[00:55] hey huwshimi how are you doing? [00:57] huwshimi: are you working on any of the gui stuff yesterday/today? [00:58] hatch: Yep, getting some of this IE stuff done. [00:58] oh ok cool, can you put your head on the cards so that we aren't duplicating work [00:58] hatch: I had a question for you, let me see if I can remember [00:58] sure, I'll be here all night [00:59] hatch: Oh yeah, I haven't been able to reproduce bug #1213260 [00:59] <_mup_> Bug #1213260: DDing a charm renders service icon under sidebar

[01:00] huwshimi: ok, umm does it render exactly where you drop it? [01:01] or is it rendered slightly off? [01:01] my laptop (which has IE10) has a low res screen [01:02] hatch: It's been rendering fine for me this morning, but I just went to double check and my VM died again. [01:05] crap, so what happens when it dies? [01:05] just crashes? [01:06] hatch: Yeah, this time it's complaining about having run out of disk space, but it has plenty [01:06] hatch: Now can't boot into it [01:07] :/ is there a 'free disk space' command? [01:07] I know parallels has that and has to be run from time to time [01:09] http://askubuntu.com/questions/219286/virtualbox-dynamic-disk-not-expanding-to-virtual-size [01:09] possibly related? [01:12] Oh it has actually run out of space, somehow it has taken up 12gb... [01:13] Not sure how to free that up... [01:16] huwshimi: maybe just increase the size of your vm [01:17] hatch: I run it off a disk and the vm has used up all the space on that disk. I'm just temporarily deleting some vm files so that I can boot the vm, clean up and then restore... [01:18] ohh [01:19] hatch: It's a new thing every other day that goes wrong :) [01:19] welcome to windows lol [01:38] huwshimi: any luck? [01:39] hatch: Not yet, it's repairing [01:39] well, it was, no feedback at the moment [01:39] oh boy - sounds like you might need to give it more space [01:39] I think mine has 32gigs [01:39] hatch: It has a few gb free at the moment [01:40] maybe it's paging things because it doesnt' have enough ram [01:42] hatch: It has heaps of ram [01:46] hatch: It's works! [01:46] Now to figure out what I can delete [01:48] hatch: The service block always appears with the top left point under the mouse [01:50] really... [01:51] hmm on a fresh trunk checkout? [01:51] hatch: Yes [01:51] hatch: I haven't seen it do anything else [01:51] pulling new trunk [01:52] gota 'make' then I'll test [01:54] ohh [01:54] you gota drop it on the big square in the middle [01:55] the thing below 'start adding charms' [01:55] actually it doesn't matter where I drop it, it's always off [01:56] hatch: Where does it up it? [01:56] wha? [01:57] hatch: put it? [01:58] ~40px from the left of the canvas under the sidebar about mid screen [01:58] I can demo it [01:58] if you want to hop into a hangout [01:58] I'm in guichat [01:59] http://bazaar.launchpad.net/~juju-gui/juju-gui/trunk/view/head:/app/views/topology/service.js#L426 is the method which handles the drop and the positioning [01:59] hatch: I'm not seeing that at all, on a fresh branch. Windows 8 IE10 [02:00] odd [02:00] checking the code [02:01] hatch: What's your screen res? [02:01] 1366x768 [02:03] hatch: Even changing my browser size doesn't change it [02:04] on this line http://bazaar.launchpad.net/~juju-gui/juju-gui/trunk/view/head:/app/views/topology/service.js#L443 put console.log(ghostAttributes.coordinates[index]); [02:04] and let me know what the two numbers are [02:12] hatch: Trying, but I think doing a disk cleanup and running IE were too much for the vm [02:12] jeesh and I thought I had hardware issues :) [02:14] hatch: 8 cores, 16gb ram is not the problem... 4gb ram and 4 cores are dedicated to the vm. Not sure why it runs so poorly. [02:15] hatch: I'm going to finish getting this VM back to normal and then I'll get back to you about the drag and drop [02:17] sure thing [02:18] huwshimi: just FYI I'm running it in a vm with 3GB of ram and 1 core [02:21] hatch: Virtualbox? [02:21] yup [02:21] V4.2.16 [02:22] I have .10 [02:22] (4.2.10 [02:22] ) [02:22] might be worth an upgrade [02:22] hatch: Did you change settings for the VM? [02:23] what do you mean? [02:23] the hard drive is 25GB [02:23] well what it thinks it's hard drive is [02:25] hatch: There are lots of little checkboxes that do things, but I don't know if I need to change any of them... [02:26] heh yeah there are [02:26] I fiddled with it long enough when I first started with it I remember [02:26] but I can't tell you what I changed and what was there [02:26] I know I had to do some command line trickery to think it had a resolution higher than like 400x600 [02:26] or something crazy [04:02] hatch: 137, 101. And it placed where I expected it to. [04:10] huwshimi: interesting - I get only negative numbers [04:10] in the am I'll have to try and get someone else to see