IRC logs for #aegir, 2016-10-27 (GMT)

2016-10-26
2016-10-28
TimeNickMessage
[11:04:31]* theMusician has quit (Quit: theMusician)
[12:55:54]* mgann has joined #aegir
[13:04:13]* realityloop has joined #aegir
[13:11:18]* realityloop has quit (Ping timeout: 245 seconds)
[13:14:43]* hestenet has joined #aegir
[13:19:34]* realityloop has joined #aegir
[13:19:52]* hestenet has quit (Ping timeout: 276 seconds)
[13:28:29]* realityloop has quit (Quit: Leaving...)
[14:37:42]* gusaus has quit (Quit: gusaus)
[14:49:41]* gusaus has joined #aegir
[14:50:48]* gusaus has quit (Client Quit)
[17:16:24]* boshtian has joined #aegir
[17:55:09]* ybabel has joined #aegir
[18:52:36]* boshtian has quit (Quit: boshtian)
[18:56:25]* boshtian has joined #aegir
[18:58:42]* ybabel has quit (Read error: Connection reset by peer)
[18:58:46]* ybabel1 has joined #aegir
[19:01:07]* ybabel1 is now known as ybabel
[19:01:36]* David_Hernandez has joined #aegir
[00:06:07]* fatguylaughing has joined #aegir
[00:07:16]* shaneonabike has joined #aegir
[01:40:35]* theMusician has joined #aegir
[01:40:38]* zombiebeard has joined #aegir
[01:55:53]* boshtian has quit (Ping timeout: 245 seconds)
[01:56:10]* hestenet has joined #aegir
[01:56:25]* ybabel has quit (Remote host closed the connection)
[01:56:26]* ybabel1 has joined #aegir
[01:58:47]* ybabel1 is now known as ybabel
[02:13:03]<colan>ergonlogic: can we move Aegir Kubernetes to the main Aegir group on GitLab? I don't believe anyone would object.
[02:13:10]<colan>also, anyone see this? https://www.previousnext.com.au/blog/enterprise-drupal-hosting-skipper ?
[02:14:00]<colan>"While Skipper is not a mass-market, self-service offering [...]"
[02:14:34]<colan>maybe we should get in touch with them & see if they want to collaborate given that aegir is already full-service, but we're still working on the other parts.
[02:15:04]<colan>Kubernetes is the crossover here.
[02:33:24]<jonpugh>colan: ergonlogic: I think helmo and I would even prefer moving it to github. We have to use github for the documentation and for travis testing, it would be nice if we could move everything to one place.
[02:41:35]* shaneonabike has quit (Quit: Leaving.)
[02:41:39]* noecc has joined #aegir
[02:47:23]<colan>jonpugh: let's do both? :) keep one as a mirror with most of the project features off.
[02:52:03]<colan>keeping github as authoritative might be best. given that gitlab has an "import from github" button, we can switch by deleted the mirror & hitting that button.
[02:52:17]<colan>if we ever want to do that one day.
[02:52:33]* shaneonabike has joined #aegir
[02:53:07]<colan>anyway, my point was more that we should move the project from the co-op to aegir.
[03:15:31]<colan>helmo: you'll be an hour off for the meeting i think.
[03:40:16]<helmo>colan: the DST change is this weekend for me .. should be in 1 hour 20 min from now
[03:46:23]* helmo is now running the release scripts for 3.8
[03:46:52]* David_Hernandez has quit (Quit: Saliendo)
[04:46:37]* memtkmcc has joined #aegir
[04:54:19]* helmo has changed the topic to Aegir hosting system 3.8 released! http://docs.aegirproject.org/en/3.x/release-notes/3.8/ | Planning for AegirNG has begun: http://tinyurl.com/aegirng | IRC Logs: https://hefring.mig5.net/bot/log/aegir | Scrum/Coop: https://plus.google.com/hangouts/_/ergonlogic.com/aegir-co-op
[04:54:58]<ergonlogic>colan: feel free to reach out re. Skipper. But bear in mind that there's nothing on that page about it being FLOSS, nor is there a "skipper" project on their github
[04:57:05]<colan>k
[04:57:34]<ergonlogic>one nice thing about git is that there isn't a need for one canonical repo
[04:57:53]<ergonlogic>so feel free to copy stuff between orgs, or whatever
[04:58:12]<colan>ergonlogic: but okay to move it out of the co-op org then?
