13:05:00 #startmeeting Ansible Collection Inclusion Review Day 13:05:00 Meeting started Wed Mar 24 13:05:00 2021 UTC. 13:05:00 This meeting is logged and archived in a public location. 13:05:00 The chair is gundalow. Information about MeetBot at http://wiki.debian.org/MeetBot. 13:05:00 Useful Commands: #action #agreed #halp #info #idea #link #topic. 13:05:00 The meeting name has been set to 'ansible_collection_inclusion_review_day' 13:05:06 o/ 13:05:15 #chair felixfontein andersson007_ dericcrago 13:05:15 Current chairs: andersson007_ dericcrago felixfontein gundalow 13:05:54 \o/ 13:07:37 could anyone remind me if there's a special procedure or not? Or we just should go through every collection submitted and say our opinions? 13:07:40 tadeboro: Your comments and explanation on NetApp were really good. One thing we (internally) spent some time discussing yesterday was "While it maybe OK for NetApp to do this (as they are one of the first collections) we will not use this as an example for others to follow. I'd hope that https://github.com/ansible-collections/netapp/issues/93 should fix most of it 13:07:47 andersson007_: first time we've done this 13:07:55 ah, ok 13:08:12 #info The full list of inclusion requests is https://github.com/ansible-collections/ansible-inclusion/discussions 13:08:23 bah 13:08:24 #undo 13:08:24 Removing item from minutes: INFO by gundalow at 13:08:12 : The full list of inclusion requests is https://github.com/ansible-collections/ansible-inclusion/discussions 13:08:35 #info The full list of inclusion requests is https://github.com/ansible-collections/ansible-inclusion/discussions/categories/new-collection-reviews 13:08:52 andersson007_: basically it's going through them, checking with the guidelines / your check list, and comment on things we find :) 13:08:52 (sorry I forgot my default that `/discussion` shows resolved items) 13:09:20 felixfontein: cool:) 13:09:37 and every entry should say whether it is OPTIONAL or REQUIRED 13:09:45 (entry = comment) 13:10:31 https://github.com/ansible-collections/ansible-inclusion/discussions/15 I did a review on equnix.metal - no feedback.. 13:11:30 and you also asked questions with no response 13:12:10 indeed 13:12:19 I'll ping them 13:12:21 I think we shouldn't review such submissions today 13:12:26 with no feedback 13:12:28 infoblox also isn't very active 13:12:39 andersson007_: agreed, move to the next one 13:12:46 +1 13:13:03 in particular because their modules won't work with ansible-core 2.11 (and ansible 4.0.0 will have ansible-core 2.11), and sanity tests currently fail 13:13:52 in which order do we want to go through them? 13:13:56 old to new? 13:14:11 the order in https://github.com/ansible-collections/ansible-inclusion/discussions/categories/new-collection-reviews will change once someone writes a comment somewhere :) 13:14:28 (or let's say the default sorting order) 13:14:50 https://github.com/ansible-collections/ansible-inclusion/discussions/14 there was a comment "Thanks, @Andersson007. we will complete you feedback and get back" 5 days ago. Not there yet. 13:15:15 should we skip it too? 13:15:32 felixfontein: FIFO sounds good 13:15:55 imo 13:16:39 I guess if we continue like that, we'll skip all ;) 13:16:56 the oldest application is infoblox: https://github.com/ansible-collections/ansible-inclusion/discussions/11 13:17:22 the second-oldest is hpe.nimble 13:17:38 (which hasn't a review yet) 13:17:55 infoblox has REQUIREDs from you unsolved, so.. 13:18:06 then comes dellemc, equinix, and then the netapp ones 13:18:08 hpe.nimble is a candidate then 13:18:49 infoblox did some work on their collection, but I'm not sure whether they addressed any of the points. at least they definitely didn't react directly to any of them. 13:19:05 ok, hpe.nimble 13:19:11 #topic hpe.nimble: https://github.com/ansible-collections/ansible-inclusion/discussions/13 13:20:15 ah, they have this 'great' repository structure again... 