[00:13] gary_poster: so good news is that I was able to replicate what I had in my head - bad news is that it won't work for what design wants [00:14] the issue huw is having is a tough one for sure [00:18] hatch, lol, ok, darn. I wonder if it approximates enough of what they want to be a benefit? a good solution might be better than a perfect solution...never? [00:18] well the issue isn't that it's not possible, it's that it doesn't work with the current markup [00:18] so it's not just a case of adding some transitions [00:18] it'll take some markup and styling refactoring [00:25] gary_poster: if you're still around can you clarify the subordinate changes that were made in jujucore - were those only in the flagged ui or were respective changes also made for unflagged [00:25] * hatch is going through the bzr log [00:26] hatch, across the board [00:26] thx [00:26] hatch, subordinate support should be good unflagged, better than it was flagged [00:27] ahh ok [00:50] gary_poster: release is complete [00:50] I'm going to guess however this doesn't auto push to jujucharms.com? [01:05] hatch awesome thanks. correct, we need to ask IS to release first thing tomorrow morning, and hope that is enough time for us to deal with any issues. we made a pact not to release on Friday...maybe this isn't quite the same? dunno. :-) [01:33] gary_poster: haha, well maybe we should roll up the release EOD Wednesday so we have Friday to iron out any bugs noticed on Thursday :D [01:37] hatch, I'll send out http://ec2-54-226-124-150.compute-1.amazonaws.com/ for anyone to qa tomorrow. it looks good to me [09:31] buongiorno, dò un'altra occhiata al tuo branch [09:31] oops :-) === teknico1 is now known as teknico [11:55] gary_poster: morning, would you like to take a look at https://codereview.appspot.com/11841043/diff/16001/server/guiserver/tests/test_apps.py before I submit? Added that test file thanks to your comment. [12:17] frankban, yes, I saw the email, thank you! sure, looking. [12:21] frankban, LGTM! thanks again [12:33] gary_poster: cool thanks [12:33] * frankban lunches [12:35] rick_h, hey. you are here today? want to see if you want to follow up on Huw's branch email or have someone else do it; and I want to see if #1205136 is another instance of an existing/known bug [12:35] <_mup_> Bug #1205136: more charm browser navigation issues

[12:35] gary_poster: I'm looking at huw's branch currently. [12:35] cool, thanks rick_h [12:35] * rick_h goes to check out the bug [12:36] rick_h, also, did http://ec2-54-226-124-150.compute-1.amazonaws.com/look good to you? [12:36] http://ec2-54-226-124-150.compute-1.amazonaws.com/ [12:41] gary_poster: yea, that's the same bug as before. Doing a search, clicking on something, url updates, end up back at search results and state doesn't 'change' from there [12:41] rick_h, ack, thanks. so you are on a precursor to that bug, and that bug itself, today and ff IIRC? [12:42] gary_poster: the ec2 url looks good. yay :flags:/ac and looks ok on a general run through [12:42] 'and ff' ? [12:42] and following [12:42] yes, I'll be working on trying to solve all these once I get huw's stuff figured out [12:42] cool perfect thanks rick_h [12:42] but I'm getting test failures on trunk and I'm seeing huw's test failures out of no where so working on getting to it [12:43] rick_h, the "ff" came up from the subconscious--had to look it up :-P http://en.wiktionary.org/wiki/ff. [12:43] (was checking whether I was crazy or not :-P ) [12:43] gary_poster: heads up, beowolf from the U1 team and myself did some chatting about the JS help email thread. I wanted to check with you and team before we volunteer us as "JS peeps" [12:43] answer: only mildly crazy [12:43] gary_poster: heh, I'm so FF == follow friday or feature flag [12:44] heh [12:44] I couldn't line either up with the sentence [12:45] rick_h, JS help mail: cool. already followed up encouraging people to come by #juju-gui. I figure you'll ask hatch if he wants to speak up. After that wasn't planning on doing