[04:58:45]<ergonlogic>my only real concern is that it's not really ready for prime time, imo
[04:59:08]<ergonlogic>but sure
[04:59:52]<ergonlogic>I personally, prefer gitlab, and will likely continue to use it for my own development
[04:59:58]<colan>right, but that shouldn't affect the parent org though. ok so i'll move it to aegir on gitlab, and then push to github aegir as well.
[05:00:14]<colan>that should make everybody happy.
[05:00:28]<ergonlogic>but I won't block consolidating to github, if that's the consensus among the core group
[05:00:28]<colan>except the bitbucket folks ;)
[05:00:47]<colan>i'm for both at the moment, personally.
[05:01:33]<ergonlogic>being floss itself is a big +1 for gitlab, imo
[05:02:13]<ergonlogic>and, which under-resourced, gitlab-ci is better than TravisCI, imo
[05:02:43]<ergonlogic>s/which/while/g
[05:02:44]<helmo>Github is mor practical for testing as we have Travis setup now ... but for teh rest +1 for GitLab
[05:02:54]<helmo>s/mor/more/
[05:03:28]<ergonlogic>gitlab-ci uses docker natively, fwiw
[05:03:38]<helmo>Scrum time :) Hi all, ergonlog1c bgm jonpugh colan gboudrias memtkmcc
[05:04:16]<ergonlogic>I guess we got a bit of an early start :p
[05:04:31]<colan>i may have another call soon, but i'll pop in here around that. might just be away for a bit.
[05:05:36]<anarcat>congrats on the release
[05:05:43]<helmo>I've seen lots of randomly failing tests these few days :( ... both Jenkins and Travis. But that seems mostly infra related
[05:05:44]<ergonlogic>helmo: way to go on 3.8 :)
[05:05:55]<anarcat>glad to see regular releases are still happening, and impressed at the kubernetes / hosting_wordpress projects
[05:05:58]<anarcat>that looks promising
[05:06:07]<helmo>I'm glad we got it out
[05:06:27]<ergonlogic>anarcat: helmo's been driving bi-monthly releases, which is great :)
[05:08:44]<helmo>I don't have much more to mention though ... I've been mostly busy reviewing and committing stuff.
[05:09:35]<ergonlogic>there was a fair ammount that had built up though, so kudos
[05:10:04]<bgm>hey, I'm around, I don't have anything to mention. Cool to see 3.8 out. Looking forward on further testing the nginx unforkening.
[05:10:24]<ergonlogic>for my part, I've been re-learning Drupal
[05:10:32]<ergonlogic>as 8 really is a different beast
[05:10:46]<bgm>I installed Gitlab on a private server. It's pretty neat. (issue boards are nice)
[05:11:01]<ergonlogic>as for the Kube stuff, I think the new architecture it'll provide is pretty promising
[05:11:33]<ergonlogic>and I'm starting on D8 server/services/tasks stuff, so that we can port the Kube prototype to D8, and go from there
[05:12:25]<jonpugh>That's awesome ergonlogic
[05:12:25]<jonpugh>Do you have a link to the code for that?
[05:12:40]<ergonlogic>I *think* I mentioned this last time, but hosting tasks can't really apply to entities, as they assume nodes throughout
[05:13:01]<ergonlogic>so the current kube stuff under D7 is a dead-end, imo
[05:13:14]<ergonlogic>at least, as far as taking it where I believe it needs to go
[05:13:29]<ergonlogic>jonpugh: there's nothing to show yet
[05:13:52]<anarcat>bgm: i like issue boards as well :)
[05:14:46]<ergonlogic>just working out things like how to build a codebase and symlink in a profile using composer
[05:15:07]<ergonlogic>phpunit and code-coverage, etc.