13:22:17 hum, I don't get why people do that 13:22:43 I guess, because you can check it out and run ansible-test without having to check it out correctly... 13:23:09 No CI (in GitHub at least) 13:23:10 it's not even ansible_collections, it's ansible_collection which probably doesn't work ? 13:23:19 also hi o/ 13:23:19 and/or because you want files outside the general collection structure in the repo that you don't have to exclude form the build process (though that sounds more like a lazy excuse) 13:23:24 is it crucial or i can start to review other stuff there? 13:23:26 hi dmsimard! 13:23:32 hi dmsimard 13:23:38 ! 13:23:38 Command "" not found. 13:23:39 dmsimard: indeed! 13:23:56 #chair dmsimard 13:23:56 Current chairs: andersson007_ dericcrago dmsimard felixfontein gundalow 13:24:05 I'll be semi afk for a while longer, internal presentation and dad ops 13:24:12 #info guidelines: https://github.com/ansible-collections/overview/blob/main/collection_requirements.rst 13:24:17 dmsimard: ack 13:24:19 #info checklist: https://raw.githubusercontent.com/ansible-collections/overview/main/collection_requirements.rst 13:24:33 darn 13:24:34 #undo 13:24:34 Removing item from minutes: INFO by felixfontein at 13:24:19 : checklist: https://raw.githubusercontent.com/ansible-collections/overview/main/collection_requirements.rst 13:24:40 #info checklist: https://raw.githubusercontent.com/ansible-collections/overview/main/collection_checklist.md 13:25:00 thanks felixfontein 13:25:14 you're reading my thoughts 13:26:31 hmm, do we allow Apache license right now? 13:26:58 in the guidelines we only mention GPLv3+ and BSD-2-clause 13:27:36 they also say "We will have a list of other open source licenses which are allowed as soon as we get Red Hat's legal team to approve such a list for us." 13:27:49 gundalow: do you know if there are news on that? 13:28:03 Yes, i've just also noticed HPE Nimble Storage Content Collection for Ansible is released under the Apache-2.0 license. 13:33:24 the test section isn't very great :) 13:33:31 - [ ] passed `ansible-test sanity` - REQUIRED: please create a ignore-2.11.txt file, after all Ansible 4.0.0 is based on ansible-core 2.11 - REQUIRED: multiple validate-modules issues are failing with the 2.11 sanity checks 13:34:02 the changelog also isn't great, there's no human-readable version and it is missing the 1.0.1 release 13:36:45 I think I went through everything that can be done quickly 13:36:54 (I've just pasted my current version of the checklist to GH) 13:37:08 now comes most of the standards and docs section 13:41:57 I see tag 1.0.0 but on galaxy there's release 1.0.1 13:43:37 galaxy.yml says 1.0.1, I guess they "just" didn't release it properly after fixing some first issues 13:43:47 I think when a new collection is submitted, we should put the references to the requirements. There can be a lot stuff fixed from a submission date to a review one 13:43:56 felixfontein: yeah 13:44:16 they don't use FQCNs in examples 13:48:07 and they use ANSIBLE_METADATA 13:49:10 true. and with support:community :) 13:51:39 i even haven't run ansible_reviewer yet.. 13:51:49 should i? :) 13:56:51 I see version_added 2.9 everywhere. It wasn't included in ansible/ansible, so it should be `1.0.0`? 13:57:13 yes 13:57:41 the documentation looks generally quite good, appart from version_added, missing FQCNs, and some formatting details 13:58:47 (or let's say, I've seen a lot worse ;) ) 13:58:50 I'd also point out that when we had `ansible/ansible` modules, we didn't want the vendor name in the module name, only the product name, so `hpe_nimble_fc` would have been instad `nimble_fc` 13:59:54 cyberpear: good point! 14:00:06 do you want to add it to the review? 14:00:58 #chair cyberpear 14:00:58 Current chairs: andersson007_ cyberpear dericcrago dmsimard felixfontein gundalow 14:01:01 I'm generally against dropping the product prefix from the module, because you'll end up colliding when you try to mix something like `aws.aws.server` and `gcp.gcp.server` in the same playbook 14:01:31 but the vendor prefix shouldn't be there, imo... I'll add to the review. 