anything unless a reply from David inspires something [12:46] gary_poster: k, beowulf started #js and I brought up the old canonical-javascripters team that's there but not really up to date [12:46] I wasn't sure on people jumping in here so cool if that works [12:47] rick_h, not sure either but firgured we could offer it, and ask for something different if it doesn't work out. :-) [12:47] or maybe they will counter with a request [12:47] we'll see [12:48] yea, I figure just spreading the word that there are people out there will help [12:48] canonical-silos :) [12:48] heh [12:48] yeah [13:12] morning [13:12] replying to js email [13:14] cool thanks :-) [13:27] abentley can we chat after the call about how to construct put_mapping for bundles? specifically how to tease out the good parts of the services dict. [13:28] sinzui: Staging is down. I assume due to the lcy02 migration. We still control the staging IP address, but we seem to have no instances at all. [13:28] bac: sure. [13:28] :( [13:31] sinzui, gary_poster: y'all coming? [13:31] yay [13:32] thanks bac [14:06] hey gary_poster can you let benji know the charm school on testing is postponed. We'll be covering a different topic instead today [14:13] Huh, ISP turned off their DNS servers last night. Oh well! [14:13] jujugui 0.8.1 is live [14:13] jujugui redirect for unversioned charms are live [14:14] Makyo: 8.8.8.8 [14:14] bac, Yeah, switched to that [14:23] rick_h: can you look at https://codereview.appspot.com/11845043/ [14:24] jcsackett: yep, loading [14:24] hatch: when you get a sec I'd like to do a quick hangout [14:24] rick_h: sure [14:24] in guichat? [14:24] hatch: sure [14:25] in there [14:37] marcoceppi, will do thanks [14:49] hatch: and can you bring this up on the review call today? [14:49] yup [14:49] hatch: because I think I've seen this before of checking a node is an 'object' [14:49] probably my fault [14:49] hatch: I wonder how many other places we've had the potential for a test that never worked :) [14:52] jujugui, could I have comments on this blog post, which I will be sending out as an email soon? I'm particularly interested in whether anyone has any suggested improvements/clarifications to the third para [14:52] hatch: added the card [14:52] http://jujugui.wordpress.com/2013/07/26/jujucharms-com-update/ [14:53] LGTM gary_poster [14:53] gary_poster: LGTM [14:53] thank you both [14:53] ELGEETEEEM [14:53] :-) [14:53] jujugui call in 7 [14:54] YES! http://canadipsum.com/ [14:54] gary_poster: seems fine [14:54] thanks bac [14:55] gary_poster: wrt the caching problem, I quite like the checksumming approach Tornado uses, I wonder if we could use it: http://www.tornadoweb.org/en/stable/overview.html#static-files-and-aggressive-file-caching [14:55] teknico, we can use it with the tornado server :-) [14:56] exactly :-) [14:56] meanwhile, I'm not sure what we are doing wrong with the Apache static file config :-( [14:59] jujugui call in 1 [15:00] gary_poster, teknico: I believe etags are already enabled in the guiserver, i see "304 Not Modified" responses after the first access [15:03] jcsackett: in test_charm_view.js can we not just load our data/browsercharm.json instead of modifying the existing charmResults? [15:03] rick_h: the charm results was already there, i wanted to keep the tests as close to the previous form as possible. [15:03] also: this whole thing is being blown away once the inspector is live, so didn't seem worth further refactoring. [15:04] jcsackett: yea, it just seems like it's a pre-built lie in the tests now. It works because it's setup to work. [15:04] jcsackett: but cool, understand [15:06] rick_h: it's a worthwhile concern. i'll do a quick replace and if it doesn't open a can of worms do it. or if the can of worms is of the sort that indicates there's something else that needs fixing. [15:07] jcsackett: yea, hobble along. I'll try to be good and look away a little bit. Though as I put in my comments, I don't get how the tests work w/o a render call [15:07] rick_h: