=== rogpeppe1 is now known as rogpeppe === schwuk_away is now known as schwuk [13:05] shhhhhhhh [13:19] frankban: ping [13:20] TheMue: hi [13:21] frankban: hello, it's me again [13:22] frankban: I've rolled back the changes regarding ServiceSet due to the API incompatability [13:23] frankban: but william told me that you're thinking about an API change make larger settings/unsettings at once [13:23] frankban: am i right? in that case we could also use this new API call to allow the explicit unsetting [13:24] TheMue: as gary_poster mentioned, for the time being we can live with that incompatibility, meaning the GUI users will set string options to an empty string, or receive an error for bool/int/float fields [13:26] frankban: the problem is that other parts of the software use the API too, so it should not get incompatible [13:26] TheMue: you mean, other parts of juju-core? [13:27] frankban: yep [13:28] fwereade_: re large settings/unsettings at once are you referring to the ServiceUpdate for multiple services? [13:36] TheMue: anyway, I believe you are right, we can take care of options unsetting when we improve the service API calls [13:38] frankban: yes, so we should stay in contact then [13:38] TheMue: sure [13:38] frankban: any ideas about it already or do we start on a green field? [13:39] frankban: and i have to find a good we to allow setting string options to empty strings by cli without changing the api behavior [13:39] TheMue: I think William already has some ideas for ServiceUpdate and ServiceDeploy, and some design decision is required [13:39] abentley, sinzui: could one of you have a look at this MP: https://code.launchpad.net/~adeuring/charmworld/spurious-failures/+merge/182901? [13:40] frankban: ok, will ask William when he's back [13:42] adeuring: sure. [13:55] rick_h: could you please take a look at https://codereview.appspot.com/13386043 ? it's the chrome dev fix. [13:57] Makyo: review done [13:58] hatch, thanks. [13:58] hatch, none of the others are sprited, can we defer that to a separate branch? [13:59] A sprite-ageddon, if you will. [14:00] Will respond to comments instead of here, sorry. [14:05] frankban, hey, sorry, I don't recall what the progress was on ServiceUpdate [14:06] frankban, I had a feeling you were doing something with it [14:10] Makyo: yep np [14:10] frankban: sure thing [14:10] fwereade_: ServiceUpdate is implemented for a single service. And for now it takes SettingsStrings as a map[string]string, so there is no way to unset an option. TheMue proposed a new API call for that. Anyway, I had the crazy idea of changing SettingsStrings to be a map[string]*string, not sure that applies well. [14:10] rick_h: thanks [14:12] fwereade_: from the GUI perspective, rather than unsetting an option, we could set the default value for that option. [14:14] frankban: unsetting an option is one part of the game, but setting a string option to an empty string another one [14:16] frankban: and this shouldn't been done with ServiceSet, because here option="" is interpreted as unsetting it. and this should not be changed [14:18] TheMue: yes I know, I was just saying that, from the GUI perspective, a default value in the juju-core db == no value. And that we can react without too may problems to possible changes in how Service(Set|Update) are implemented. [14:19] frankban: ok [14:22] rick_h: you around? [14:22] hatch: yep [14:22] with that lovely post-dentist "I've been touched in strange places" feeling :P [14:23] just a very small FYI on your zIndex branch - the sidebar now shows above the 'modal' blackout when removing relations and the dialogue pops up [14:23] I hate modals so I would remove the modal [14:23] buuuuuut [14:23] hatch: ok, well, they should be adjusted. I tried to set a 'standard' for where z-indexes should lie and they can be adjusted clearly now. [14:24] hatch: I added a doc chunk to the top of the stylesheet.less walking through it [14:24] should be a two-liner [14:24] ahh cool cool [14:25] I just noticed it in QA but it was definitely not blocking lol [14:25] hatch: k, yea my bad. I tried to look over anything with a manual z-index to make sure it looked safe, but was hard to tell [14:25] yeah that's dynamic I think [14:26] frankban: looks good, small note on the doc comment [14:26] working from the laptop today....boy is the console ugly [14:26] * hatch can't wait for rick_h to fix it for me [14:26] bac, sorry for missing time. May I move our call to 2 PM [14:27] hatch: console is ugly? [14:27] oh, your tmux/zsh issues? [14:27] yeah I tried to customize it....and broke it [14:27] oh bac not here right [14:27] forgot [14:27] gary_poster: one review+qa is also valid for charm branches, correct? [14:28] rick_h: thank you! [14:28] frankban, sure. we haven't agreed to it yet, but sure, give it a try and report back. ;-) [14:28] ok thanks :-) [14:29] gary_poster: and, if it's not clear, I am +1 on that proposal, including possible team reviews of parts of the code, as discussed in the juju ml [14:29] frankban, cool :-) [14:31] frankban: thank you for fixing it up. My use of chrome dev can