14:01:33 well, only if you use the cursed `collections:` keyword ;) 14:02:00 cursed ♥ 14:02:10 :D 14:14:24 ok I am released \o/ 14:14:29 are we still looking at hpe.nimble ? 14:14:47 I think I'm through with it 14:14:51 andersson007_: how are you doing? 14:15:25 I'm looking at the nimble 14:18:18 ok I will jump on the next one 14:19:33 dellemc? 14:20:35 though I guess I'll look only at some points, since we're still waiting for them to adjust andersson007_'s feedback from a week ago 14:20:55 I'll wait until meeting switches over to the next one so I don't go out of turn :p 14:25:10 I'm currently giving the hpe collection a quick review, basically some things that I see when looking at it - once andersson007_ is done with nimble we can choose/continue with the next one 14:26:30 feel free to start next one. I'll finish soon 14:30:26 it's a bit annoying that some collections have a review for already some time now, but don't really react on it 14:30:53 I guess the next would be equinix, but also there hasn't been any feedback to the existing review from a week ago 14:33:29 I'll look a bit at equinix until andersson007_ is done :) 14:34:03 yeah. I've just finished nimble. I think i should go to sleep a bit otherwise I won't be able to be at the irc meeting:) 14:34:19 lol ok :) 14:34:25 :) 14:39:15 couldn't find anything that looks bad on the first glance (at least nothing that andersson007_ hasn't mentioned yet) 14:39:34 I guess next come the netapp collections? 14:40:24 netapp.azure, netapp.cloudmanager and netapp.um_info 14:41:15 andersson007_ already reviewed cloudmanager, jillr already um_info 14:44:44 let's do azure if it hasn't been looked at before ? 14:45:03 sounds good 14:45:20 a bit of a problem is that all three collections are in the same repository, so some comments affect all of them 14:48:48 oh, it's that one :) 14:49:36 yes ;) 14:49:58 we only have 'special' ones today it seems - we just finished the only non-special (hpe.nimble) 14:51:24 so https://github.com/ansible-collections/ansible-inclusion/discussions/17 it is ? 14:51:38 #topic netapp.azure https://github.com/ansible-collections/ansible-inclusion/discussions/17 14:52:43 that's ... not semantic versioning 14:54:30 popping in real quick before I have to run into meetings for a while, all the netapp collections looked to have the same remaining items on the checklist needing fixed so I sort of took the comments on each as blanket reviews to all of them 14:54:53 all the other collections in the list looked to have blocking items outstanding from other folk's reviews already as well 14:55:18 jillr: thanks for looking at them, appreciate the help :) 14:58:27 #chair jillr 14:58:27 Current chairs: andersson007_ cyberpear dericcrago dmsimard felixfontein gundalow jillr 14:58:34 jillr: thanks! 15:00:39 dmsimard: thanks for changing the topic, I got a bit distracted by one of our cats :) 15:00:45 tadeboro already mentioned the lack of semantic versioning here: https://github.com/ansible-collections/ansible-inclusion/discussions/16#discussioncomment-506143 so I won't repeat it 15:01:03 now she is lying behind my chair, so when I roll back I'll flatten her.... apparently not the most intelligent cat :) 15:01:33 heh 15:03:15 now I moved her away, and then... she comes back. -.- 15:03:48 suicide cat... :) 15:10:48 not coming across a whole lot that hasn't already been said 15:10:59 it does have a dependency on azure.azcollection which is already in the package 15:11:41 the dependency is missing >=1.0.0 15:11:53 and there's a tgz file in the repo 15:12:02 yeah I didn't know that "*" even worked 15:12:20 felixfontein: yeah I noticed the tgz file but it's not in the actual collection tarball (I downloaded and checked) 15:14:48 "*" is the default value I think 15:14:54 (which means I don't care which version) 15:15:33 I thought so 15:18:21 I've added a general comment on CI (not all sanity tests were ran) at