php56) for packages 
extending "rh" SCL (rh-php56)
redkilian reetspissedoff reLad RemiFedora
(05:32:03 PM) DV left the room (quit: Remote host closed the connection).
(05:32:04 PM) langdon: just the problem is it *isn't* upstream.. it is 
just one option.. it is not the *only* rpm-scl-ified vagrant 
collection.. as it would be if it was a "real" rpm..
(05:32:06 PM) dcaro is now known as dcaro_mtg
(05:32:06 PM) hhorak: RemiFedora: that's also something we should do the 
(05:32:18 PM) hhorak: (in other similar extending collections)
(05:33:18 PM) ***hhorak wondering whether kbsingh is round, he's always 
liked the prefix discussion :)
(05:33:22 PM) alphacc: I think we did agreed in term of sclo rebuild 
starting with rh- should be what is shipped by redhat scls. same 
relation centos <-> rhel.
(05:33:22 PM) langdon: RemiFedora "should" be using something like 
rh-php56-more though..  because it is *still* based on the "rht defined 
idea of the php collection"
(05:33:39 PM) jstribny: langdon, but that implies there is never upstream
(05:33:45 PM) langdon: jstribny, exactly
(05:33:59 PM) RemiFedora: langdon, your wrong, rh-php56 take upstream 
from "remi" repo ;)
(05:34:36 PM) langdon: RemiFedora, ohh.. i was thinking of your "extra 
packages" that are not in rhscl..
(05:35:07 PM) RemiFedora: langdon, in "remi" you have eveything, then 
"rh" pull some, and the others go to "more"
(05:35:14 PM) jstribny: langdon, but i create vagrant1, i made the 
project, i am upstream....you can still do my-own-vagrant-scl as 
alternative vagrant collection.....
(05:35:17 PM) langdon: i think RemiFedora's version should just be 
called rh-php56 which is the upstream of rh-php56 productized in rhscl 
(if that is what it is)
(05:35:43 PM) langdon: jstribny, right.. hence.. you could use jstribny-
(05:35:52 PM) RemiFedora: langdon, not possible, I already have pĥp70 
SCL, I don't know yet if it will go to rh ...
(05:36:41 PM) jstribny: langdon, but this would be CentOS official scl 
so probably not jstribny- (although i would love to be so famous)
(05:37:12 PM) langdon: i am not sure why this is so hard to explain.. 
but.. the short version is.. i think we should be careful to not "claim 
namespace territory" that we don't intend to claim.. so. i really don't 
care that much :) and am happy to say "call it whatever" just be aware 
of the implications being made...
(05:37:37 PM) erdic left the room (quit: Ping timeout: 246 seconds).
(05:37:39 PM) hhorak: hm, I don't think we should use names in centos, 
we should find nice identification of the whole group instead (I still 
like sclo the most)
(05:38:16 PM) erdic [~erdic at unaffiliated/motley] entered the room.
(05:39:00 PM) jstribny: i am fine with sclo
(05:39:24 PM) lars_kurth left the room (quit: Ping timeout: 264 seconds).
(05:40:40 PM) avij left the room (quit: Ping timeout: 258 seconds).
(05:40:41 PM) avij [~anssi at leopardi.miuku.net] entered the room.
(05:40:54 PM) ***langdon abstains :)
(05:43:01 PM) hhorak: so, let's move on, I confess I'm biased but sclo 
seems like least problematic :)
(05:43:19 PM) hhorak: so let's go with sclo at least for vagrant...
alaird alivigni alphacc
ksb kstange
(05:45:39 PM) JHogarth left the room (quit: Quit: JHogarth).
(05:45:41 PM) hhorak: Another point on the agenda was mainly about 
https://bugs.centos.org/view.php?id=9151 -- I was going to ask kbsingh 
whether this ticket is fine or some information is missing..
