[00:03] <blackboxsw> and drop the super deprecated modules part 2 https://github.com/canonical/cloud-init/pull/134
[00:03] <blackboxsw> ok EOD for me
[08:32] <mkrai> blackboxsw, Hi I get error running clour-init query all command http://paste.openstack.org/show/787808/
[14:03] <otubo> I have to work in such a different timezone, any of the maintainers here can help with the migrate-lp-user-to-github script?
[14:03] <otubo> smoser: blackboxsw perhaps are already up :
[14:04] <otubo> *I hate to work, I can't even blame the autocorrect, it's my mind that want's to leave on vacation already :-D
[14:26] <otubo> Ha! No worries smoser blackboxsw, I just figured out the problem I was having. That's my launchpad-github mapping https://github.com/canonical/cloud-init/pull/135
[14:27] <otubo> Also, I run the script with --dryrun enabled but it ended up doing the whole proccess. Perhaps there's a bug there. :-)
[14:41] <Odd_Bloke> blackboxsw: I'm seeing a bunch of automated "This bug is believed to be fixed in cloud-init in version 19.2-67" bug mail from you, with the last number changing in (what looks like) each one.  Is that intentional?
[14:43] <Odd_Bloke> Oh, I see the PR now.
=== hggdh is now known as QuosqueTandem
[15:10] <smoser> Odd_Bloke: i just did https://github.com/canonical/cloud-init/pull/135
[15:10] <smoser> approved it, but if you can land it.. i'm not sure (i need to figure out) what the landing process is.
[15:13] <otubo> smoser: Odd_Bloke rharper blackboxsw guys, I'm leaving for the shutdown now. Thanks for the good work this year. See ya in the next decade. :-)
[15:23] <Odd_Bloke> otubo: Thanks, have a good break! :)
[15:25] <Odd_Bloke> smoser: I can't see o'tubo in our CLA spreadsheet, so I'm checking internally if we have a non-individual signature that covers them.
[15:29] <danmux> Hi, is it ok to ask for help on using cloud-init in here, it's not just reserved for development?
[15:32] <Odd_Bloke> danmux: Yep, please do!
[15:34] <danmux> ace! specifically multipart user-data and jinja2
[15:35] <danmux> Should I be using `Content-Type: text/jinja2` in place of `text/cloud-config` ?
[15:36] <danmux> (also can i confirm `cloud-init devel render` does not handle muilti-part files, just the raw template bit)
[15:38] <danmux> in which case how best can I test my multipart `user-data.txt`
[15:40] <Odd_Bloke> Hmm, good questions.
[15:41] <Odd_Bloke> I haven't really dealt with multipart user-data much myself.
[15:41] <Odd_Bloke> smoser: blackboxsw: Perhaps one of you have context to answer (some of) the above without digging into the code too much?
[15:54] <danmux> I've been digging around the code a bit, and it looks to me like it is either or thing but I'm not 100%
[15:54] <danmux> I can't tell if it would go on and honour the `template` directive - for example...
[15:54] <danmux> ```Content-Type: multipart/mixed; boundary="MIMEBOUNDARY"MIME-Version: 1.0--MIMEBOUNDARYContent-Transfer-Encoding: 7bitContent-Type: text/cloud-configMime-Version: 1.0## template: jinja#cloud-configwrite_files:  - path: /usr/local/etc/nomad/client.hcl```
[15:55] <danmux> (oops)
[15:58] <danmux> for info `/usr/bin/cloud-init 19.2-36-g059d049c-0ubuntu2~18.04.1`
[16:25] <blackboxsw> danmux: it's been a while since I've touched that part of the code. will run a test. I think you might have had to pass text/jinja2 as content type. But, will validate and get you a working example
[16:26] <danmux> blackboxsw hang fire a sec - im able to test now
[16:26] <blackboxsw> sweet
[16:27] <danmux> hmm ok it failed
[16:28] <danmux> oh should i be editing the `.i` file ?
[16:30] <danmux> blackboxsw - ok I was running a single module again, but i think the templating is done in `init`?
[16:33] <danmux> ah - im doing something dumb
[16:45] <blackboxsw> danmux, here's what works for me
[16:45] <blackboxsw> ... sorry getting a paste together
[16:46] <danmux> (y)
[16:48] <blackboxsw> danmux: at long last https://pastebin.ubuntu.com/p/GzbR23WjQg/
[16:50] <danmux> tyvm - you couldn't paste the `mime-msg` as well blackboxsw :pray:
[16:50] <blackboxsw> to retest on your system that had already booted cloud-init. you can run the following to clean you system, force a reboot and re-run cloud-init with :     sudo cloud-init clean --logs --reboot
[16:51] <blackboxsw> ahh will do now danmux
[16:52] <blackboxsw> danmux: https://pastebin.ubuntu.com/p/XKwSf7x4bX/
[16:52] <danmux> great ty for the test advice - otherwise I should just setup a container as per your example.