https://github.com/ansible-collections/ansible-inclusion/discussions/16#discussioncomment-525272 15:24:10 I pushed the ansible-4.0.0alpha2 tarball yesterday night. I'm working on the announcements now. 15:26:04 \o/ 15:26:19 (and good morning :) ) 15:33:21 Good morning :-) 15:33:34 felixfontein: (et al) Last nioght, I added support for doing this: 15:33:36 python -c 'from ansible_collections.ansible_release import ansible_version; print(ansible_version)' 15:33:41 4.0.0a2 15:33:57 I've seen the PR and added a :+1: I think ;) 15:34:34 I'm not happy with the name (ansible repeated three times) but I don't want to be unclear.... ansible_collections is a shared directory structure and the version only applies to the ansible package. 15:34:42 it would be great if `ansible --version` would show that as well, but I think the core team doesn't want that 15:34:43 Do you have any suggestions? 15:35:10 I think the repetition is good and warranted, for basically the reason you mentioned above :) 15:35:28 Cool :-) I'll leave it as is and add a note about it to the release notes. 15:36:08 :+1: 15:40:02 dmsimard: I added https://github.com/ansible-collections/ansible-inclusion/discussions/17#discussioncomment-525381 15:40:37 it's too bad that we don't have more collection applications in a state for a proper review :/ 15:42:50 Hi all. I just finished my last meeting for today. Did you leave anything for me to review? (I am still reading the backlog, it will take me a while to read it through.) 15:43:53 tadeboro: we mostly noticed that there isn't much to review, since most collections already have reviews which were waiting for feedback for some time 15:44:18 it feels a bit wasteful to review them again without the first review(s) being addressed first 15:44:52 Ugh, not an ideal situation. And I agree: reviewing things that did not fix already identified issues makes little sense. 15:45:46 I think I wrote at least a comment for every collection with a few things I spotted that might not have been mentioned yet (or I overlooked in the other reviews), but I'd rather wait with a full review 15:48:55 I see that semver is still too hard for some people. Oh well, it is what it is ... 15:50:16 :) 15:50:24 also having changelog entries for all releases is not trivial 15:50:30 or even tags for releases 15:50:50 my understanding from looking at the netapp things 15:51:11 is that they have everything in one repo so that it's more convenient for them but also release all their collections with the same (monthly calver) tag every month 15:51:31 I suppose having multiple repos and each collection having their own versioning would be a non-zero overhead 15:51:41 doesn't mean that we have to be cool with it, but I can see the reasoning 15:52:58 dmsimard: they can still use semver for all collections at once 15:53:21 some collections would get major releases without breaking changes, but that's better IMO than not having semver 15:54:08 yeah don't get me wrong, I would tend to stand by our requirement on semver because otherwise it would become too complicated for us to manage versions boundaries across versions of the ansible package 15:56:41 felixfontein: added a first run of the checklist on netapp.azure 15:57:07 I suggest we take a break, team will be in meeting soon 15:57:45 sounds good :) 15:58:33 at least the person from netapp seems to be actively working on it 15:58:46 (https://github.com/ansible-collections/ansible-inclusion/discussions/16#discussioncomment-525413) 15:59:04 shall we close the meeting ? maybe summarize what are the reviews we looked at 16:07:16 dmsimard: sounds good 16:07:23 (sorry got distracted again :) ) 16:07:58 though the reviews are mostly clear from the minutes I think 16:08:08 #topic Summary 16:10:48 #info Reviewed two collections and added comments to some more 16:11:06 #info All collections looked at need improvements 16:11:40 #info We did not look more closely at collections that have reviews with issues that need addressing but haven't been addressed for some time 16:11:43 #endmeeting