19:00:44 #startmeeting Go SIG meeting 19:00:44 Meeting started Mon Jan 3 19:00:44 2022 UTC. 19:00:44 This meeting is logged and archived in a public location. 19:00:44 The chair is jcajka. Information about MeetBot at https://fedoraproject.org/wiki/Zodbot#Meeting_Functions. 19:00:44 Useful Commands: #action #agreed #halp #info #idea #link #topic. 19:00:44 The meeting name has been set to 'go_sig_meeting' 19:00:55 Hi there 19:01:02 Hello and welcome to the first Go sig meeting of 2022 19:01:08 #topic Roll Call 19:01:19 #chair alexsaezm 19:01:19 Current chairs: alexsaezm jcajka 19:01:24 Happy new year everyone :) 19:04:34 it seems we are just two here 19:05:03 * mroche lurks in the bg 19:06:29 Happy new year! 19:06:44 Eighth_Doctor, copperi[m], gotmax ping 19:06:49 mroche[m], same :) 19:06:54 Oh wow. I'm actually here for one. 19:07:20 Hopefully more will pop in, there seems to be FESCO meeting going on in parallel. 19:07:49 There are https://pagure.io/GoSIG/go-sig/issues two tagged issue for the meeting. 19:08:36 any preference on which to start with? 19:09:15 can we start with the #36 while I ping dbenoit? 19:09:31 sure 19:10:03 #topic Fedora 36 changes https://pagure.io/GoSIG/go-sig/issue/36 19:10:44 there has been question related to the go 1.18 change on the last meeting 19:10:57 oh sorry I wasn't able to attend 19:11:16 there has been few breakages after the introduction of go1.18 beta in to the rawhide 19:11:33 one was golist, thing that Go macros are using 19:11:45 AFAIK rebuild fixed it 19:12:02 and there is one BZ, let me look it up 19:12:13 * Eighth_Doctor waves 19:12:20 * Eighth_Doctor is in fesco and this meeting at the same time 19:13:05 https://bugzilla.redhat.com/show_bug.cgi?id=1980193 is the BZ affected by the update 19:13:33 oh right I saw the needinfo flag, still need to check it 19:14:43 * alexsaezm is doing a quick check 19:16:08 Side note on that: the currently changes/golang1.18 lists reverting to 1.16.x as the contingency, whereas the ticket says 1.17 19:16:26 yes 19:17:05 my initial idea was to revert to 1.16 because 1.17 never landed on an stable release but as people told me in the devel list, it should be fine to revert to 1.17 because it has been for a while in rawhide 19:17:13 I should update the proposal tho 19:17:19 thanks for the reminder 19:17:33 :thumbsup: 19:18:10 * alexsaezm can't find the package repository :/ 19:19:16 wait the package is still unpublished, am I correct? 19:19:32 yes it is review 19:21:57 hmmmm I'll take a look but the package seems to be unmaintained 19:22:07 I mean the last update is from May 2021 19:22:14 anyway I'm not sure if that I have covered all that has been risen last meeting 19:23:14 it is possible that it is a bug in the package, not uncommon with new version of Go uncovering some issues, but it goes both ways :) 19:24:34 I think there has been some bug fixes in the cryptography part of 1.18 so I'll take a look at it 19:25:32 thank you for taking look at it 19:25:47 jcajka, I was checking the last meeting logs and I don't see anything else apart from what we already talked 19:26:12 ack, it seems to me the same way :) 19:26:23 I guess we can move to the next topic 19:26:38 #topic EPEL 9 https://pagure.io/GoSIG/go-sig/issue/37 19:26:59 I sent a message to dbenoit because I cc'ed him in that issue but seems to be public holiday in the USA so he is not online :( 19:27:29 but I know he is talking about the issue in an email so I guess we can wait there 19:27:39 reading the ticket I think update in ticket can work. What do you think? 19:28:15 do you mean writing an update of the current state in the ticket? 19:28:55 yes 19:29:18 Sure, will do, I'll try to clarify the current state of the situation 19:31:06 thank you 19:31:30 are there any questions, comments to this topic or can we move to open floor? 19:32:31 not from my side 19:34:43 #topic Open Floor 19:35:16 Hopefully a quick one: should the deadline for the Fedora beta freeze approach and a new 1.18 beta/rc/ga not drop, should an updated build of beta1 be performed to fix my incorrect versioning? 