render is called in the view's init. which seems wrong, but i'm tampering as little as possible. [15:07] * rick_h runs away [15:07] * jcsackett laughs [15:07] jcsackett: can you note that then vs the render? comment line please [15:07] * jcsackett nods [15:18] orangesquad: could one of you revierw this mp: https://code.launchpad.net/~adeuring/charmworld/charm-revision-files/+merge/177160 ? [15:18] * sinzui looks [15:50] hazmat: phantomjs@1.9.1-0 and mocha-phantomjs@2.0.0 [15:51] hazmat: that's the most recent phantomjs but mocha-phantomjs is currently at 3.1 on npm [15:52] * hazmat downgrades mocha-phantomjs [15:55] hatch, thanks that fixes it [15:56] excellent [15:56] I'll update the docs [15:56] hazmat: you wouldn't happen to have a copy of the error message still? [15:56] hatch, incidentally which docs.. HACKING or docs/hacking.rst [15:57] I was going to do docs/hacking.rst - maybe we should remove HACKING [15:57] hatch, https://github.com/metaskills/mocha-phantomjs/issues/76 " TypeError: 'undefined' is not an object (evaluating 'Mocha.process.stdout')" [15:57] thanks [15:57] jujugui - thoughts on removing HACKING in favour of hacking.rst? [15:58] hatch: docs/hacking.rst is a symlink to HACKING [15:58] it's the same file :-) [15:58] teknico: oh haha - I was just looking at it in sublime [15:58] hazmat: ^^ [15:58] thanks :D [15:59] hatch, fwiw, HACKING is the expected unix-y location, & docs is where we want it for our generated docs [15:59] hatch: yeah, sublime should have a way of showing that [15:59] gary_poster: alright - I'm creating a card as well to upgrade the various mocha dependencies to the most recent [16:00] thanks hatch [16:00] hatch, gary_poster symlink [16:00] ah [16:00] cool [16:01] (yeah, is now) [16:04] gary_poster: https://canonical.leankit.com/Boards/View/102529849/106315624 [16:17] jujugui does anyone have any other additions/modifications they would like to see in the hacking docs before I commit this? [16:17] hatch, i'd suggest moving the improv stuff far down [16:17] hatch, its not really required anymore [16:18] its really only useful for pyjuju testing.. maybe split it out into a separate file [16:21] hatch, hazmat, important for hacking for as long as we have to support pyjuju. keep it in same file till we don't need to any more (13.10?_ [16:21] ) [16:22] hatch, ack on mocha card thx [16:22] gary_poster, is that 13.10 or 2017.04 ;-) [16:22] jk [16:22] hazmat, :-P here's hoping 13.10 :-) [16:23] gary_poster, so the thing is .. its basically describing how to do a dev environment in HACKING.. really for py or go.. for testing/dev we really just do the same process of deploying the charm against a branch [16:24] the whole improv thing was a dev tool which afaik we don't really use anymore in favor of in memory env [16:24] whoops [16:24] hazmat if you mean we only use it for qa, yeah, you are right [16:24] really for py or go.. we test (or should be) against a real env [16:26] ok so I'll just make the single line change about the mocha stuff for now [16:35] gary_poster: what treadmil do you have? Do they make ones without the uprights? [16:35] hazmat, for go we do for obvious reasons. :-) for py, improv is great for branch qa, and it helps find and dupe issues fast. I've never had improv approve something that pyjuju rejected. for release qa, agree on py env. so anyway, I still think it has a place, for as long as we care about py. [16:35] hatch, "sole" something or other [16:35] so how do these work without the tops? heh [16:35] gary_poster, fair enough [16:35] hazmat, tops, you mean the handles? they are beneath my desk [16:35] controls are somewhat awkward to see [16:35] but buttons are on the handles so I can control by sound [16:35] oh so the screen thing goes behind the desk? [16:35] gary_poster, what brand did you go with? [16:35] http://www.soletreadmills.com/ [16:35] hatch, yes. hazmat, treadmill is sole, and desk is trekdesk [16:35] ohhh very cool haven't heard of these guys [16:35] I always thought you needed to take