From Fedora Project Wiki
* Jeff_S here, still sleepy though | 08:00 | |
-!- smooge changed the topic of #fedora-meeting to: EPEL Sig meeting -- Meeting rules at http://www.fedoraproject.org/wiki/Extras/Schedule/MeetingGuidelines -- Init process | 08:02 | |
smooge | MSG #fedora-devel Meeting ping: Everyone interested in EPEL -- EPEL meeting in #fedora-meeting now! | 08:02 |
---|---|---|
smooge | MSG #epel Meeting ping: Everyone interested in EPEL -- EPEL meeting in #fedora-meeting now! | 08:02 |
-!- fab [n=bellet@monkey.creatis.insa-lyon.fr] has quit ["Leaving"] | 08:03 | |
smooge | Hi everybody; who's around for the EPEL meeting? | 08:04 |
smooge | .... watches the worlds shortest EPEL meeting occur | 08:04 |
* inode0 comes and is all alone | 08:04 | |
smooge | sorry inode0 | 08:05 |
inode0 | so I may have plenty of mike time :) | 08:05 |
* Jeff_S here | 08:05 | |
smooge | I think the dayolight savings time is killing various humans | 08:05 |
smooge | mike time? | 08:05 |
-!- mccann [n=jmccann@66.187.234.199] has joined #fedora-meeting | 08:05 | |
inode0 | time to stand at the mike and talk | 08:05 |
smooge | ah ok .. I was thinking of all the other mike's I have known | 08:06 |
smooge | they usually are murdersome people | 08:06 |
smooge | ok status reports | 08:06 |
-!- smooge changed the topic of #fedora-meeting to: EPEL Sig meeting -- Meeting rules at http://www.fedoraproject.org/wiki/Extras/Schedule/MeetingGuidelines -- Status Reports | 08:06 | |
-!- hanthana [n=hanthana@124.43.75.4] has quit [Read error: 113 (No route to host)] | 08:07 | |
smooge | Build system: We are still on the plague system. Will be until the world shifts somewhere | 08:07 |
smooge | Builds pushed: David Gilmore made the push I think for November. We had several CVE's pushed this month for uw-imapd and some other items. | 08:08 |
-!- schlobinux [n=xavierb@nat/yahoo/x-9f6d552ac49e8de5] has joined #fedora-meeting | 08:08 | |
smooge | Orphans dealt with: Micheal and Michael are working out how to get the broken deps script working 100%. | 08:09 |
-!- Sonar_Gal [n=Andrea@fedora/SonarGal] has joined #fedora-meeting | 08:09 | |
-!- warren [n=warren@redhat/wombat/warren] has joined #fedora-meeting | 08:09 | |
smooge | Rebranding: Seems that few people really care about rebranding EPEL to something pronounceable or what image it conveys... so I think this is a dead item. | 08:11 |
-!- giallu [n=giallu@fedora/giallu] has joined #fedora-meeting | 08:12 | |
* smooge waits for his walkin to leave | 08:12 | |
-!- smooge changed the topic of #fedora-meeting to: EPEL SIG Meeting | Free discussion around EPEL | 08:12 | |
smooge | ok I think we can go over free discussion | 08:12 |
smooge | inode0, the mike's all yours.. | 08:13 |
-!- jnettlet_ [n=jnettlet@c-76-118-159-90.hsd1.ma.comcast.net] has quit ["Ex-Chat"] | 08:13 | |
inode0 | I thought EPEL was pronounced ep-el? | 08:14 |
-!- jnettlet [n=jnettlet@c-76-118-159-90.hsd1.ma.comcast.net] has joined #fedora-meeting | 08:14 | |
inode0 | more seriously I can volunteer a comment or two about the issue of including layered products in EPEL | 08:14 |
smooge | well there is E-pel e-pel ep-el Ep-el | 08:15 |
-!- hanthana [n=hanthana@124.43.53.11] has joined #fedora-meeting | 08:15 | |
smooge | what we need for layered products is how to do it a) safely, b) continuely, and c) ... | 08:16 |
inode0 | my instincts have always been against doing so as that will make EPEL less friendly to some of the customers it is aimed at, many of whom use some layered products | 08:16 |
-!- warren [n=warren@redhat/wombat/warren] has quit ["Leaving"] | 08:16 | |
inode0 | I've been rethinking that though listening to others who see it differently | 08:16 |
-!- warren [n=warren@redhat/wombat/warren] has joined #fedora-meeting | 08:17 | |
-!- JSchmitt [n=s4504kr@fedora/JSchmitt] has joined #fedora-meeting | 08:17 | |
* nirik is here now | 08:17 | |
