19:00:03 #startmeeting ansible meeting 19:00:03 Meeting started Tue Aug 1 19:00:03 2017 UTC. The chair is thaumos. Information about MeetBot at http://wiki.debian.org/MeetBot. 19:00:03 Useful Commands: #action #agreed #halp #info #idea #link #topic. 19:00:03 The meeting name has been set to 'ansible_meeting' 19:02:16 * nitzmahone lurks 19:02:33 #chair bcoca nitzmahone 19:02:33 Current chairs: bcoca nitzmahone thaumos 19:02:50 * bcoca pokes at thaumos to open august ticket 19:03:03 i was debating that. 19:03:05 lol 19:03:37 * jtanner brb 19:03:40 O/ 19:03:47 #chair newswangerd jtanner 19:03:47 Current chairs: bcoca jtanner newswangerd nitzmahone thaumos 19:06:07 * mkrizek lurks 19:06:22 so ... what's up 19:06:56 #topic https://github.com/ansible/ansible/pull/18662 19:07:25 ^ crab fixed ticket to desired functionality, but we still have to decide on name of the option (hard part) 19:07:40 thanks bcoca 19:07:46 august issue is now created. 19:07:47 i didn't want overlap with existing names (he used vars, also mentined extra_vars) 19:08:31 ^ mostly bikshedding, anyone got better name? 19:08:40 we can table for later while you think 19:08:42 or just ignore 19:09:39 Hi, sorry for my tardiness 19:10:40 #chair abadger1999 19:10:40 Current chairs: abadger1999 bcoca jtanner newswangerd nitzmahone thaumos 19:10:45 I am reading the issue over again. 19:12:20 lookup_vars? 19:12:57 lookup('template', './some_template.j2', vars=dict(x=42)) <= passing vars to template lookup 19:13:20 since you can use more than one instance per task, vars: is not a good way, only thing i think name should not be 'vars' as it is waaay overloaded 19:13:25 right so I mean lookup('template', './some_template.j2', lookup_vars=dict(x=42)) 19:13:30 additional_vars 19:13:41 +anything that is not taken 19:13:47 these vars are specific to lookups, are they not? 19:13:57 to the template lookup 19:14:04 not lookups, specifically this one 19:14:07 ah ok 19:14:13 they're specific to the invocation. 19:14:15 template_vars 19:14:23 yeah I was thinking template_vars too 19:14:30 That works in my head. 19:14:35 It's parameters that get passed through our lookup all the way to the jinja templating engine. 19:14:44 +1 template_vars 19:14:52 +1 template_vars 19:15:01 +1 19:15:56 k, done, since no one else seems to care 19:16:07 #action bcoca updates ticket with resolution 19:16:44 thanks @bcoca 19:17:17 * bcoca passes baton to thaumos 19:17:20 #topic ansible/ansible#27519 19:17:24 #link https://github.com/ansible/ansible/pull/27519 19:17:31 @bcoca, I believe you're already reviewing this one, right? 19:17:46 no, just finished their previous ones 19:18:00 ah, okay. I thought you started to based on your other chat msg. 19:18:39 @bcoca this is a 2nd module following the cyberark_authentication one 19:18:45 hoices: [details, present, update, addtogroup, absent] <= choices for 'state' already tell me they did not look at the other modules they had ready after all our feedback on previous 19:19:04 #chair cyberark-bizdev 19:19:04 Current chairs: abadger1999 bcoca cyberark-bizdev jtanner newswangerd nitzmahone thaumos 19:19:06 cyberark-bizdev: ^ those kind of states are things we try to avoid 19:19:30 update shoudl be implied in present (diff between changed=true or false 19:19:32 ) 19:19:53 addtogroup and detais are imperative not declarative 19:20:03 ok 19:20:16 bcoca... how do we implement those then ? 19:20:22 you probably want a _facts module instead of details 19:20:52 cyberark-bizdev: as we discussed in the previous module, handle the changed states internally, let user use simple declarative language 19:22:00 the return also has same issue, you show the fulls tructure in sample ... but its supposed to be documented outside of it 19:22:08 @cyberark-bizdev, I really would like to ask you to work with us in #ansible-devel 19:22:22 This back and forth in meeting is not working. 19:22:41 ok... this is the first time I submit this PR for discussion here. 19:22:46 main reason we do one module at time is so systemic issues like the ones above can be corrected on subsequent submissions 19:23:09 if there was no change to the modules you were submitting based on previous reviews ... kind of missing the point 19:23:20 @cyberark-bizdev, you are missing the point of the meeting. 19:23:25 I understand that out, and this module has other functionality which was not present in the previous one 19:23:36 the other one only had present, and absent 19:23:47 so i expect new problems 19:23:59 well, previous one had add/edit features that we hammerd out early 19:24:08 vis the password chagne 19:25:26 yes... I have followed, and updated according to feedback 19:25:30 okay, for time's sake... I am asking this to move to #ansible-devel. @cyberark-bizdev, will you please work with us there. 