(05:45:41 PM) hhorak: but since he doesn't seem to be around, alphacc or 
bstinson -- anybody involved in dist-git repos creating?
(05:46:21 PM) hhorak: I was also thinking about creating request for 
other components from rhscl at once, there's basically no need to wait 
with that..
(05:46:36 PM) hhorak: so if nobody stops me, I'll do it tomorrow :)
(05:47:23 PM) DV [~veillard at 2001:41d0:a:f29f::1] entered the room.
(05:47:55 PM) asad_ [~asad_ at c048.fem.tu-ilmenau.de] entered the room.
(05:48:30 PM) bstinson: hhorak: are these the final dist-git repos that 
you need? i.e. the ones that go in the RPMS/ project?
(05:48:32 PM) jstribny: please do
(05:48:53 PM) jstribny: but for vagrant you probably need a list of 
components right>
(05:49:49 PM) xaeth is now known as xaeth_afk
(05:51:26 PM) bstinson: if they're the staging ones you should be able 
to create those repos under the sclo project yourself
(05:51:29 PM) hhorak: bstinson: these should still be under sclo/ 
project and the plan is still build from srpm.. we can't build from 
rpms/ (i.e. from scm) until we have lookaside cache working, right?
(05:51:38 PM) hhorak: jstribny: yes, can you send me the list, please?
(05:52:37 PM) lalatenduM: I would suggest a github organisation as of 
now till we have the dist git working
(05:54:30 PM) lalatenduM: or whatever u guys think is right :)
(05:54:44 PM) sbonazzo left the room (quit: Quit: Leaving.).
(05:54:56 PM) gamename [~gamename at c-67-177-166-184.hsd1.tn.comcast.net] 
entered the room.
(05:55:00 PM) jstribny: hhorak, http://etherpad.corp.redhat.com/V4kXtF6ixQ
(05:55:00 PM) hhorak: bstinson: just tried to create sig-sclo/Judy and 
received 'your personal repository must be located at "~hhorak"'
(05:55:11 PM) hhorak: jstribny: thx
(05:55:30 PM) hhorak: bstinson: I must do something wrong
(05:55:53 PM) langdon: lalatenduM, https://github.com/sclorg ?
(05:56:08 PM) xaeth_afk is now known as xaeth
(05:56:18 PM) lalatenduM: langdon: yeah , that sounds fine
(05:56:25 PM) langdon: lalatenduM, it exists :)
(05:56:48 PM) bstinson: hhorak: ah, something going on with permissions 
then. that's a kbsingh question
(05:57:02 PM) hhorak: bstinson: ok, will ask him
(05:58:39 PM) hhorak: lalatenduM: langdon: dist-git should be actually 
ready, at least the sig-sclo ... so I don't think we must do it (commit 
sources) twice, once to github, then to git.c.o
(05:58:41 PM) oere left the room (quit: Quit: Leaving.).
(05:58:57 PM) lalatenduM: hhorak: agree
(05:58:59 PM) langdon: hhorak, ack.. i was just answering lalatenduM's q
(05:59:15 PM) hhorak: great :)
(05:59:33 PM) hhorak: ok, anybody else something to discuss today?
(05:59:58 PM) hhorak: (I have some new todo items, so will sleep well :) )
(06:00:06 PM) alphacc: hhorak: drop me the full list of collections when 
you can.
(06:00:24 PM) hhorak: alphacc: just the collections or all the packages 
as well?
(06:00:48 PM) alphacc: hhorak: just the collections that need tags in 
koji :)
(06:01:13 PM) hhorak: alphacc: ok, will do.
(06:02:16 PM) hhorak: so, closing in a minute :)
(06:02:25 PM) alphacc: hhorak: what's the plan for rhscl 6 rebuild at 
this time ?