[16:52] <blackboxsw> the tool in cloud-init ./tools/make-mime.py -a <your_yaml>:<mime-type>
[16:53] <blackboxsw> will generate that content for you
[16:53] <blackboxsw> if you have cloud-init source tree
[16:53] <blackboxsw> https://github.com/canonical/cloud-init/blob/master/tools/make-mime.py
[16:54] <danmux> yup - got it - ours is being generated by terraform. But with your info I'm sure i can make some progress!
[16:54] <blackboxsw> good luck :)
[16:54] <blackboxsw> hope it helps
[17:23] <danmux> Woohooo! blackboxsw that nailed it! tyvm - no I can sign off for Christmas (UK timezone) - Merry Christmas to you and to Odd_Bloke for the referral (y)
[17:23] <blackboxsw> Schweet! danmux
[17:23] <danmux> s/no/now
[17:23] <blackboxsw> good to hear
=== hjensas is now known as hjensas|afk
[18:05] <meena> ah, the joy of freelancing. no holidays, no vacations, no time off, no nothing
[18:06] <meena> (i think i'm mostly joking, but i can't really tell)
[18:12] <meena> https://github.com/canonical/cloud-init/pull/61#pullrequestreview-335429532 merge?
[18:18] <blackboxsw> hah!
[18:27] <blackboxsw> powersj: minor suggestion on your doc branch and I think we can land it https://github.com/canonical/cloud-init/pull/109#pullrequestreview-335437127
[18:27] <powersj> blackboxsw, fixed
[18:27] <blackboxsw> meena: I'll let Odd_Bloke close out on that review
[18:32] <blackboxsw> approved waiting on CI
[18:45] <Odd_Bloke> meena: Goneri: It's landed!
[18:54] <Odd_Bloke> blackboxsw: https://github.com/canonical/cloud-init/pull/136
[19:17] <meena> cool, cool, cool
[19:17] <meena> now, it's time for some refactoring
[19:20] <blackboxsw> Odd_Bloke: on it for 136. and powers I replied to your doc https://github.com/canonical/cloud-init/pull/104/files
[19:20] <blackboxsw> ooh meena, putting the fun back into refactoring :)
[19:23] <meena> "can't spell refactoring without fun" — blackboxsw
[19:24] <blackboxsw> :)   -> lunch -> coffee
[19:25] <blackboxsw> or -> coffee -> lunch .... can't decide
[19:28] <meena> i find lunch - > coffee to be slightly more conducive to having a proper post lunch sleep and then wake up refreshed
[19:29] <Odd_Bloke> "refunctoring" means "rewriting in Haskell", right?
[19:35] <meena> I'm partial to any functional language i can write, haskell i can only read
[20:28] <Odd_Bloke> powersj: blackboxsw: Let's kick off the holidays (and the decade) the right way: https://github.com/canonical/cloud-init/pull/137
[20:28] <blackboxsw> woot!
[20:30] <blackboxsw> Odd_Bloke: in the theme of dropping all the dead code:  https://github.com/canonical/cloud-init/pull/134
[20:30] <blackboxsw> your above is +2'd just waiting on CI
[20:30] <Odd_Bloke> blackboxsw: The PR I've already Approved, you mean? ;)
[20:30] <blackboxsw> hah
[20:31] <blackboxsw> ahh I'll wait for your py.27 to land and I'll close out 134
[20:31] <Odd_Bloke> Thanks for the speedy reviews!
[20:32] <blackboxsw> unfractured attention makes for light work :)
[20:49] <Odd_Bloke> Landed. \o/
[20:59] <blackboxsw> Odd_Bloke: can you confirm breakage of make config/cloud.cfg
[21:00] <blackboxsw> or powersj
[21:00] <blackboxsw> on focal it doesn't work for me as it tries to interpret $user from config/cloud.cfg.tmpl line 17
[21:00] <blackboxsw> as a required substitution param
[21:01] <Odd_Bloke> blackboxsw: Breakage of what exactly?
[21:01] <blackboxsw>     return str(selected_params[key])
[21:01] <blackboxsw> KeyError: 'user'
[21:01] <blackboxsw> make: *** [Makefile:79: config/cloud.cfg] Error 1
[21:02] <blackboxsw> no emitted 'rendered' cloud-config
[21:02] <blackboxsw> I'm unable to make that target (to generate cloud.cfg from template)
[21:04] <blackboxsw> and if I do provide a bogus user: ME to the templater.render_string(...., tpl_params)  then I get a rendered output file that subtitutes ME for $user on line 17 of  config/cloud.cfg.tmpl
[21:06] <blackboxsw> it's tools/render-cloudcfg is called by setup.py during package creation. I'm wondering why that hasn't broken our package builds too
[21:08] <blackboxsw> interesting. this works on xenial: ./tools/render-cloudcfg --variant ubuntu config/cloud.cfg.tmpl
[21:08] <blackboxsw> not on focal
[21:08] <Odd_Bloke> I have never seen these Make targets before. :p
[21:08] <blackboxsw> yeah I hadn't until I started fgreping.   so something else in the template engine is getting hung up on the $user in the config template file
[21:09] <blackboxsw> I think it may be focal only
[21:09] <Odd_Bloke> blackboxsw: What do you want me to run to reproduce?