19:25:37 I will 19:25:40 thank you 19:26:19 delete user is not handling changed 19:26:46 also making changed=false in check_mode ... not really what check_mode means 19:27:11 okay, changing topics 19:27:18 #topic ansible/ansible#27537 19:27:41 #likn https://github.com/ansible/ansible/issues/27537 19:28:06 house decided to solve himself 19:28:11 @bcoca, was this merged affter?... okay 19:28:14 so we cool? 19:28:37 hmm, not really as we still have message in other case 19:28:48 wanted to discuss if this should really be the way we handle roles 19:28:57 k, gotcha 19:29:28 as jmainguy points out, you can easlily write a role that work across several 'ansible wrappers' and ansible itself 19:29:38 having these warnings seems uneccesary in ansible itself 19:30:30 I guess the point is, why would we warn that this is a container role vs a infra role... I don't think we should have a warning. A role is a role. It should be able to work anywhere/everywhere. If it doesn't it's broken imo 19:31:14 agreed 19:31:23 jimi|ansible, abadger1999, anyone else? 19:31:52 chouseknecht: any input from you on why we have a warning in the first place? 19:32:19 bcoca: I can't think of a reason. Would like to hear if there was rationale, though. 19:32:49 unshipit 19:33:31 i believe it was mostly that they added this thing for 'ansible-container' (a meta entry) and taht roles flagged with it were meant for container .. did not know they added warning, i see no real reason for it in ansible (ansible-container might want to warn for roles that DONT have the key) 19:33:43 I assume it was a nicety for a-c. 19:33:54 what bcoca said 19:34:07 even having it as a meta entry doesn't make sense to me 19:34:44 he removed 1 warning, but kept another (i see container now has 2 keys) 19:38:54 soo .. no answer? we can table for now, lets try to ping him offline 19:39:36 yeah .. I'm tentatively for removing but since I don't know why it was added, I hate to just remove arbitrarily. 19:39:46 i agree, a role should be a role 19:40:07 ^^ seems like sound logic to me 19:40:09 #action capture and interrogate chouseknecht 19:41:02 using only NATO approved techniques 19:41:17 but .. trump said its ok! 19:41:24 oh, okay then 19:42:56 chouseknecht already removed the message though 19:43:11 so what's there to interrogate or being tentative about removing? 19:43:52 I think bcoca just wanted to hash out why this was in Ansible in the first place, or am I completely lost as usual? 19:43:56 he removed 1/2 of the messages 19:44:00 and yes, that too 19:44:05 ah, only half 19:45:11 I removed the one geerlingguy objected to 19:45:36 oh hai! 19:45:40 #chair chouseknecht 19:45:40 Current chairs: abadger1999 bcoca chouseknecht cyberark-bizdev jtanner newswangerd nitzmahone thaumos 19:46:10 chouseknecht: question is why warn at all, the roles shoudl be executable from ansible w/o issues 19:46:27 and the other warning is still in there 19:46:49 They kinda don't make sense outside of AC though 19:47:05 that seems to depend on construction 19:47:25 even if role has extra files for container .. it could still be used normally 19:47:31 For container enabled, yes 19:47:33 as Jmainguy's ticket pointed out 19:47:49 Jmainguy? 19:48:01 ^ might have wrong alias, but ticket you closed 19:48:03 geerlingguy? 19:48:09 ^ possibly ... 19:49:54 It really doesn't matter what the role contains... Whether it has a file for AC or a file for whatever else. To me a role can contain whatever the hell the person wants to put in it. 19:50:48 The fact that is has actual stuff in it that pertains to AC is cool, though! At least it's valuable and not taking up disk space 😄 19:54:45 Okay... if there's no other reasons the warning should exist, then take it out? 19:54:48 so... Is there any reason why we wouldn't remove everything else? 19:54:49 yeah I am not familiar with this specific one, thing you got the wrong main guy 19:55:01 lol @Jmainguy 19:55:48 yeah a role can contain a huge number of tasks, for it to alarm at the role level seems wrong 19:55:53 if that is indeed what is happening here 19:56:20 all you 'guy's look alike! 19:56:27 =) 19:57:43 okay, so I say let's take this offline and discuss more. 19:58:13 #action continue discussion offline 19:58:17 any other quick items to address? 19:58:25 if not, I'm going to endmeeting 19:58:46 #endmeeting