(06:02:39 PM) alphacc: (last one and I let you go:)
(06:03:25 PM) langdon: hhorak, and, from me, do we have a rough eta on 
any of this? I use the vagrant scl a fair amount and would like updates :)
(06:04:29 PM) alphacc: Becasue we rebuild rhscl at work and if it does 
not appear soon I'll have to kick off a rebuild ; however I'll prefer to 
do it part of sclo.
(06:05:14 PM) hhorak: alphacc: hm, no plans for now, we said we start 
the proof of concept builds on rhel-7 only... but the PoC was kind of 
done and it might be more effective to do it at once right?
(06:05:32 PM) jstribny: langdon, once ruby scls are rebuilt and i have 
access to branches, i could build it in two-three days
(06:05:37 PM) alphacc: hhorak: at once meaning when we have dist-git ?
(06:06:42 PM) alphacc: hhorak: or we can consider a srpm rebuild to have 
something quickly ?
(06:07:00 PM) shaunm left the room (quit: Ping timeout: 244 seconds).
(06:07:03 PM) hhorak: alphacc: I meant building rhel6 builds kind of 
together with rhel7 ones..
(06:07:10 PM) sayan [~sayan at fedora/sayan] entered the room.
(06:08:03 PM) hhorak: alphacc: yeah, I think we can have srpm builds for 
rhel6 as well, I don't see why we couldn't..
(06:08:29 PM) gsanchietti left the room (quit: Quit: Ex-Chat).
(06:08:48 PM) alphacc: hhorak: ok, ping me if it's a matter a sending 
srpm in correct order I can help.
(06:09:27 PM) hhorak: langdon: I'm always too positive about eta, so if 
I say having vagrant for rhel7 built till end of august.. I may be too 
positive again.. or might not..
(06:09:50 PM) mdittmeier [~mdittmeie at ip68-108-96-112.lv.lv.cox.net] 
entered the room.
(06:09:57 PM) langdon: hhorak, lol
(06:10:10 PM) hhorak: alphacc: if we use the same way (import scl.org 
packages into cbs), which we probably must anyway... the order should be 
(06:11:03 PM) alphacc: hhorak: if you use the bootstrap repo everywhere
(06:12:41 PM) hhorak: alphacc: do you think it is problem? I thought we 
could use it everywhere to save troubles with bootstrapping and circular 
dependencies.. these are not scm builds anyway :)
(06:15:18 PM) alan_ [~alan at 93-139-204-195.adsl.net.t-com.hr] entered the 
(06:15:41 PM) alan_ is now known as Guest11649
(06:17:16 PM) talori left the room (quit: Quit: Leaving.).
(06:18:24 PM) jstribny left the room (quit: Quit: Leaving).
(06:18:54 PM) apevec left the room (quit: Ping timeout: 260 seconds).
(06:19:37 PM) davidep left the room (quit: Quit: Leaving.).
(06:19:53 PM) alphacc: hhorak: not a problem but I'd like to rebuild 
against what we have at a point
(06:20:40 PM) hhorak: alphacc: fine with me, that should be just another 
round of scripted builds... theoretically :)
(06:21:14 PM) alphacc: hhorak: again just my opinion run it thorugh 
everybody :)
(06:21:17 PM) Guest11649 left the room ("Leaving").
(06:21:49 PM) ***alphacc &
(06:21:53 PM) alphacc: thanks all
(06:22:05 PM) hhorak: so, I'll prepare the list of scl.org repos for 
rhel6 and we can create one testing rhel6 scl first.. sounds fine?
(06:22:14 PM) hhorak: I'll need to leave as well, thanks all !!!
(06:23:00 PM) hhorak: #endmeeting
(06:23:00 PM) centbot: Meeting ended Wed Aug  5 16:22:59 2015 UTC. 
Information about MeetBot at http://wiki.debian.org/MeetBot . (v 0.1.4)
(06:23:00 PM) centbot: Minutes: 
(06:23:00 PM) centbot: Minutes (text): 
(06:23:00 PM) centbot: Log: 

More information about the CentOS-devel mailing list