if they can repro [04:11] huwshimi: are you in the 'Desktop' version of IE? or the Metro version? [04:14] hatch: Desktop [04:14] well what the heck hah [04:15] I suppose it's entirely possible mine is caching something wrong [04:17] :/ [04:17] I think you're lie'n to me so you don't have to fix it lol [04:21] heh [04:28] hatch: It's possible it's broken, I just can't reproduce it here [04:28] yeah but how's that possible haha [04:29] both on trunk same version of ie [04:45] hatch: Do you get the same thing in rev 940? [05:10] huwshimi: let me check [05:13] huwshimi: not exactly - that revno was before the fix for dropping the service on the center message [05:13] they are positioned incorrectly still however [05:14] I'm guessing it works correctly there? [05:16] hatch: You changed some positioning stuff the revision after that so I was hoping it might have worked back then... [05:16] yeah doesn't look like it - I'll get someone else to test it out, it could jsut be that mine is broken for whatever reason [05:16] I'm going to take off but if you could push up/email me with whatever you get done so we can merge it in that would be awesome [05:17] hatch: Sure. Night. === rogpeppe1 is now known as rogpeppe [13:09] bac, benji: when you have time, could you please review https://codereview.appspot.com/12927049 ? [13:09] sure [13:09] jcsackett: ping [13:09] frankban: sure. hey sorry i didn't get to your review yesterday before your EOD [13:10] bac: no problem and thanks [13:10] benji: thank you [13:15] jcsackett: fyi, not sure where you are on the bug, but updated it with notes since I thought it might be my fault (though kind of an accidently worked thing) #1214058 [13:15] <_mup_> Bug #1214058: Cannot open Browse if sidebar is minimized

[13:16] hatch: ping when you get in. Want to chat about where we left off last week if you've got the time. === robbiew1 is now known as robbiew [13:43] bac, benji: thanks for your reviews! [13:54] hey I gota run and take the car into the shop so I'll probably be back in 30-45 mins [13:54] rick_h: I normalized the constraints stuff so you're free to go on the ghost constraints [13:54] hatch: k, will unblock the card then [13:54] hatch: thanks and have fun at the dealer [13:55] lol dealer [13:55] ppl take their cars to the dealer? [13:55] :P [13:55] hatch: yea, under warranty and all :P [13:55] rick_h: saw your notes, had come to the same conclusion yesterday, but thanks for the validation. :-) [13:55] riiiiight, my cars are too old haha [13:55] (not so much about what had happened, but about what was going on) [13:55] ok gota run be back in a few [13:56] jcsackett: cool, wasn't sure where you were with it but I was curious as it seemed really strange any changes recently broke that :/ [13:56] rick_h: yeah, and the truth is it never really should have worked. :-P [13:56] jcsackett: right, that made more sense :) [13:57] in a ...strange...kind of way [13:57] rick_h: i'm glad you updated me though. i could see what was wrong but my update at standup was def going to be "i have no idea why this ever worked". [13:58] jcsackett: sorry my z-index stuff borked a couple of other things on you [13:58] jcsackett: it's the one line 'fixes' that kill [13:58] * jcsackett laughs [13:58] you mean that floating sidebar icon thing? [13:59] at least it was an easy fix too. [13:59] yea, as a drive IE fix I had to bump the z-index and then saw the bugs come out of it [13:59] ...of course another one-line, so we'll see what that borks. :-P [13:59] but since it was a one-liner and I TRIVIAL'd it...WCPGR [14:00] ...i have no idea what that last acronym means. :-P [14:01] jcsackett: that's SteveK's famous "What Could Possibly Go Wrong" /me misses that [14:01] aaaah. [14:01] yeah, we need to bring that back. [14:23] back [14:24] rick_h: on the weekend I installed cruise and a stereo into the mrs car and now it's off to get a new windshield - I typically do all of my own repairs but it's cheaper to have someone else install the windshield :) [14:25] hatch: defintely. special tool to pull the gasket around there and such [14:25] hatch: was blown away by how fast a pro could do it [14:26] yep! [14:26] so who has IE running? [14:26] huw said he couldn't repro this bug https://bugs.launchpad.net/juju-gui/+bug/1213260 and was looking to see if it's fixed for anyone else too [14:26] <_mup_> Bug #1213260: DDing a charm renders service icon under sidebar