[05:15:50]<ergonlogic>for the d8 re-write, it's crucial that we build-in TDD from the get go, imo
[05:16:52]<ergonlogic>which is mostly a matter of discipline, but that should be greatly aided by code-coverage nagging
[05:17:19]<bgm>having clear APIs helps for that, imho
[05:17:31]<bgm>(d7 and previous culture didn't tend to have APIs, but hooks)
[05:18:08]<ergonlogic>found this recently:
[05:18:11]<ergonlogic>https://www.drupal.org/project/rest_api_doc
[05:18:58]<bgm>makes sense, civicrm also does something similar
[05:19:08]<ergonlogic>exactly
[05:20:19]<helmo>looks promising
[05:21:31]<ergonlogic>so, my plan, for the moment is to build out basic entities for servers, services and tasks, and hook them up to Ansible on the back end (mostly to provide the remote/SSH capabilities that Drush does for us now)
[05:24:08]<ergonlogic>I think we can avoid lots of complications by just serializing all relevant context into the task, rather than relying on anything like Drush aliases as a backend data store
[05:24:42]<ergonlogic>or relying on Ansible to generate such data via Facter/setup/etc.
[05:24:59]<jonpugh>Absolutely. I like the way symphony console config component can save anything to yml files
[05:25:33]<jonpugh>I've been thinking that's the way to go for years now, that's how terra cli and Drupal console work
[05:25:49]<ergonlogic>yeah, yaml's really easy to work with, and is simple to serialize
[05:26:23]<ergonlogic>I'm not planning to implement a CLI at this point
[05:26:45]<ergonlogic>but presumably it'd mostly just create tasks in the queue
[05:28:26]<jonpugh>Yeah, at this stage, most important part is clearly defining all entities and their properties
[05:28:40]<jonpugh>So good luck :)
[05:30:16]<ergonlogic>I figure to make most of them bundle-capable, to help define the intrisic properties of (for example) tasks, versus the specific of a given implementation
[05:30:51]<ergonlogic>that way, they can be extended by creating a bundle, adding fields, etc.
[05:31:13]<ergonlogic>and exported via Features
[05:31:47]<ergonlogic>which then gives us somewhere convenient to hang implementation specific code
[05:32:46]<ergonlogic>that should probably be the case pretty much across the board
[05:33:32]<ergonlogic>so, our "application" entity could be the base for a Drupal bundle, a WP bundle, etc.
[05:33:41]<colan>i don't think you need Features in d8, can just stick stuff in the module's code and config/install dir.
[05:34:56]<ergonlogic>well, the point of Features would be two-fold, imo
[05:35:39]<ergonlogic>first, to make it easier for people to start extending it with new application bundles or whatever
[05:35:54]<ergonlogic>so, basically, a module generator
[05:36:18]<ergonlogic>but also to be able to get the revert/override stuff
[05:36:32]<colan>fair enough.
[05:36:42]<ergonlogic>I guess config_manager or something may provide some of that
[05:37:09]<colan>core provides all or nothing. if you want split up config, then that's what Features is for.
[05:37:23]<colan>config_manager makes Features work, essentially.
[05:37:28]<ergonlogic>right
[05:38:19]* shaneonabike has left #aegir ("PONG :barjavel.freenode.net")
[05:38:23]<ergonlogic>anyway, the idea is, if you want to support some new app, you shouldn't have to know how to attach a field to a bundle in code, imo
[05:38:43]<ergonlogic>or a field instance, I guess
[05:39:06]<ergonlogic>there's nothing forcing you to use Features to do so, of course, either
[05:39:15]<colan>i've been doing a lot of work with configuring d8 sites as i want, and then exporting all of it as install profiles. works great: https://www.drupal.org/docs/8/creating-distributions/how-to-write-a-drup...
[05:40:08]<colan>would be a good way to generate the new hostmaster, or whatever it's called.
[05:40:36]<ergonlogic>yep, that's what I'm working from
[05:40:59]<jonpugh>I vote to keep it named hostmaster. there's not a good enough reason to disrupt the naming convention for the "drupal install profile" component of aegir
[05:41:05]<ergonlogic>but, I don't think that provides a way to update config, does it
[05:41:06]<ergonlogic>?
