21:00:38 #startmeeting EPEL (2020-07-24) 21:00:38 Meeting started Fri Jul 24 21:00:38 2020 UTC. 21:00:38 This meeting is logged and archived in a public location. 21:00:38 The chair is tdawson. Information about MeetBot at http://wiki.debian.org/MeetBot. 21:00:38 Useful Commands: #action #agreed #halp #info #idea #link #topic. 21:00:38 The meeting name has been set to 'epel_(2020-07-24)' 21:00:40 #meetingname epel 21:00:40 The meeting name has been set to 'epel' 21:00:41 #chair nirik tdawson bstinson Evolution pgreco carlwgeorge 21:00:41 Current chairs: Evolution bstinson carlwgeorge nirik pgreco tdawson 21:00:43 #topic aloha 21:00:59 howdy 21:01:00 hi 21:01:15 Hi carlwgeorge 21:01:17 Hi pgreco 21:02:41 morning 21:03:29 Hi nirik 21:04:24 One more minute and then we'll get started 21:04:37 hello 21:04:46 Hi smooge 21:05:13 #topic Old Business 21:05:15 .epel 105 21:05:16 tdawson: Issue #105: Implement rhel module name/stream avoidance - epel - Pagure.io - https://pagure.io/epel/issue/105 21:06:03 My investigation hasn't been too fruitful. Most of my searches keep turning up end user information, instead of how to create bodhi tests. 21:06:58 might try the folks in #fedora-ci... ? 21:07:19 Although, I will admit that I only started my investigation today, and the people I know would know have already gone home. 21:08:16 nirik: oohh ... even just going to the page they have linked to gives me more than I found before. :) 21:08:40 thats something. :) not sure if it's enough, but something. ;) 21:08:52 I will try there ... not just the page, but the IRC channel. 21:09:10 but now of course you might hit the 'everyone is weekendward' 21:09:48 True, but I'm really hoping that next week I'll have more time for this. 21:11:09 If I get the test started, smooge nirik would you two know where I would be able to find a list of RHEL modules, that is accessable from the fedora infrastructure? 21:11:10 thank you for persuing this. :) 21:11:38 I'm not sure. we have pkgs... but I don't think we have a modules list. 21:11:50 I can make you a list 21:12:05 OK. 21:12:23 should be identical to CentOS list? 21:12:27 It might be that we have to update it with each release, but I was hoping we'd be able to automate it. 21:12:52 cyberpear: yeah, or at least very close... 21:13:20 that would certainly be an easier place to check. 21:13:43 currently the build system sees 42 modules 21:14:17 tdawson https://paste.centos.org/view/282e53df 21:14:39 basically what I am doing is logging into batcave, cd /mnt/fedora/app/fi-repo/rhel/rhel8/koji/latest/x86_64/RHEL-8-001 21:15:43 ls -1 | awk ' $0~/:/{split($0,a,":"); print a[1]}' | sort -u 21:17:04 note that I think rhel never removes older modules, but centos just ships the newest of each? 21:17:12 OK, I'll see what I can do with that information. ... I'll also need the version number, but I think that got trimmed off. 21:17:17 to see what streams they are... add a[2] so https://paste.centos.org/view/9da7913f 21:17:41 nirik: You are correct. 21:18:28 There is a bug open that I'm still watching, dealing with not being able to turn off, or not see, old RHEL modules. 21:19:28 But what we need is just the module name and stream name, and I think I can get that, either from smooge's way, or some way with centOS 21:21:11 This is all very good information, and I've got it written down. 21:21:39 Give me another week, and this week I believe I'll have the time, and hopefully I'll have at least a prototype. 21:22:10 Anything else on this before we move on? 21:23:11 #info EPEL-6 is End of Life in 2020-11. It will be moved to archives in 2020-12 21:23:12 #info THIS IS NOT A DRILL. 21:23:37 #topic EPEL-7 21:23:58 I didn't see any problems or issues for EPEL7 this week. Did anyone else? 21:24:18 nope 21:24:37 looks like I just came to say hello today :) 21:24:49 #topic EPEL-8 21:24:52 pgreco: Ha! 21:25:06 pgreco: But we're glad, even when you come to just say hello. :) 21:25:19 ;) 21:26:16 for epel8 I have a question actually 21:26:32 sure 21:26:53 is it still true that repoclosure doesn't work that well when there are modules? 