continue! [14:31] nice [14:32] :-) [14:33] hatch: ooh, new YUI and Template landed [14:33] ok so at our rate in about 3 months I'll get to upgrading it :P [14:34] rick_h: juju-core + lxc + juju switch is really cool. and, for when you need pyjuju, a virtualenv + virtualenv wrapper + a single symlink work well ("workon pyjuju" and you are done) [14:47] hi bac: Would you have time today to review https://code.launchpad.net/~sinzui/charmworld/support-gui-charm-urls/+merge/182924 [14:58] brb running to drop the car off at the shop, back in 15 [15:51] jujugui call in 9 kanban now [15:56] jujugui, I'll attend http://summit.ubuntu.com/uds-1308/meeting/21899/servercloud-s-juju-new-user-ux/ which conflicts with our standup. :-/ I feel like I need to be at both. Makyo, would you mind running and taking some basic notes of anything important not covered by the kanban? [15:56] gary_poster, sure. [15:56] thank you Makyo [15:58] jujugui call in 2 [16:12] abentley, benji : Would you have time today to review https://code.launchpad.net/~sinzui/charmworld/support-gui-charm-urls/+merge/182924 [16:17] jujugui can I get a review/qa https://codereview.appspot.com/13249045/ [16:22] bcsaller: did you QA? [16:22] thanks for the review [16:22] I didn't [16:24] you should :D [16:28] hatch: I can qa, looking at now [16:28] cool thanks [16:36] rick_h: I was thinking the same thing as your perf comment and figured that it was virtually insignificant either way [16:38] hatch: yea, I guess I just wanted to raise it. [16:41] jcsackett: your next dev environment: http://www.youtube.com/watch?feature=player_detailpage&v=8SkdfdXWYaI#t=545 [16:43] benji: that's real cool === schwuk is now known as schwuk_away [16:59] benji: that is awesome. [16:59] :) [17:00] rick_h: so a couple days of use with Sense 5 and I can tell you that it has some cool features but I muuuuuuch prefer stock android [17:01] hatch: yea, I'm all nexus/stock any more [17:01] I'm hoping I can get the Zoe camera app apk for the stock OS [17:01] but I won't do that until I'm back home [17:02] dont' want problems while away hah [17:14] I wish tracking numbers had a delivery time [17:21] yuss finally someone else who agrees with me that the search is broken :P [17:21] our power is growing [17:23] gary_poster: I was thinking of tackling 1214087 but the comments are a little confusing [17:23] https://bugs.launchpad.net/juju-gui/+bug/1214087 [17:23] <_mup_> Bug #1214087: GUI fails to deploy keystone [17:24] it reads to me that it's a charm problem but then you want us to only send changed config values? [17:30] hatch, yes [17:30] hatch, you see why? [17:30] yep [17:30] so should this ticket be renamed then? [17:30] hatch, fwiw, openstack is quite arguably our most important Juju story right now [17:31] and I'm guessing keystone is part of that? [17:31] heh [17:31] * hatch googles keystone [17:31] :-) yeah [17:31] oh yeah that is a pretty important part [17:31] :-) [17:32] ok I'll add that to the config save [17:32] hatch I wouldn't rename bug because both should happen: we can only send changes, because that is reasonable; and keystone charm should not be so persnickety :-) [17:32] cool hatch. ghost config is more important than post-deploy config if you have to choose [17:33] thank you [17:34] https://bugs.launchpad.net/charms/+source/keystone/+bugs odd this picks up the wrong status for our ticket [17:41] gary_poster: in attempting to reproduce the issue using rapi I get an error that it can't understand the constraints...not the config error [17:42] I'd like to reproduce the error first so I know that it's fixed or not [17:42] hatch ack on call. could be that the "undefined" thing is a problem [17:42] ok I'll peek into that first [17:57] ok the error is `juju.errors.ConstraintError: Could not interpret u'mem' constraint: need more than 1 value to unpack` [17:58] I should mention `u` prefixes every json key value [18:01] the data being sent is correct [18:03] hatch: that's a python thing, ignore it [18:03] hatch: basically saying it was a string 'mem' [18:03] u'mem' means unicode string [18:04] ohh ok [18:04] the syntax looks to match the syntax in the python tests [18:05] rick_h: looks like we can't deploy anything on rapi in the new inspector [18:05] because of the constraints [18:06] do you have time to pair on this? [18:06] hatch: sure thing [18:06] * rick_h can catch up [18:06] cool guichat [18:30] gary_poster: here is the constraints issue https://bugs.launchpad.net/juju-gui/+bug/1218555 working on it now [18:30] <_mup_> Bug #1218555: Constraints are sent in the incorrect format with new inspector to rapi [18:44] $50 to get the unlock code from my carrier..... [18:44] Main laptop battery: ~40 mins. Backup laptop battery: ~20 mins. I hope the flight to London involves power outlets :| [18:44] * hatch searches for a cheaper code [18:44] Makyo: :'( [18:45] ack and thanks hatch [18:45] * gary_poster not really here for a bit, because of early morning/late day Thursday [18:56] benji, abentley . I deployed your revisions to staging. Mine works, but I did not notice the bread crumbs changed: http://staging.jujucharms.com/~gnuoy/precise/apache2-1 [18:57] I... don't think I know anything about bread crumbs. [18:57] sinzui: having bad internet weather here. [18:57] benji, abentley. Maybe I should put "revision" in the breadcrumbs. [18:58] abentley, floods again? [18:58] sinzui: No, just random slowness. Sunny today, actually. [18:58] the "Apache2 1" at the end is suspect [18:59] benji, my current branch will reverse the breadcrumbs and put them in the page title, I don't think revision is important, though we might want to keep it because it distinguishes it from the versionless charm [19:00] BC: Home/ Gnuoy/ Precise/ Apache2 revision1 [19:00] yeah, I think including the revision helps people know that they're looking at a particular revision which would seem to be important [19:01] Title: Apache2 revision 1 : Precise : Gnuoy : Juju Charms. [19:02] Oh, we still say "Charm Browser" in the heading. That must say Juju Charm Browser or just Juju Charms [19:09] bcsaller: I saw Makyo reviewed. So if we can't dedupe the code then is it cool to leave current feedback as more tests required then? The code on things like sub relation indicators, exposed indicators, positioning tests, x/y annotation verification, etc need to be added? [19:18] rick_h: thats cool, yeah. In the end more of this will be broken out into smaller functions and the level of reuse can go up, and there are somethings like sizes that should become configuration rather than hardcoded values [19:20] bcsaller: k, I don't want to hold up things and wanted to double check what the plan is from here then. [19:21] orangesquad or benji: Could you please review https://code.launchpad.net/~abentley/charmworld/prefix-bundles/+merge/182992? [19:21] rick_h: Honestly I expect that the drawing code in there will undergo quite a bit of change once UX sees it, there is very little info on what people want, this is more a 'show capabilities' branch for now [19:22] abentley: I'll take a look [19:23] benji: Thanks. [19:51] benji: What trailing slash do you mean? [19:52] abentley: the import_filter is a prefix of the branch path, right? [19:53] benji: right. [19:54] abentley: therefore, a filter of "XXX/foo" would match "XXX/foo/bar" and (erroniously) "XXX/foozzle", right? [19:54] benji: I wouldn't consider the second erroneous. [19:55] hmm, maybe I misunderstood the purpose of _filtered_payloads. I thought it was meant to match on path segments, not just on a character-by-character basis [19:57] benji: No, it is just a prefix. [19:57] abentley: so, would it ever make sense to use it to match part of a path (like my second example)? [19:57] benji: In reality, if it starts with "f", it has to be "feisty", anyhow. [19:58] ok, I guess it's not worth worrying about then === sinzui is now known as sinzui_walking [22:04] so hows everyones afternoon? [22:05] the shed base is reinforced, shelves setup, and more crap shoved in it. now to cool off in the basement, phew [22:05] gota love some good time in the yard [22:05] or garage [22:05] :) [22:05] yea, sprinkler folks moved the head under the shed out of the way so can put things back together [22:06] yea, I can't wait until winter to get back into the woodshop [22:06] I'm getting an aching for some woodworking. Need more hours in the day [22:06] yeah - winter is when I can work on the cars [22:07] summer should be the time, but who has the time! [22:07] right, as much fan as camping has been, looking forward to finishing up our last trip in Oct and winterizing it [22:07] I thought you park it indoors? [22:07] yea, but it's not heated. [22:07] just a large warehouse style building [22:08] so protected from the elements, but no promise it won't get cold in there [22:08] ahh gotcha [22:11] hatch: is your branch up for review? /me missed that go by [22:12] I am just lboxing now [22:12] from my laptop though so it'll take 4 evaaaa [22:12] ah, ok. [22:19] rick_h: I'd like it if you could QA this because it needs a QA on all three backends but I realize it's past your EOD https://codereview.appspot.com/13252045/ so tomorrow would be fine too [22:28] Morning [22:31] mornin huwshimi [22:58] hi huwshimi. my wife is still at dr. I'll ping you when I am available. hopefully soonish [22:59] gary_poster: No problems, any time. [23:15] bcsaller, did you ever get your necessary reviews? [23:15] gary_poster: Happy to postpone if it's getting to late for you. We'll be in the same room on Monday :) [23:15] gary_poster: one LGTM and now we are down to one with requests for tests paralleling the service module [23:15] huwshimi, heh, true, but I think we should talk. can do it now. you ready? [23:16] gary_poster: Yep [23:16] bcsaller, oh ok, so two reviews but one response in progress, cool thx [23:38] gary_poster: whenever you have a moment - do we have a mockup for the way the unit relation view is supposed to look like? Or just make it look decent? [23:39] hatch, no mockup. decent, based loosely on existing design (which is engineer design, but whatever :-) ) [23:39] hey, I've always said engineers develop the best UI because it's the most efficient lol