[07:37] <nfdeswqa> Haha, wow! What a fun time I had tonight. Turns out Kylo Ren is Han and Leia's son, Ben but was seduced to the dark side. He even kills his own dad at the end.. It was really tense. Oh and Rea finds out she has jedi powers and does a mind trick on a storm trooper to escape captivity. She kicks Kylo Ren's ass with a lightsaber too! Luke Skywalker only shows up for 20 seconds at the end though which is kind
[07:37] <nfdeswqa>  of lame. Oh well.
[14:01] <bloodearnest> heya folkas
[14:02] <bloodearnest> I have a very random question: can I control the mongodb oplog size with the manual provider?
[14:02] <bloodearnest> the code supports it (the local provider sets it to 1MB)
[14:02] <bloodearnest> but I can't see if there is anyway for me to configure it
[14:05] <bloodearnest> motivation: in an empty environment, oplog takes up 1.5G on it's own
[17:39] <mbruzek> marcoceppi: pmatulis has some questions about the use of "schema" in the action context in our new documents.
[17:39] <marcoceppi> mbruzek: what do you mean?
[17:40] <mbruzek> my understanding is it is just a format for the yaml file.  But there seems to be a --schema flag to the juju action commend.
[17:42] <marcoceppi> mbruzek: I don't see a --schema flag?
[17:43] <marcoceppi> mbruzek: you mean on defined?
[17:43] <mbruzek> https://github.com/mbruzek/docs/blob/developer-guide/src/en/developer-actions.md
[17:43] <marcoceppi> mbruzek: tldr, the actions.yaml file is a schema, it's a jsonschema
[17:44] <marcoceppi> mbruzek: what's the problem? all of that is correct
[17:44] <mbruzek> pmatulis: can you clarify your question?
[17:45] <pmatulis> a couple of questions now
[17:46] <marcoceppi> okay
[17:47] <pmatulis> ① is there such a thing as a "Juju schema"?
[17:47] <pmatulis> ② must actions.yaml have a json-schema format? the doc i have says "actions support json schemas" as if it also supports something else
[17:47] <pmatulis> (re ① or "Juju actions schema")
[17:49] <marcoceppi> pmatulis: the actions.yaml is a "jsonschema" which follows a specific format but supports the inclusion of additional jsonschema directives. an action does not have to define additional schema, but it can (IE, everything under params) an action without params is just an action but the actions.yaml follows a schema
[17:50] <marcoceppi> the fact that it's a "jsonschema" is misleading, since YAML is a superset of json and we use yaml in juju we simply declare it as the schema
[17:52] <pmatulis> wow ok. i'll have to knead that into something that an average user can understand
[17:52] <marcoceppi> pmatulis: this isn't for an average user
[17:52] <marcoceppi> it's a schema, you follow it, it's for developers
[17:52] <marcoceppi> it would be like saying "this is the actions.yaml format"
[17:53] <marcoceppi> the fact that the word schema is used is irrelevant
[17:53] <marcoceppi> except that it defines the constraints on the format used
[17:53] <marcoceppi> we didn't make our own format up like in config.yaml we use an existing format
[17:54] <pmatulis> right, that was my original question to mbruzek. but then the '--schema' flag i found threw me. is that also used colloquially?
[17:55] <marcoceppi> pmatulis: also, fwiw, this is what is /currently/ in the docs. https://jujucharms.com/docs/stable/authors-charm-actions
[17:55] <marcoceppi> pmatulis: the --schema flag simply dumps the actions.yaml file when running juju actions defined
[17:56] <pmatulis> marcoceppi, mbruzek: ok, my questions have been answered. thanks guys
[17:57] <marcoceppi> pmatulis: cool, let us know if we can be of any other help!
[18:47] <tvansteenburgh> was the 'machines' block in a bundle file introduced in v4 bundles? or did it exist before that?
[18:51] <rick_h_> tvansteenburgh: it came about in v4
[18:51] <rick_h_> tvansteenburgh: it was part of the new format to add that and the machine view in the GUI
[18:51] <tvansteenburgh> rick_h_: thanks!