[05:41:09]<colan>took me a while to figure that stuff out then; i worked on that doc. ;)
[05:42:00]<colan>well, you mean update config on the site, and re-export to get the updated profile? you can do that, yes.
[05:42:13]<ergonlogic>personally, I think we should move everything under the "aegir_" namespace
[05:42:47]<colan>i'll second that. findability++
[05:42:48]<ergonlogic>colan: I mean, once you have a running site, built from a profile, how do you propagate changes to config?
[05:43:05]<ergonlogic>like, an updated view, for example
[05:43:42]<colan>re-export the config, overwriting the old profile, release a new version of the profile.
[05:44:09]<jonpugh>hey I have to go so I have to give my updates now while you're around
[05:44:17]<colan>shoot
[05:44:20]<ergonlogic>jonpugh: go for it
[05:45:14]<jonpugh>all the docker/testing/travis stuff is very challenging and I've burnt myself out on it for the moment, so I need to ask for more help on that again.
[05:45:29]<jonpugh>there are people like tommycox who have come forward on the drupal issue i posted...
[05:45:51]<jonpugh>so instead of that stuff the last few weeks was all working on updating hosting_tasks module for dynamic updates using vue.js
[05:46:49]<jonpugh>but, like you said, the testing stuff is crucial for a better cycle for drupal8 aegir, so...
[05:46:49]<helmo>Also greatly appreciated
[05:47:11]<jonpugh>Maybe now that we have more kubernetes work done and more docker work done we can lure more people back to the project to help
[05:47:33]<jonpugh>which brings me back around to what I mentioned in #aegir-coop, http://aegir.atrium.devshop.build
[05:47:51]<jonpugh>I think we can setup a simple blog on this page just to give updates on the interesting stuff we are doing, even if it's experimental.
[05:48:09]<jonpugh>and submit that to drupal planet, which will help us attract more help
[05:48:19]<ergonlogic>fwiw, I think the aegirNG stuff will be different enough that it'll require a different test suites
[05:48:40]<jonpugh>which is fine, my point is we need infrastructure ready for testing as we get started.
[05:49:18]<ergonlogic>agreed
[05:49:18]<jonpugh>and it's running decently for 7.x-3.x, we should be able to easily adapt it for what you are working on
[05:50:00]<jonpugh>but maintaining the whole stack is a beast
[05:50:07]<jonpugh>dockerfiles need to be created for new releases
[05:50:11]<jonpugh>updated for better features
[05:50:21]<jonpugh>README's, behat tests repos...
[05:50:23]<jonpugh>it's a lot
[05:50:37]<anarcat>i love how drupal.org now has a header so big that only the title of docs page show up now.... wtf.
[05:50:44]<ergonlogic>I think we should try to simplify that as much as possible, and build in release processes as we go too
[05:51:14]<jonpugh>if we target offering an official docker image, we have a good goal that will force us to do that stuff
[05:51:38]<jonpugh>Which they now moved to the "Docker Store"
[05:51:39]<jonpugh>https://store.docker.com/images/9c2c5426-0cca-4a30-a450-b2961541c6dc
[05:51:44]<jonpugh>PHP image for example.
[05:53:05]<colan>definitely agreed on the blogging to bring the community back.
[05:53:54]<jonpugh>ok
[05:54:05]<colan>been building a list of things to write about for my blog, already on planet. we should cross-post stuff to the aegir blog, which absolutely needs to be on Planet.
[05:54:10]<jonpugh>We could think about importing the users from the old community site.
[05:54:18]<helmo>yes, on our own it's hard to get content for a blog ... but maybe together we can
[05:55:19]<jonpugh>I talked to schnitzel at DrupalCon New Orleans, who was promoting Amazee.io, their new private "cloud hosting". He said "Aegir is still a thing?" :d
[05:55:34]<colan>write for your own, then publish on the project one as well. hopefully that'll generate some interest.
[05:55:43]<jonpugh>yeah, we can cross post, no problem
[05:56:01]<jonpugh>ok, so I'll work on that
[05:56:01]<colan>jonpugh: yeah, every time i talk to anyone, i get "Oh, i should look at that again then"
[05:56:46]<colan>our lack of communication is feeding that mentality.
