16:00:23 #startmeeting Go SIG meeting 16:00:23 Meeting started Tue Jul 9 16:00:23 2019 UTC. 16:00:23 This meeting is logged and archived in a public location. 16:00:23 The chair is jcajka. Information about MeetBot at http://wiki.debian.org/MeetBot. 16:00:23 Useful Commands: #action #agreed #halp #info #idea #link #topic. 16:00:23 The meeting name has been set to 'go_sig_meeting' 16:00:36 #topic Roll Call 16:00:39 * decathorpe is here as well 16:00:52 #chair decathorpe 16:00:52 Current chairs: decathorpe jcajka 16:00:57 * jcajka waves 16:03:02 * jcajka is on the move so please excuse if I drop out 16:04:01 .hello2 16:04:02 nim: nim 'None' 16:04:11 #chair nim 16:04:11 Current chairs: decathorpe jcajka nim 16:06:02 .hello qulogic 16:06:03 QuLogic: qulogic 'Elliott Sales de Andrade' 16:06:09 #chair QuLogic 16:06:09 Current chairs: QuLogic decathorpe jcajka nim 16:11:38 * jcajka_ is back 16:12:32 let me fire one question how is it looking with the status of rawhide? are we currently waiting on compose? 16:14:09 * jcajka_ is still seeing broken deps locally/on mirrors 16:14:45 I just fired off the builds to add Obsoletes on the last set of renames eclipseo did 16:15:02 so at the very least, upgrades won't be working just yet 16:15:35 some of my packages were probably renamed but haven't been retired (yet?) 16:16:04 I would wait for eclipseo for exact details though 16:16:33 thanks looks better ;), this will require successful compose too, not sure what status of them is, but at least buildroot in koji should be better now 16:16:56 right away 16:19:08 QuLogic: I just checked, there's still obsoletes missing for some of my renamed packages 16:19:29 decathorpe: such as? 16:19:52 for example, golang-github-AudriusButkevicius-go-nat-pmp got renamed to https://src.fedoraproject.org/rpms/golang-github-audriusbutkevicius-nat-pmp 16:19:53 I don't have a list; I'm just going by commit notifications 16:21:35 hmm, I recognize a bit of that name, but I don't have it here 16:22:00 the other three are: 16:22:05 decathorpe, QuLogic: IMHO it would be good to track those that look like "missed" in the BZ either in https://bugzilla.redhat.com/show_bug.cgi?id=1715534 or separate bugs so they don't get lost 16:22:24 ah, alright. I'll comment there as well 16:22:59 golang-github-oschwald-geoip2(-golang), golang-github-oschwald-maxminddb(-golang), https://src.fedoraproject.org/rpms/golang-github-sasha-s-(go-)deadlock are the other three 16:23:44 I don't seem to have an commit notification on those 16:24:49 once or twice, fedmsg gave up on fully sending notifications, so I may have not correctly cross-referenced the commits 16:25:00 yeah, fedmsg is not 100% reliable 16:25:02 https://bugzilla.redhat.com/show_bug.cgi?id=1715534#c10 16:27:03 QuLogic, what do you use to detect what needs an obsolete? 16:27:47 QuLogic, the code I posted last month during the meeting does a dnf repoquery --whatprovides 'golang(*)' --qf "#%{name}|%{source_name}|%{evr}|%{provides}" 16:28:16 QuLogic, and then you just need to check all the golang() provides on the new repo 16:28:28 QuLogic, are owned by the same package in the old repo 16:28:39 QuLogic, otherwise it needs an obsolete 16:28:48 yea, I could do that, but I need a compose for that 16:28:56 QuLogic, that should not depend on fedmsg 16:28:56 or downloading all the packages from the side tag 16:29:24 QuLogic, you don't need a compose, you can query directly the repo in koji 16:29:43 QuLogic, all you need to download is the koji repo metadata 16:30:23 sure, I can run it now to make sure nothing was missed 16:30:42 (of course that will only tell you what needs an obsolete, not if the obsolete has already been filled) 16:31:04 QuLogic: that would be awesome 16:31:13 the advantage is you have all the time between the package build and the compose to fix things 16:32:08 QuLogic, as a bonus dnf works on multiple repos, so you can querry as many old repos as you want to build the from list 16:32:48 decathorpe: ohh, you mean eclipseo hasn't pushed a retire at all; yea that's why I didn't add Obsoletes 16:33:30 exactly, these 4 packages were renamed but not retired 16:33:49 QuLogic, this kind of repoquery is also handy to detect when multiple packages try to provide the same golang(foo) 16:33:49 this should be done now 16:38:23 QuLogic: are you fine if I taf you with action item to clean up the rawhide after the side tag mere? 16:38:28 taf-tag 16:38:50 decathorpe: done 16:39:34 jcajka: yea ok 16:39:40 #action QuLogic will clean up missing obsoletes and retires after the side tag has been merged 16:39:42 thanks 16:41:09 QuLogic: great, thanks 16:41:16 should I retire the old packages? 16:41:56 ideally after the obsoletes are in place so there is update path 16:42:03 I'll do it when the builds pass 16:42:14 QuLogic: thanks 👍️ 16:43:14 wonderful 16:43:32 (of course obsoletes won't fix upstream API breaks) 16:46:29 eclipseo, QuLogic, nim: I have one thing that I would like to ask you, could you please remove most of the mentions of Go SIG in the https://fedoraproject.org/wiki/Changes/Adopt_new_Go_Packaging_Guidelines as this change is not endorsed/lead by Go-SIG, but just you three 16:46:42 .hello 2 16:46:43 eclipseo: Sorry, but you don't exist 16:46:49 .hello2 16:46:50 eclipseo: eclipseo 'Robert-André Mauchin' 16:46:53 #chair eclipseo 16:46:53 Current chairs: QuLogic decathorpe eclipseo jcajka nim 16:47:36 QuLogic:you are still seeieng errors?? 