the top off [16:35] (of the treadmill) [16:35] * hazmat wonders the same [16:35] wayyy cheaper than the height adjustable desks [16:35] but plus treadmill :) [16:35] the controls are in the top and handles for the tread.. does the desk fit over that, or is that a something you rigged [16:35] hazmat fits over. I have maybe a foot clearance. it does push against the back display but has not been a problem for me with length. advice I read for treadmill desk treadmill is to look for long deck [16:35] I think I'd like a treadmill desk but what if I want to sit down heh [16:38] maybe put the chair on the treadmill lol [16:38] sinzui: \o/ thanks for sorting out those links [16:38] hatch, yeah. either (a) use a laptop and a chair now and then, (b) get a stool and put it on the treadmill when you need to, or (c) don't sit down. :-) [16:38] very quick single line review https://codereview.appspot.com/11918044/ [16:38] I usually opt for (c) with a side of (a) [16:38] jcastro, np [16:39] hatch LGTM land it [16:39] hatch self-review would have been fine [16:39] landing [16:39] alright :) I QA'd to make sure the command worked to copy/paste/run too [17:25] gary_poster: so a large part of ben's branch is separating the viewlets out into their own separate files [17:25] so maybe I'll try and land that today [17:25] as of right now it's very broken [17:25] so this was probably just a first pass is what I'm guessing [17:41] ok thanks hatch === matsubara is now known as matsubara-lunch [18:24] Makyo: thanks for the branch :) I'll take one of the reviews [18:25] hatch, Cool. I was getting chunk mismatches on it, so I just reproposed. [18:25] I see you only fixed your little(big) monkey patch ....copying the new code caused issues? [18:27] code looks good - reviewing [18:27] hatch, yeah, down that path lay madness. [18:27] lol alright [18:28] Think I might come up with a better layout this weekend since we no longer need to take radius into account. [18:28] s/reviewing/qaing === schwuk is now known as schwuk_away [18:37] Makyo: it no longer pops up to ask if I want mysql to be a slave/master [18:37] when linking wordpress and mysql [18:38] oh...actually jujucharms doesn't ask either [18:38] bug? [18:38] hatch, there are no ambiguous relations between those two services. [18:38] hatch, try mediawiki and mysql. [18:38] ahhh sorry [18:38] I thought it was wordpress [18:39] Or, if you're feeling brave, hadoop and hadoop. [18:39] I don't see anything wrong with the QA [18:39] just gona check the flagged now [18:46] Makyo: lgtm'd with a doc request [18:46] Cool, thanks. [18:47] ohh one question [18:47] bzr ignore - isn't that going to stop the files from being versioned? [18:48] or does bzr ignore work differently than gitignore? [18:49] hatch, It works the same. [18:49] we're ignoring the stuff pulled down from npm [18:49] so won't that stop the new files from being added into the repo? [18:49] ohh [18:49] Which are now named d3.js instead of d3.v2.js [18:54] hatch, would this be an @module? [18:55] It'd be nice to have it show up in generated docs until it goes away. [18:56] jujugui one more review, please, mostly mechanical, but QA would be nice: https://codereview.appspot.com/11938044/ [18:57] Makyo: i'll do it [18:57] bac, thanks. [18:57] sinzui: can you review https://code.launchpad.net/~bac/charmworld/indexer/+merge/177201 [18:58] * sinzui looks [19:08] bac: the refactoring looks good. I am confused though by one addition to the rules. [19:08] which? [19:09] bac line 51: 'bundle_properties['config']' starts a block that appears to describe a charm's options dict [19:09] the default and description attrs are indexed for charms. [19:09] I don't think bundles have that [19:09] They do have services though [19:11] Do i misunderstand what *properties['config'] does for charms and bundles? [19:11] sinzui: no you're correct about that [19:11] i'll delete it [19:11] regarding services, i talked to aaron and he indicated the services would be indexed and did not need to be listed