19:35:18 https://src.fedoraproject.org/rpms/golang/pull-request/18 19:36:19 Assuming 1.18 is greenlit for F36. 19:36:38 mroche[m]: could you elaborate? 19:36:40 oh right that happened to me before and I always forget about the naming in the betas :( 19:37:17 oh I see it 19:37:23 AFAIK we cannot remove the build but chances are, we are going to have a beta2 soon 19:37:31 so we can fix it in the next beta build 19:38:27 :thumbsup: sounds good! I looked through the Go archive list and some versions had multiple betas, whereas others went from b1 to rc, so I wasn't sure about the timeline this time around, considering the GA and freeze are around the same time. 19:38:38 oh... no I'm afraid that this will require epoch bump 19:38:51 as 1.18beta1-1.fc36 > 1.18-1.fc36 19:39:04 or some other creative version 19:39:10 versioning 19:39:40 why is that? 19:40:03 rpm version rules 19:40:13 Is this not accurate then? "This might be fine for Rawhide where old builds are dropped from composes, but would cause trouble if the final release is not out early enough and 1.18beta1 ends up in the Beta." 19:41:16 I think here it would be good to ask rel-eng, what would be best solution to fix this 19:41:43 for the record the comparison has been obtained via "rpmdev-vercmp" 19:41:54 https://docs.fedoraproject.org/en-US/packaging-guidelines/Versioning/#_handling_non_sorting_versions_with_tilde_dot_and_caret 19:43:20 but that doesn't help with current issue 19:43:27 but shouldn't it be 1.18.0-1 the very first stable release? 19:44:09 hm.. that might work 19:44:17 Hypothetically, if B2 came out today and was built with the proper 1.18~beta2 versioning, and nothing until the freeze, would B2 be the build to make it into the freeze compose? I've never had to think about this before, so I don't know how it's handled. 19:44:30 rpmdev-vercmp 1.18beta1-1 1.18.0-1 19:44:30 1.18beta1-1 < 1.18.0-1 19:44:57 * alexsaezm never understands how to use rpmdev-vercmp 19:45:16 alexsaezm: that could work, I have always stick the upstream versioning go1.18 19:45:25 Only diff is upstream just doesn't like using .z release for initial GA. 19:45:51 mroche[m]: not really as koji/dnf/... would keep using the higher NVR 19:46:05 AFAIK^^ 19:46:15 alexsaezm: but that is not dogma 19:49:26 oh I see 19:49:37 so it seems that wayout will be to use 1.18.0 for the GA version 19:49:49 you mean that they are going to release 1.18 and then 1.18.1 19:50:00 yes 19:51:23 from .gitignore in golang repo, go1.16.src.tar.gz for example, upstream has been usually omitting the zero 19:51:36 in GA release 19:52:50 right right.. hmmmm I guess is not the end of the world to use the .0 for referencing the GA (personally I would prefer that they release X.Y.0, I find it more pleasant to the eye :D ) 19:53:08 :) 19:54:38 I have another question if this is kind of sorted out 19:55:08 go ahead 19:55:15 https://src.fedoraproject.org/rpms/golang/pull-request/19#request_diff 19:55:39 do you know if is it still relevant to the 1.18? 19:56:25 alexsaezm: should be, at least re-enabling tests on ppc64le, I will rebase it 19:57:00 thanks a lot 19:57:36 From my side I would like to mention the epel7 golang. As I have mentioned I'm rather time constrained nowadays, and I have never truly maintained Go in EPEL. It would be much appreciated if anyone would help there, update should be strait merge from 1.17/1.16 branch, most difficult will be CVE tracking. 19:58:05 But I will be happy to guide on it. 19:58:06 right I saw your email jcajka , I have no problem taking that although I never did anything in EPEL 19:58:30 it is essentially the same as Fedora ;) 20:00:04 well, in that case, unless anyone wants it, I can totally work on it 20:00:29 more folks working on it will not hurt :) 20:00:57 thank you again 20:01:15 my pleasure 20:01:22 we are on top of the hour, if there are no more topics, concerns, comments 20:01:32 I will end the meeting in a minute 20:01:48 not from my side 20:02:24 All good here 20:03:10 #endmeeting