[05:06] morning all [05:07] facubatista, you don't need 'install' because if you are on 2.7 it will only call 'start' and if you are on 2.8 it will only call 'dispatch' [08:32] moin moin [08:44] morning Chipaca, how goes? [08:45] jam: a lovely day to just sit by a lake and watch the waves lick the shore [08:45] not feeling particularly energetic [08:45] Chipaca, get pumped! [08:45] did that help ? :) [08:46] * Chipaca reaches for the bicycle pump [08:46] and then do _what_ with it?!? [08:46] Chipaca, doesn't that already make you feel more energetic? [08:46] y [08:47] :-p [08:47] jam: how are you btw? [08:47] (disclaimer: do not inject a pump directly into your skin, do not look sideways at pump, do not engage with BicyclePump(tm) without written consent from your doctor) [09:07] Chipaca, model_name = '' would be valid, though I like None as breaking stuff rather than silently not doing what you want [09:33] Though if you go back to the thread about "when should we error" there may be a reason to not raise exceptions [09:40] Chipaca, I just noticed. I'm doing well. I did run into something odd [09:40] So in StatusBase, we actually don't populate the map until you instantiate an instance (the population is done in __new__) [09:40] which means that your new test it test_testing.py fails if you run *only* that module [09:40] ./run_tests test/test_testing.py [09:41] do we need a Meta class so that inheriting from StatusBase adds it to the map instead of creating an instance? [09:41] I'm a bit surprised we don't run into problems at runtime with someone checking Unit.status and it not know about 'unknown' or some other status type [09:45] Chipaca, this https://paste.ubuntu.com/p/fHJsQ66Ydm/ "fixes" it, but I think it is just papering over the problem [09:49] Chipaca, I filed https://github.com/canonical/operator/issues/288 so I don't forget about it. [09:51] jam: thanks [09:51] my brain is in a different castle rightnow [09:52] Chipaca, np. btw I had a response to "do we need an install symlink" which I don't think we do, bring it up sometime when you're ready to think about it [10:18] jam: why don't we need an install symlink? [10:28] Chipaca, the version of Juju where 'install' is called is also the version that supports 'dispatch', IIRC [10:29] so you need install or dispatch, but not both [10:29] Chipaca, I know we have the meeting with Tom in a bit, do we need to do another prep session for that [10:30] jam: that's just in k8s though, right? [10:30] jam: outside of k8s it's install and upgrade-charm [10:30] inside of k8s its start and upgrade-charm [10:30] for pre-2.8 [10:31] Chipaca, right, 'install' for non-k8s and 'start' for k8s [10:31] jam: so without knowing more, you need all of 'em [10:32] Chipaca, yeah, I was thinking through how to communicate it. It feels bad to populate a lot of things "just in case" and then have the framework be populating the rest. [10:33] Chipaca, btw, for the meeting with Tom, should we be doing prep work for that? [10:33] jam: this is a first meeting (tom is proxying paul collins) [10:34] gotcha, I see that now in the notes [10:34] Chipaca, are we intending to copy that over to the charmcraft calendar? [10:35] jam: i was hoping to get that handed over so we could move it there, yes [10:35] copying is not useful as it doesn't carry moves etc over [10:36] Chipaca, sure. "replacing with a copy". (I'm not able to edit the description to add the notes doc, etc.) [10:37] actually, let me try to sort that now :) [10:38] mthaddon: would you mind passing ownership of the calendar event for today's meeting to me? that's inside Edit, under More actions (beside save), at the bottom [10:39] Chipaca: sure, done [10:40] of course then i need to add /b/1 to the URL the calendar tells me to click :) [10:41] mthaddon: thanks! [10:41] jam: all sorted [10:41] added the doc [10:41] Chipaca, thanks [10:57] Muy buenos días a todos! [11:01] Chipaca, coming? [11:01] morning facubatista [11:01] hola jam :) [11:02] jam: yeah, somebody came to the door [11:16] WOO I GET TO HAVE LUNCH [11:16] * Chipaca hugs mthaddon [11:17] o/ [12:53] jam, facubatista, the reason we don't have a postgres meeting next week is because we're down two people on the monday :) [12:53] Chipaca, I responded to 208, I'm curious whether we want to close it/tag it/something else [12:55] jam: tag it enhancement for now at least [12:55] later we might have an opslib tag [12:58] jam, facubatista, actually we're down 3/3 next monday [12:58] may 25th international holiday (of mystery) [13:00] Chipaca, standup? [13:00] yep 1 sec [13:52] (and I've been doing reactive stuff, so no point) [13:53] * Chipaca hugs stub [13:53] stub: we still like you even if you're working on other stuff [13:53] … for now [13:54] I'm less trouble when I'm working on other stuff ;) [14:38] stub: I cannot comment on that [14:38] facubatista: you got 5? [14:38] Chipaca, yes [14:38] he might even have 6 [14:38] gasp [14:39] jam: hey if you've also got 5, let's meet in the standup thing [14:39] omw [16:47] jam, Chipaca, I don't like "directory" as the option name for the charm's project base directory, "directory" is way too generic [16:59] facubatista: FWIW that one seems strange to me WRT having it in the config file [17:00] "basedir"? "base-dir"? "charm-dir"? (this may be confused with the "src" dir which is not what we mean) "project-dir"? "project-basedir"? "project" [17:00] ? [17:03] facubatista: let's see if gustavo has opinions on this :) [17:04] facubatista: --from , --at [17:04] facubatista: charmcraft build --what ./this [17:04] the "from" is nice after the "build" verb [17:04] yep [17:05] --at has the same thing, but it's less directional which might be confusing (is it going to read from that, or put the charm there?) [17:12] Chipaca, I started the "config files" issue and end up writing a lot there :/ [17:12] Chipaca, https://github.com/canonical/charmcraft/issues/5 [17:13] facubatista: ok, I'll read it later :) now I'm going to run away [17:13] facubatista: or maybe I'm going to go away, and run [17:13] facubatista: you decide :-) [17:13] * facubatista decided