in the exact index section [19:12] sinzui: we also discussed the fact the rules could be tweaked next week since bundle searching will not be used anywhere yet. [19:12] bac: I think we can add services later. It looks valuable, but benji and abentley might say the info is to arbitrary to be trusted [19:13] too arbitrary [19:13] right [19:13] okay, bac, I have no other concerns, so r=me with the removal of that block [19:14] rt [19:14] thx [19:17] bac. I am not certain that tarmac will be happy. staging is dead at the moment because of upgrades to canonistack. We can let it try to do the merge. We will rebuild/deploy staging next week. [19:18] sinzui: i'm confused. if tarmac is dead shouldn't i just merge it manually? or you think it is only hobbled? [19:19] staging is dead [19:19] I don't think tarmac is on canonistack. Am I wrong? [19:20] i have no idea [19:20] is diogo running your tarmac? [19:23] how can you tell if a charm is a subordinate in the left hand side/browesr? [19:24] also why doesn't rsyslog-forwarder show up on a search for rsyslog.. that sounds like missing tokenization on name [19:27] hazmat: yeah I brought up the search issue yesterday [19:31] hazmat: I think jc said that it was a search issue and not a gui issue [19:32] hazmat, we are not doing substring matching. We were getting bogus results [19:33] hazmat, substring matching is address with autocomplete, which is feature flagged on comingsoon [19:36] sinzui: my branch just merged so tarmac must be doing its thing [19:36] sinzui, this isn't substring matching per se.. its tokenization when indexing [19:36] bac: fab [19:37] sinzui, ie rsyslog-forwarder when tokenized results in two words.. rsyslog and forwarder [19:38] ah [19:38] hazmat, subordinates: I don't think there is way, in sidepanel or fullscreen. I'll get it in the overview--after location and before links [19:39] hazmat, expand filters an uncheck reviewed cahrms [19:41] * sinzui is not sure why the approved rsyslog-forwarder is not appearing for the search for rsyslog [19:41] maybe it is text matching content not in the charm name [19:42] sinzui, that sounds like missing tokenization on name [19:44] hazmat, we are explicit about doing exact matching on name [19:46] sinzui, you still want to tokenize that field [19:46] sinzui, this is not about substring/wildcard query [19:47] its about properly splitting a word [19:47] * hazmat peeks at src [19:49] hazmat, try this url http://comingsoon.jujucharms.com:8086/:flags:/ac/ [19:50] so line 193 of search.py .. its doing the schema for the charm.. its not analyzing title or name [19:50] I think it is solving your concern. [19:50] No it is not [19:50] please my urls [19:50] please try my urls [19:53] autocomplete makes it a little better, but doesn't actually address the issue [19:54] hazmat, I think your issue is for users who paste search terms or are using the API to get data [19:54] sinzui, this is a bug in the indexing.. title and name should be analyzed [19:55] title is especially bad.. its never going to match unless they do an exact search on the whole sentence [19:56] file a bug then. We are playing with search next week so we can make changes to suit you. [19:56] its not about suiting me.. its about making it work better. [19:57] hazmat, I agree it will be better. [20:01] oh.. its just title for the bundle. [20:02] bug filed. 1205477 [20:06] http://gawker.com/canadian-man-sorry-for-chugging-eight-beers-and-swimmin-888283225 [20:06] bahahaha [20:07] hazmat, filed 1205479 and scheduled for next week [20:07] jujugui have to run. have a great weekend [20:07] you too! [20:07] You too [20:08] Makyo: sorry, i got distracted. doing qa now. it all looks good so far [20:08] bac, cool, thanks. [20:08] done [20:27] * bac early dog walk. bbiab. [20:52] * Makyo dogwalks. The pups are restless. [21:36] hatch: you still around? [21:37] of course [21:37] got a second to chat? [21:37] sure [21:37] in guichat [21:37] heading there now. [22:02] bye y'all. have a good weekend. [22:04] cya bac