21:27:08 today I noticed that we have appliance-tools, which I use 21:27:21 but dumpet was never released, and it is a hard dep 21:27:56 I'm not worried about that in particular, since I'm trying to talk to the people involved 21:28:07 but I'm worried we didn't catch it automatically 21:29:24 pgreco: So, possibly adding tests for packages in bodhi? Making sure they are installable before they go to stable? 21:29:42 no repoclosure does not work with modules 21:29:43 Instead of my manual tests ... which I haven't ran on RHEL8 yet. 21:30:31 I am not sure 'that well' is even a term to use with how repoclosure 'works' with modules 21:30:39 I don't know what tools we have available, but yeah, something like that 21:31:01 smooge: I didn't want to say "at all", but that was my understanding :) 21:31:38 when I was dealing with it for the EPEL-8 bringup.. it was 'its a hard problem' and 'someone will need to write a replacement someday' 21:32:19 my first repoclosure tests were "yum install \*" and see what breaks 21:32:22 There is a new "does it install" test going around in Fedora ... does that deal with modules correct? Or is it only working right because there are no default fedora modules? 21:32:59 I am going to expect 2 21:35:08 * tdawson just realized that he just done writting a program that checks if packages buildroot will install. It would be a very simple switch to have it look at regular packages. 21:37:16 well, looks like I didn't just come to say hello, I came to complicate people's lives... :P 21:37:47 We'd need to clean up all the packages that currently don't install ... but then I could see how well this does in bodhi, and we could put it in. 21:38:04 It doesn't use repocloser, rather it tries to install the package. 21:38:31 in the end, that is the more important task ;) 21:38:37 Yep 21:38:41 repoclosure is theory, 21:38:45 install is fact 21:39:02 Yep again 21:39:16 yeap, I've found cases that repoclosure being happy made me sad... 21:39:28 That's why I had to do it. All my repoquery and repoclosure stuff was jsut "amost" getting things right. 21:39:37 soname bumps tend to do that 21:40:08 https://memegenerator.net/instance/43977765/sexy-carl-sagan-thats-a-nice-theory-you-have-there-it-would-be-a-shame-if-someone-were-to-test-it 21:40:15 if the old version of the package is still available, repoclosure says it is all good, but then you try to install it and it fails 21:40:25 hehe 21:41:08 ha. 21:41:32 although really he would be happy if everyone tested it... 21:41:42 I do't mean to put more stuff on my plate ... but let me convert my script over, and see what sort of results I get this week. 21:41:59 sorry about that link.. i didn't know it also came with pop-ups 21:45:41 OK, I've got it on my list of things to do "convert buildroot-generator.py to will-it-install.py" 21:46:18 And, since I'll be dealing with bodhi, see how easy/hard it will be to make that a test. 21:47:03 tdawson: let me know if I can help with it, or if you want to push that script somewhere public for me to take a look 21:47:30 pgreco: Will do. 21:47:45 same here. I am down to only 10 hour days with the move 21:47:56 I've already got buildroot-generator.py on a public git repo, I can probrubly put the other on there as well. 21:50:43 I think that would be great, having help with it. I'm not the best python programmer, so I'm sure others can make it much better. 21:51:19 I think I told carlwgeorge a few days ago, my python is the google translate of my C thoughts 21:51:44 but that doesn't prevent me from doing stuff :) 21:51:44 * tdawson laughs 21:52:44 Looks like the time is going by, any other RHEL8 stuff? 21:53:14 not here 21:53:22 i got nothing 21:53:25 #topic General Issues / Open Floor 21:54:17 anything else for the open floor? 21:54:33 nothing here 21:55:22 thanks eveyrone 21:55:30 I got nothing 21:55:45 Yep, thanks everyone for being here, and especially thanks for the input. 21:56:11 We'll talk to you next week, or on IRC and email. 21:56:36 #endmeeting