inode0 | in a way, thinking abstractly, it seems to violate the spirit of the don't whack base RHEL packages rule | 08:17 |
smooge | I think the big issue is on how to help customers deal with repositories for their systems. I mean to some people everything could be seen as a 'layered' product over what they have. How to help them deal with this | 08:17 |
inode0 | but it will likely help a lot more users than it will hinder I think now | 08:18 |
-!- wwoods [n=wwoods@nat/redhat/x-fe8d8098e85eb162] has quit ["new kernel time"] | 08:18 | |
inode0 | but I also think some accommodation needs to be made for all the major layered products including those that trample on something from base | 08:19 |
-!- mdomsch [n=Matt_Dom@cpe-70-124-62-55.austin.res.rr.com] has quit ["Leaving"] | 08:19 | |
-!- sternecg [n=sternecg@ohnat.bristolwest.com] has left #fedora-meeting ["Leaving"] | 08:19 | |
smooge | yeah.. I think in some ways I can actually see a perverted logic to the mini-repos that the RHEL-5 was broken into | 08:19 |
smooge | so what I am trying to say is "I agree more with you than I did a week or so ago when it was brought up." | 08:20 |
inode0 | I'm not a big fan of separate repos if there is another way | 08:21 |
smooge | and I think you are saying "you agree with more what I was saying in #rhel" | 08:21 |
smooge | the big issue is I don't see a way to do it without seperate repos with yum technology. I guess skvidal would need to come in and let us know what is the right way | 08:22 |
inode0 | suppose package X tromps on a Red Hat layered product Y but not on any base RHEL packages | 08:22 |
inode0 | I don't really have a concern any more about that going into EPEL proper | 08:22 |
-!- wwoods [n=wwoods@nat/redhat/x-cfc6881def5ce981] has joined #fedora-meeting | 08:22 | |
-!- sonargal [n=Andrea@fedora/SonarGal] has quit [Connection timed out] | 08:23 | |
inode0 | I'm still concerned about package Z that tromps on a Red Hat layered product and some base package as well though | 08:23 |
nirik | so the change would be from 'no packages that are in RHEL anywhere in EPEL' to 'no packages that are in Base RHEL in EPEL' ? | 08:23 |
smooge | actually this is where I would love to see a 'pulp' solution. You take the upstream repos and you have a cool tool to mirror what you need and manage/let you know about breakage. | 08:23 |
smooge | nirik, its more like is Red Hat Directory Server RHEL? Is Red Hat IPA RHEL? ... | 08:24 |
nirik | any complicated solution is bad, IMHO... most people aren't going to tweak things carefully, they are going to say "hey, foo is in EPEL, let me just install it" | 08:24 |
nirik | yeah, what is "base RHEL" anymore... Client and Server? | 08:25 |
smooge | Client/Server/Virtual | 08:25 |
smooge | well except I think they have a further virtual product.. but I am not sure where genome fits into that | 08:26 |
inode0 | as a rhel user I would say anything in a base channel on RHN is base rhel | 08:26 |
nirik | so if you have a subscription you get everything? or only some channels? or ? | 08:26 |
smooge | you get Client/Server/Virtual | 08:26 |
inode0 | you don't get layered products, and you may or may not get what smooge just said | 08:27 |
smooge | pretty much a subscription gives you everything in CentOS and the Java rpms | 08:27 |
inode0 | some would get those though so they all need to be covered | 08:27 |
smooge | and if you get the layered products you might get stuff that completely replaces items in base. | 08:27 |
inode0 | that is the really sticky point in my mind | 08:28 |
nirik | yeah, messy. ;( | 08:29 |
smooge | which is where I would have to say that sort of layered product needs to go in a seperate repo if its going to be offered. I know that the RH-mrt stuff wouldn't go into EPEL because it replaces kernels and stuff | 08:29 |
inode0 | I haven't done a survey but I believe spacewalk would whack apache at least (the layered versions do anyway) | 08:30 |