[18:52] <rick_h_> tvansteenburgh: how goes using the new stuff?
[18:52] <rick_h_> tvansteenburgh: things going ok/making sense/going to work out?
[18:53] <tvansteenburgh> rick_h_: so far so good, i've been pestering urulama when i have questions
[18:53] <pmatulis> marcoceppi, mbruzek: in developer-actions.md, there is an example that uses some parameters that have not been mentioned/defined
[18:53] <rick_h_> tvansteenburgh: cool, make sure to shout if things can/should work better/different
[18:53] <pmatulis>   required: [outfile]
[18:53] <pmatulis>   additionalProperties: false
[18:53] <tvansteenburgh> rick_h_: will do, thanks!
[18:54] <pmatulis> marcoceppi, mbruzek: any reason why that should be there? i would like to remove it
[18:55] <marcoceppi> pmatulis: why do you want to remove those?
[18:55] <mbruzek> pmatulis: someone else wrote that document, I wouldn't delete the content, those are valid examples as far as I know
[18:55] <pmatulis> marcoceppi: like i said above
[18:56] <marcoceppi> pmatulis: required and additionalProperties is like description, they're part of the parent schema
[18:56] <mbruzek> pmatulis: perhaps the better answer is to define/mention those parameters
[18:56] <marcoceppi> the actions schema*
[18:56] <pmatulis> marcoceppi: why put stuff in there when they don't mean anything at that time
[18:56] <pmatulis> readers hate that
[18:56] <marcoceppi> pmatulis: literally the next paragraph defines those keys
[18:56] <pmatulis> looking
[18:56] <marcoceppi> JSON-Schema defines special keys such as required and additionalProperties, which may be given for the whole action at the same level as description and params, or within nested schemas at the usual level.
[18:57] <marcoceppi> I suppose we could remove it from that example, since it's repeated below
[18:57] <pmatulis> yeah
[18:58] <marcoceppi> but it's really just a more complete example
[18:58] <marcoceppi> also, this doc is almost a 1:1 match with what is currently in the repo
[19:01] <pmatulis> marcoceppi: right, i'm fighting a battle on 2 fronts, not easy
[19:46] <cory_fu> I am finding JUJU_REMOTE_UNIT is set to an empty string during the -relation-broken hook, leaving me no way to tell with whom I have lost the relation.  Is this expected behavior?
[20:11] <marcoceppi> cory_fu: I believe so
[20:43] <pmatulis> what are the minimum requirements for a charm action?
[20:43] <pmatulis> so far, it seems like
[20:43] <pmatulis> ① /actions/binary file
[20:43] <pmatulis> ② top-level key named according to binary name in actions.yaml
[20:43] <pmatulis> is there anything else required in actions.yaml for an action to work?
[20:57] <marcoceppi> pmatulis: that is the bare minimum
[20:59] <pmatulis> marcoceppi: thanks
[23:03] <mbruzek> tvansteenburgh: or cory_fu: how do I activate a python virtual env in a Makefile?  source .venv/bin/activate does not seem to work
[23:04] <cory_fu> Is this for running tests?
[23:04] <mbruzek> make lint
[23:04] <cory_fu> mbruzek: I'd highly recommend using toxx
[23:04] <marcoceppi> mbruzek: a more direct answer, you really can't "activate" a virtualenv in a Makefile, but you can run things like .venv/bin/python or .venv/bin/pip
[23:05] <mbruzek> ack
[23:05] <cory_fu> mbruzek: Take a look at https://github.com/juju-solutions/charms.reactive/blob/master/Makefile and https://github.com/juju-solutions/charms.reactive/blob/master/tox.ini
[23:05] <mbruzek> thanks cory
[23:05] <cory_fu> np
[23:06] <mbruzek> cory_fu: the goal was to self contain this thing, I don't see the makefile installing tox
[23:07] <cory_fu> mbruzek: You're right, I assume tox.  You can just apt-get install python-tox though