[09:01] <Chipaca> moin moin moin
[09:01] <bthomas> Mornig Chipaca
[09:01] <Chipaca> bthomas: how're you doing?
[09:03] <bthomas> Chipaca: I am good. I am reading through Operator Framework module in an effort to understand and put doc strings where they are missing. If I figure out things left as TODOs in hook docs I will update that also. You may get a question or to when I am utterly confused :).
[09:03] <Chipaca> ok :)
[10:20] <bthomas> Chipaca: pushed a very rough first draft for charm.py (https://github.com/balbirthomas/operator/tree/improve-docs) . Starting on framework.py after a short break. No need to review now. Just a heads up in case you want to guide and steer my efforts in any particular direction.
[10:21] <Chipaca> ack
[11:18] <facubatista> ¡Muy buenos días a todos!
[11:19] <bthomas> morning facubatista
[11:19] <facubatista> hola bthomas
[11:19] <Chipaca> facubatista: 👋❗
[11:28] <facubatista> hola Chipaca
[11:33] <facubatista> @niemeyer, one drawback of moving the python dependencies out of requirements.txt to a charmcraft.yaml (with the rest of the config options) is about running tests... normally people would have a requirements-dev.txt with dependencies to run the tests (which would automatically include those in requirements.txt)
[11:34] <niemeyer> facubatista: How do we run tests when the setup is all being described inside charmcraft.yaml?
[11:34] <facubatista> niemeyer, unless we provide a way to run the tests through charmcraft itself... like `charmcraft test`
[11:35] <facubatista> but we need `charmcraft test` to be fast, it can't create a VM for every run
[11:35] <niemeyer> facubatista: Right.. without something like that, it means the environment you use to run tests, and the environment you use to build and ship, are completely different
[11:35] <niemeyer> facubatista: Snapcraft and Multipass already have logic to make that kind of thing fast
[11:36] <niemeyer> facubatista: Ask Sergio about how machines are reused for fast iteration.. I don't know where that is today, but we discussed and something was implemented back then
[11:38] <Chipaca> it would be nice if 'test' were just another stage
[11:38] <Chipaca> (and that you could tell charmcraft to just run that one)
[11:39] <Chipaca> that plus per-stage requirements and you're there
[11:40] <facubatista> niemeyer, Chipaca, I opened this to track the ideas https://github.com/canonical/charmcraft/issues/160
[11:42] <Chipaca> aye
[11:43] <Chipaca> even tox is too slow for some people
[11:46] <Chipaca> how is it 12:45
[11:50] <facubatista> Chipaca, are you preparing a fix for the flaky test about times checking?
[11:51] <Chipaca> facubatista: i forgot! i'll do that right after lunch
[11:51] <Chipaca> facubatista: sorry
[11:51] <facubatista> Chipaca, ack
[11:51] <facubatista> thanks
[12:43]  * Chipaca runs the test suite 999 times
[13:27] <justinclark> Hello again!
[13:28] <bthomas> Hi justinclark
[13:28] <Chipaca> justinclark: welcome back!
[13:47] <facubatista> hola justinclark!
[14:18] <axino> Chipaca: re "harness.charm._stored.foo = 42 should dtrt?" - can this work if I want to test what's in the __init__ function of my CharmBase subclass ?
[14:19] <Chipaca> axino: sorry, i don't follow :-/
[14:20] <axino> Chipaca: I have https://pastebin.canonical.com/p/9RjtYGsCXP/ and I want to test both cases of the "if" line 19
[14:21] <axino> Chipaca: I'm not sure I can rely exclusively on harness for this
[14:23] <Chipaca> axino: unless i'm missing something harness.charm._stored.reldata will be {} after just begin, right?
[14:23] <axino> Chipaca: for the first ever hook run on a unit, yes. But for subsequent hooks, it could have another value (which is the whole point of StoredState() :) )
[14:24] <Chipaca> ah no because _init_postgresql_relation is called from by __init_
[14:24] <Chipaca> so .relation will be at least {'pg': {}}
[14:24] <Chipaca> reldata*
[14:26] <Chipaca> axino: i agree, i think the harness doesn't help you for this
[14:26] <Chipaca> axino: if you know how we could help, let me know :)
[14:26] <axino> Chipaca: alright :) I think I basically have to patch StoredState and instanciate the class and test the value of _stored
[14:27] <axino> Chipaca: I'll try things on my own for a bit - thanks
[14:28] <Chipaca> axino: doing .begin twice might work?
[14:29] <axino> Chipaca: "RuntimeError: cannot call the begin method on the harness more than once"
[14:29] <axino> :)
[14:29] <Chipaca> 1 sec
[14:35] <Chipaca> axino: aw, a second harness does not pick up the stored state from a previous one
[15:29]  * facubatista -> lunch
[15:55] <Chipaca> ok, i'll bbl
[15:55] <Chipaca> going for a run while the sun shines :-D
[16:39]  * facubatista begs for review: https://github.com/canonical/charmcraft/pull/163
[16:40] <mup> PR charmcraft#163: Two small details before integrating help messages to main <Created by facundobatista> <https://github.com/canonical/charmcraft/pull/163>
[16:40] <bthomas> facubatista: having a look
[16:48] <bthomas> done
[18:04]  * facubatista -> tennis
[18:17]  * Chipaca blames facubatista for https://github.com/cloudflare/cobaul
[20:53]  * facubatista is back
[20:54] <facubatista> Chipaca, what did I do?
[21:27] <facubatista> jam, what is the proper name for the "human operator" in the Juju world? I mean, the person that executes juju commands to deploy and maintain the system...
[21:29] <jam> facubatista, it used to be "operator". The best word that I've found to replace it is the Admin/Administrator
[21:29] <facubatista> jam, administrator it is, thanks!
[21:56]  * facubatista eods