wallyworldkelvinliu: no rush, here's a small juju run fix for the new actions UX https://github.com/juju/juju/pull/1197802:59
hpidcockkelvinliu: when you get a chance, no rush https://github.com/juju/juju/pull/1197905:01
stickupkidare we always required to have a metadata in a charm?08:52
stickupkiddo we have any documentation about what a charm is, at it's raw basic form?08:56
stickupkidso i can crib from it08:56
stickupkidachilleasa, https://github.com/juju/charm/pull/31809:47
stickupkidif I can't find it, then it might help if I write what I found09:48
achilleasaapproved; you also need to land against V209:59
achilleasaah... nvm it's charm not bundlechanges... doh!10:00
achilleasastickupkid: https://github.com/juju/bundlechanges/pull/6710:20
achilleasawith lots of extra tests :D10:20
stickupkidachilleasa, https://github.com/juju/bundlechanges/pull/67/files#diff-f447f119094ea333680948095a672515R76910:24
stickupkidcan't you using thisSet.Difference()?10:24
stickupkidor are you wanting to exit early?10:24
achilleasadifference allocates and populates a new map. Given that the comparison process for the grouping is pretty much O(n^2), I opted for early exit plus 1 less allocation10:28
achilleasathough in principle it shouldn't matter much as the number of endpoints is usually small10:28
stickupkidachilleasa, i'd say, lets use difference and if performance is an issue swap to that10:29
stickupkidpremature optimisation etc10:29
achilleasaok, let me amend the last commit10:33
stickupkidok, why does GUIArgs change when doing the exposed related stuff?10:38
achilleasait seems that the general pattern is that we add new fields to guiargs as well10:40
achilleasaquestion is whether that is being used now by the new gui or not10:40
stickupkidGUIArgs are dead aren't they10:40
stickupkidjam, will know10:40
achilleasanot sure :D10:41
achilleasaso I just add extra things to be consistent10:41
stickupkidconsidering we're bumping bundlechanges, we should consider if GUIArgs is required anymore?10:44
achilleasaas this is V3, we can land this if it looks OK and I can raise it in standup. If we don't need it we can kill it in master10:45
stickupkidachilleasa, got one for you https://github.com/juju/charm/pull/31911:15
stickupkidachilleasa, this encompases the changes we discussed yesterday (dropping the ch: prefix for urls)11:15
stickupkid* when using String()11:16
achilleasastickupkid: you bumped the application facade to V13 on develop right? (to add CharmOrigin)12:01
stickupkidachilleasa, no idea why they didn't just go with https://github.com/juju/charm/pull/319#discussion_r48571716915:48
stickupkidachilleasa, they would have been named as well to make it much more readable15:49
achilleasaI bet this will come back and bite us some time in the future and we will have to use versioned URLs...15:56
stickupkidachilleasa, we have 3 already16:08
jamstickupkid, achilleasa : 'username@" wouldn't work there, because that would be "who am I connecting to the store as" not "who owns this charm".16:37
achilleasajam: couldn't that simply be another query param? e.g. &author=foo16:38
jamachilleasa, it certainly could, though in the end we're moving to the Snap model where the namespace itself doesn't show off the author16:39
jameg, it isn't jameinel/juju it is just 'juju' and the owner is off in the details of the project16:39

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!