[05:57:06]<jonpugh>I was clicking through Atrium itself, it's very powerful, I think we can use the community space for the blog and other content like case studies
[05:57:26]<jonpugh>and the public coop space for our public content like meeting minutes
[05:57:41]<helmo>on that note ... I just sent mail to out announce list about the new release ... it's a documented step which we've skipped for too long
[05:57:43]<jonpugh>and the private coop space for members only discussions, maybe even polls for voting
[05:58:01]<jonpugh>fantastic
[05:58:17]* colan wonders how many are on that list
[05:58:20]<jonpugh>we should promote the mailing list more :)
[05:58:27]<jonpugh>yep that too
[05:58:35]<colan>or at least the project :)
[05:58:37]<helmo>107
[05:58:38]<hefring>http://drupal.org/node/107 => Bad regex pattern for username validate [#107] => 7 comments, 1 IRC mention
[05:58:55]<colan>1
[05:58:56]<hefring>http://drupal.org/node/1 => About => 0 comments, 3 IRC mentions
[05:59:00]<helmo>LOL
[05:59:18]<jonpugh>helmo: Release process is documented where? in the source code?
[05:59:23]<jonpugh>not seeing it on docs.aegirproject.org
[05:59:32]<helmo>jonpugh: on http://docs.aegirproject.org/en/3.x/community/release-process/#3-generat...
[06:00:20]<jonpugh>oh it's under community
[06:00:32]<jonpugh>probably should move it to "developing aegir"?
[06:00:48]<ergonlogic>yep
[06:00:59]<ergonlogic>we only just added that recently
[06:01:19]<jonpugh>I got it...
[06:01:22]<helmo>The process took less then an hour today ... It's just that the reviewing and testing before takes an infinite amount of time ;)
[06:02:05]<helmo>But travis has helped a lot there already to build confidence
[06:03:10]<colan>helmo++
[06:03:28]<jonpugh>Progress :)
[06:03:31]<colan>and jonpugh, great work on the task timer stuff.
[06:03:47]<jonpugh>thanks
[06:03:56]<jonpugh>I just got sick of having my own hack in devshop
[06:04:19]<jonpugh>I've had a secondary ajax going this whole time because fixing hosting_tasks was too intimidating at the time
[06:04:40]<jonpugh>It feels like vue.js integration needs to go deeper in it's own module.
[06:05:05]<jonpugh>All you have to do is define your data objects and output a template, i have a strong feeling it wouldn't take long to add it as base integration in views and fields, for example.
[06:05:40]<jonpugh>Not to get ahead of myself...
[06:06:28]<jonpugh>My plan is to include the patch in the upcoming devshop release so we can give it some real world testing in time to include in 3.9
[06:07:40]<jonpugh>again, something I really need help with! :D
[06:07:42]<jonpugh>so much...
[06:08:19]<ergonlogic>anything else for the scrum? or should we switch over to #aegir-coop?
[06:08:31]<ergonlogic>jonpugh: I'll be in touch as I get to that point with the NG stuff
[06:08:51]<jonpugh>When there's any code at all it will make a great blog post :)
[06:08:57]<jonpugh>and I'm sure people will want to help
[06:09:32]<anarcat>oh, sorry for the noise, didn't realize you had a scrum!
[06:09:35]<ergonlogic>well, the initial code will likely be mostly "minimal", plus dev/test harness
[06:09:49]<ergonlogic>anarcat: no worries. you're still an alum :)
[06:10:00]<jonpugh>whatever it is, people will still be excited to hear it's being worked on
[06:10:06]<anarcat>ergonlogic: thanks :)
[06:10:16]<ergonlogic>ok
[06:10:23]<anarcat>ergonlogic: ... assuming "alum" means "that old dude in the back who can rant on whatever", i'm in :p
[06:10:33]<anarcat>just don't make me write a new service file :p
[06:10:39]<anarcat>s/file/module/
[06:10:53]<ergonlogic>anarcat: I hope we've had to write our last one :p
[06:11:15]<jonpugh>yeah, they're classes now >D
[06:11:21]<jonpugh>oh wait...