16:48:01 eclipseo: hmm, on what? 16:49:13 no nothing I misread 16:49:35 jcajka, you're not being constructive, and if we acted the same way WRT your activity, the SIG would soon cease to exist 16:50:05 nim: I would kindly ask you to not be aggressive and personal 16:50:57 jcajka, I didn't spend months working within the SIG, in meetings, day and night, for you to spam all Fedora channels to disown this work 16:51:16 we all have agreed on some rule set and you have transgressed that 16:51:38 jcajka, you're not being constructiv 16:52:06 jcajka, if you insist in looking at things only your way, two can play the game 16:53:07 jcajka, do you realise that you're creating single-handledly a FUD campain against the work others are doing in the SIG? 16:53:42 jcajka, do you think I don't have other things to do that pointing several times in the week to actual state in koji and other Fedora infra 16:53:55 nim: I would kindly ask you to be constructive and not personal 16:53:58 jcajka, every time you claim false things in Fedora channels? 16:55:11 jcajka, that's what I'm asking you to do 16:56:02 jcajka, and I'm not the one posting negative things on other SIG members work every time I have an occasion to 16:56:13 eclipseo, QuLogic: would you be so kind to do that? 16:58:06 I'm sorry to bring this, this late but at time the change hit FESCO I have been offline on PTO and haven't paid as much attention that I should have 16:58:44 nim: IMO deception shouldn't be something that is normal and acceptable in Fedora 16:59:17 you should ask eclipseo; I'm only on there because of fixing golist 16:59:41 but IMO we're discussing it during Go SIG meetings, so the distinction seems moot 16:59:50 I don't aggree with this, the change is the SIG's work and wouldn't have happened without it. I don't understand why you are making difficulties about this. 17:00:08 jcajka, like when you claimed we broke something that was not even shipped in Fedora? I had to post the pointer to koji state for FC30 17:00:13 The work is done anyway so this is a moot point 17:00:14 nim: I don't wan in any extend down play the magnitude of work that you have done, but we haven't got consensus on the change 17:00:26 eclipseo: seems like new k8s broke prometheus :( 17:00:32 want 17:00:46 jcajka, or when you claimed we were blocking you ? I had to point you were admin of the git containing the files you "could not" change 17:01:10 nim: I have not in the context that you are describing 17:01:15 jcajka, it is getting quite tiring 17:01:23 jcajka, and you've been exalating lately 17:01:48 eclipseo: I disagree if this should be norm that I can propose anything by your name 17:02:59 We didn't porpose anything by your name and I don't understand why making a fuss of such a trivial point 17:03:30 eclipseo: sorry, but you did, original proposal has had my name on it 17:03:46 QuLogic: have a link? I have yet to rebuild the final piece of k8s because random error during build 17:04:08 eclipseo: https://koji.fedoraproject.org/koji/taskinfo?taskID=36152701 17:04:44 _build/src/github.com/prometheus/prometheus/discovery/kubernetes/client_metrics.go:239:23: cannot use f (type *clientGoWorkqueueMetricsProvider) as type workqueue.MetricsProvider in argument to workqueue.SetProvider: *clientGoWorkqueueMetricsProvider does not implement workqueue.MetricsProvider (missing NewDeprecatedAddsMetric method) 17:05:07 jcajka: because I thought you were on board with us, now it's removed. I don't understand your persistence of not wanting to be associated with anything we do 17:05:23 (I assume this is k8s anyway since it's the main upgrade) 17:05:50 eclipseo, QuLogic: could you please leave the k8 technical discussion after the meeting 17:06:09 anyway, back on topic, eclipseo how much is left to build? just apps? 17:07:51 eclipseo: nothing personal, just technical, in my opinion the change has rather under valued its extend and scope 17:09:00 and it has been rather ignore when I have been pointing it out 17:09:16 QuLogic: left to build K8s compose on k8s, Unboostrapped containerd, all binary apps packages 17:10:35 some packages are still bootrstrapped too because Unbundling them completly would have taken too much effort, so k8s for example would need some work at a later time to be fully unbootstrapped 17:10:53 eclipseo, QuLogic: can I please get from you answer, will you remove it or not? 17:16:28 like I said, it's not my writeup 17:18:59 anyway we are well on top of the hour 17:19:34 QuLogic, eclipseo: really don't get me wrong, I very much appreciate all the hard work that you have done on the execution of the change 17:19:39 jcajka: I'll remove 17:20:02 eclipseo: you need any help with the rest? 17:20:11 QuLogic: updating to 2.11 should fix the compatibility with k8s 17:20:22 eclipseo: I very much appreciate that 17:20:54 yes I don't have time to do the rest, cause new work so I'll appreciate if you could look into it 17:21:12 do you have a list of remaining items? 17:21:18 I need to mass assign the new package to the SIG too and add them to Koschei 17:21:22 FYI there has been new point release of Go, I have just landed go1.12.7 in rawhide 17:21:48 any more topic or stuff that should go on record? 17:21:55 topics 17:22:16 I'll also need to double check if I haven't missed any library as I worked from a Grokmirror clone which is a bit unreliable 17:23:13 eclipseo, QuLogic are you fine with ending meeting now? 17:23:18 QuLogic: any stuff that don't start withh golang- on that list https://fedoraproject.org/wiki/Changes/golang1.13#Dependencies 17:23:26 jcajka: ok 17:23:43 jcajka: yep 17:24:17 #endmeeting