-!- balor [n=balor@gimili.plus.com] has joined #fedora-meeting | 08:30 | |
smooge | it would replace apache and thats the problem in that spacewalk is one of the things people keep asking for in EPEL | 08:31 |
inode0 | rhel base could be redefined for our purposes though | 08:31 |
* nirik wonders why it replaces apache? | 08:31 | |
smooge | %base :) | 08:31 |
inode0 | rhel base == anything on RHEL installation DVDs that Red Hat layered products don't whack :) | 08:32 |
inode0 | nirik: mostly so it has squid and stuff configured I think | 08:32 |
smooge | nirik, I think the issue is tha tthe layered groups don't look at needing to back-port stuff. Their idea is that if you are going to use this, well you will want what we tested against. | 08:32 |
smooge | so the layered products usually are built against whatever X first had a feature they needed/wanted. | 08:33 |
smooge | apache, openldap, nss, etc | 08:33 |
smooge | which basically usually means you are almost assured that glibc is going to be the same, but anything else might have changed | 08:34 |
smooge | which is where putting layered products into the base of EPEL makes life hard since you could be building alpine and end up pulling in fedora-directory-server | 08:35 |
smooge | [not sure that would happen, but I was thinking that might be the nearest one.] | 08:35 |
inode0 | well, one suggestion was just a single split point | 08:35 |
smooge | ? | 08:36 |
inode0 | base EPEL as it now is and a layered-EPEL repo with all upstreams falling into the stomp on layered products and/or base packages | 08:36 |
smooge | not sure what a single split point means | 08:36 |
inode0 | that way at least users are aware of the slight elevation in risk for the layered products messing something up in base | 08:37 |
smooge | I was thinking that myself. Basically if a person wanted to develop something for EPEL the guide would be: If its in centos .src.rpms it has to be developed into a layered sub-project | 08:37 |
smooge | or one could go with Release @base -> CREEPINGCRUD-REPO -> {EPEL,EPEL+DS/IPA, EPEL+SPW, etc} | 08:39 |
* nirik has a DB server falling over... back when it's back working. | 08:39 | |
smooge | ok thanks nim-nim | 08:39 |
smooge | s/nim-nim/nirik | 08:39 |
inode0 | as a user I'd prefer the isolated repos to be honest, but understand that is more work and it is likely that many layered products will play nice anyway | 08:41 |
inode0 | and it might be more confusing to some as well | 08:41 |
inode0 | matching up RHDS to FDS and RHN Proxy Server to Spacewalk isn't always obvious | 08:42 |
smooge | yeah... that can be uhm interesting at times | 08:43 |
smooge | I think the DS one is easier to match up than Spacewalk with RHN :) | 08:44 |
inode0 | on the other hand using EPEL+Spacewalk for my satellite server is clean and gives me confidence no other layered stuff can creep in | 08:44 |
smooge | hmmm | 08:45 |
inode0 | clean from my user perspective, not your EPEL maintainer perspective necessarily | 08:46 |
inode0 | would it be that hard to have the release package drop in layered disabled repos? | 08:47 |
smooge | so in some ways its going to be working with the various upstreams on how they want to be 'included' or not into offerings. | 08:47 |
-!- alexxed [n=alex@dyn-86.105.67.31.tm.upcnet.ro] has joined #fedora-meeting | 08:47 | |
smooge | well I need to start winding this down. | 08:48 |
smooge | thankyou for your time inode0 | 08:48 |
smooge | it is good to have users who are not EPEL developers perspectives | 08:48 |
smooge | is there anything else you would like to add before I close/reset the meeting? | 08:48 |
-!- sdziallas_ is now known as sdziallas | 08:48 | |
* inode0 is done | 08:49 | |
smooge | thanks. I will get this posted to the the list by end of day | 08:50 |
Generated by irclog2html.py 2.6 by Marius Gedminas - find it at mg.pov.lt!