[06:11:30]<anarcat>ergonlogic: woot
[06:11:40]<jonpugh>everything in drupal8 and symfony is a "service"
[06:11:44]<jonpugh>good times
[06:11:44]<anarcat>jonpugh: i believe they were classes before as well
[06:11:47]<anarcat>ugh
[06:12:00]* anarcat is looking forward to the closure rewrite of drupal
[06:12:07]<anarcat>clojure, sorry
[06:12:10]<anarcat>no pun intended
[06:12:18]<ergonlogic>ok, now youre trolling :p
[06:12:36]<anarcat>ergonlogic: that's what you get when you make people alums :p
[06:12:43]* anarcat goes back to fixing 4 year old security issues
[06:12:51]<anarcat>https://security-tracker.debian.org/tracker/CVE-2016-8859
[06:12:51]<colan>heh
[06:12:52]<anarcat>fun times
[06:12:53]<jonpugh>Yeah what about those?
[06:13:02]* noecc has left #aegir ("pax")
[06:13:17]<jonpugh>anarcat: keeping us safe!
[06:13:22]<jonpugh>anarcat++
[06:14:16]<anarcat>who the hell knows about tre anyways
[06:14:29]<anarcat>but apparently it's part of the musl libc, which is used in all of your Macs :p
[06:14:33]<helmo>anarcat++
[06:14:36]<anarcat>i'm just mopping behind
[06:14:51]<anarcat>hey enough karma, i'm paid for this shit, it's not like i'm donating my time ;)
[06:15:16]<anarcat>but thanks helmo / jonpugh / ergonlogic :)
[06:15:22]* anarcat wonders if there's still something tracking karma here
[06:15:51]<helmo>hefring: karma?
[06:16:19]<colan>karma++
[06:16:53]<jonpugh>anarcat?
[06:17:14]<anarcat>karma?
[06:17:21]<anarcat>plokta?
[06:17:44]* anarcat shrugs
[06:18:14]* ybabel has quit (Quit: ybabel)
[06:19:14]<memtkmcc>karma anarcat?
[06:19:26]<memtkmcc>hefring: karma anarcat
[06:19:26]<hefring>anarcat has karma of 20.
[06:19:28]<ergonlogic>hefring: karma anarcat?
[06:19:28]<hefring>anarcat has karma of 20.
[06:19:33]<memtkmcc>hah
[06:19:48]<anarcat>oh neat
[06:19:54]<anarcat>so it is a thing
[06:20:19]<memtkmcc>anarcat++
[06:20:30]<anarcat>karma anarcat
[06:20:33]<anarcat>karma anarcat?
[06:20:43]<memtkmcc>hefring!
[06:20:45]* anarcat kicks hefring
[06:20:46]<memtkmcc>:)
[06:20:59]<anarcat>hefring: karma anarcat?
[06:20:59]<hefring>anarcat has karma of 21.
[06:21:01]<anarcat>hefring: karma police?
[06:21:01]<hefring>police has karma of 0.
[06:21:05]* anarcat wins
[06:21:10]<memtkmcc>hefring: karma anarcat
[06:21:10]<hefring>anarcat has karma of 21.
[06:21:31]<memtkmcc>hefring: botsnack
[06:21:31]<hefring>sudo make me a sandwich!
[07:17:57]* mgann has quit (Quit: My iMac has gone to sleep. ZZZzzz…)
[07:25:30]* theMusician has quit (Ping timeout: 250 seconds)
[07:48:49]* memtkmcc has quit (Quit: Leaving.)
[08:06:10]* fatguylaughing has quit (Quit: fatguylaughing)
[08:56:20]* zombiebeard has quit (Quit: zombiebeard)
[09:26:04]* mstenta has quit (Ping timeout: 248 seconds)
[09:50:23]* mstenta has joined #aegir
[09:51:59]* gusaus has joined #aegir
[10:56:19]* gusaus has quit (Ping timeout: 256 seconds)
[10:59:09]* gusaus has joined #aegir