[21:09] <blackboxsw> checking eoan /bionic I might have a trivial patch
[21:09] <blackboxsw> Odd_Bloke   ./tools/render-cloudcfg --variant ubuntu config/cloud.cfg.tmpl
[21:09] <Odd_Bloke> And what are these used for?  (Should we remove them?)
[21:09] <blackboxsw> should emit output to stdout with the rendered template
[21:09] <Odd_Bloke> It does.
[21:09] <blackboxsw> Odd_Bloke: they are used in setup.py to build the package
[21:10] <blackboxsw> but we probably can drop the makefile target
[21:10] <powersj> blackboxsw, if these were broken, how did you upload to focal?
[21:10] <Odd_Bloke> OK, so we probably shouldn't remove them. :p
[21:11] <blackboxsw> powersj: I think something in the python environment used though dh_python is succeeding there whereas direct script running through python3 does not.
[21:11] <powersj> ah
[21:11] <Odd_Bloke> blackboxsw: I can't reproduce a problem on focal.
[21:11] <blackboxsw> ok so something from where I sit
[21:11] <blackboxsw> hrm
[21:12] <blackboxsw> ahh I'm on eoan :/
[21:12] <blackboxsw> checking eoan and focal containers
[21:13] <blackboxsw> and upgrading my laptop
[21:14] <blackboxsw> ok success on focal
[21:18] <blackboxsw> ok; dirty local dev environment. Fresh eoan no problems. something I've inevitably pip installed that is colliding w/ the renders
[21:18] <blackboxsw> Odd_Bloke: powersj n/m thanks for validation
[21:18] <powersj> :D
[21:18]  * blackboxsw should work in a clean dev container to avoid thrashing 
[21:18] <meena> someone post an image link or a github thingy pls
[21:22] <blackboxsw> https://imgflip.com/i/3k3kft
[21:31] <meena> excellent, previews are working.
[21:31] <meena> now, to get back to refunctoring.
[21:37] <meena> whoa
[21:37] <meena> Python2.7 is dead?!
[21:38] <Goneri> <Odd_Bloke> meena: Goneri: It's landed! <3
[21:39] <Odd_Bloke> meena: Yep! \o/
[21:40] <meena> Odd_Bloke: can i get some starting help with the refactor?
[21:40] <blackboxsw> long live python 3
[21:40] <meena> just to get the class(es) started, i can do the rest myself.
[21:41] <meena> blackboxsw: ETA for Python 4? :P
[21:44] <Odd_Bloke> meena: This is wrong point in the decade to ask for help, I'm afraid. ;)
[21:45] <Goneri> Odd_Bloke, thank you, now I will refresh the netbsd branch :-)
[21:45] <meena> Odd_Bloke: in that case, i'll just repeat all of Distros mistake, and you get to fix them next decade.
[21:45] <Odd_Bloke> meena: I would experiment with moving one of the two "*_on_{freebsd,linux}" functions onto either the Distro or Renderer classes.
[21:45] <Odd_Bloke> Which one they go on will probably depend on what makes sense in the structure of the code.
[21:45] <Goneri> I think we need a distros/linux.py and some inheritance
[21:46] <Goneri> and a lot of linux specific content should be move from net/__init__.py to this new distros/linux.py
[21:46] <meena> Goneri: that's what distros/__init__.py is.
[21:46] <Goneri> distros/__init__.py is linux specific, it should be agnostic
[21:47] <Odd_Bloke> On the Distro would seem more natural to me, as the functions aren't to do with the rendering of network stuff.
[21:47] <Odd_Bloke> I would suggest that refactoring them onto an existing class would make sense.
[21:48] <Odd_Bloke> Refactoring to have a Linux Distro and an agnostic, shared super-class is something I would support, but we would need to discuss that as a project.
[21:48] <Goneri> a large part of net/__init__.py is actually just some Linux specific code.
[21:48] <Goneri> Odd_Bloke, yes maybe, it's indeed already a good step in the right direction.
[21:49] <Odd_Bloke> As a reminder, the intent of the exercise is to gather data about what refactoring looks like it would make sense.
[21:49] <Odd_Bloke> So we don't need it to be Done (TM).
[21:50] <Odd_Bloke> We just need to be able to see roughly what it would look like, and modifying the existing distros.Distro to add functionality and overriding it in freebsd.Distro would do that, I think.
[21:50] <meena> Good point.
[21:50] <meena> points.
[21:50] <Goneri> IMO, a good metric is the number of is_FreeBSD in the code base
[21:50] <Goneri> the less, the better
[21:51] <Odd_Bloke> Agreed.
[21:59] <meena> Odd_Bloke: how do we access the distro from cloudinit/net?
[22:03] <Odd_Bloke> meena: That will be one thing that will need to be worked out.
[22:05] <meena> Odd_Bloke: >_>
[22:05] <meena> okay.
[22:08] <Odd_Bloke> (=
[22:12] <meena> i wonder if we can do that with decorators… maybe. probably not.
[22:49] <meena> maybe the easiest way would be to
[23:02] <meena> hmm… a class?
[23:19]  * meena gives up for tonight o/~