[14:26] hatch: looking [14:28] hatch: yea, not doing it here [14:28] hatch: even if I just drop it 'right' past the edge of the sidebar, the left side of the service block is a few px east of the sidebar edge [14:30] well I'm glad....but slightly irritated because that means there is something wrong with my IE :/ [14:31] hatch: updating the bug with my screenshot of a tiny IE window working right. Re-open if you can dupe or find something. [14:31] rick_h: mind closing the ticket as cannot reproduce? Just so we have a record of it [14:31] oh I can dupe the issue for sure [14:31] but noone else can lol [14:31] hatch: well, marked it invalid with the notes about 'cannot reproduce' [14:32] oh right...that's how it works [14:32] did anyone get an email from huw about what he was working on? [14:33] * rick_h didn't see anything [14:33] allllright then [14:33] jcsackett, 1x1? [14:34] rick_h: so the constraints util method should all be gtg now, and all you need to do is create another wrapper around the constraints partial to loop through the fields [14:34] hatch: looking [14:34] sinzui: yes. [14:35] sinzui: i'm in the hangout attached to the appt [14:39] rick_h: remember the performance discussion we had about etags and mobile? You should watch these talks it outlines the issues in detail in the first video https://plus.google.com/118445028821328031751/posts/hz2XpU76xzN [14:40] hatch: will put it on for lunch time viewing. thanks [14:40] it has some really great information in it - there is like 3.5h of talks in the playlist hah === BradCrittenden is now known as bac [15:09] Makyo: as far as core is concerned what happens if the constraints aren't specified? [15:09] hatch, Aren't specified? As in, you send a set-constraints with no constraints, or as in you deploy without specifying? [15:10] both I suppose - assuming the constraints object is empty [15:10] right now there are no errors it appears? [15:10] doe stha tmeans it picks the lowest possible? [15:10] wow I can't type [15:11] hatch, Don't know off the top of my head, I'd assume it just allocates whatever size is specified in your ~/.juju/environments.yaml, which is a machine class, like tiny, small, etc. [15:11] alright so with none specified in the gui a good name would be 'Default' ? [15:12] I'd make sure with luca. That makes sense to me, but others might not know where that default is specified - I've been using Juju for a year, so I've got that advantage. [15:13] good point [15:13] luca: are you there? [15:13] *poke poke* [15:13] Makyo: hatch heya [15:13] when 'scaling up' the dialogue shows you the current constraints [15:13] if none are specified....what shoudl be shown? [15:13] hatch: the defaults [15:14] we can't get the defaults [15:14] all the info we have is that there isn't anything specified [15:14] luca, if you deploy without constraints, it uses what's specified in your environments.yaml file, which we don't have access to. [15:14] right now I have Default Ghz Default GB ..... etc [15:15] Makyo: hatch right. What do you think we should show? [15:15] I like 'Default' [15:15] although then people won't know what those are [15:15] but it's kind of a chicken/egg issue there [15:16] I suppose I'd expect a 'use defaults from environment.yaml' checkbox that disables fields, maybe? But yeah, I have experience with juju [15:17] I'd say that could be in the 'set constraints' section [15:17] but in the 'these are your current settings' section [15:17] maybe we just have Default for now, and then when we add 'help bubbles' they can say where the defaults are set [15:19] Yeah, prowling through the machine and unit objects in the db from improv, we don't have any specs. [15:19] Don't know how meaningful improv is, not checking against core yet. [15:39] jujugui lf two quick reviews and an IE QA on https://codereview.appspot.com/12987045/ plz [15:39] hatch, on it [15:39] Got the IE bit. [15:39] hatch: looking [15:39] right arm! [15:40] I needed to get these landed so that I can finish the upgrade ux heh [15:40] darn IE [15:40] hatch: what do I need to do to see the original bug? [15:41] scale up the units without editing the constraints so it'll say 'Default CPU Default GB ... [15:41] if it doesn't wrap then you're good [15:41] and as far as the X's just focus an input, if there is no X in the post deployment inspector then you are good [15:42] hatch: ok, the bug states on 'inputs' and so I was checking on the config inputs [15:42] yeah I left the X on ones which aren't databound [15:42] I didn't realize IE did the X clear on non type="search" inputs [15:42] just incase IE ppl like that X for some reason... [15:43] and yea, the config ones are actually textareas so bad check [15:43] shift + cmd + left is so much faster than trying to find the damn X with the mouse :) [15:48] rick_h: did you want me to comment in the css why the part I removed is no longer required? [15:48] I'm a little confused by the comment [15:51] jujugui call in 9, kanban now [15:51] hatch, want to run it today, then I'll get W/Th, you get F? [15:52] sounds like a plan [15:52] * hatch gets out the whip [15:52] * hatch just pulling alittle stewie there [15:55] benji: after the call a review of https://code.launchpad.net/~bac/charmworld/json-for-deployer/+merge/181087 ? [15:56] bac: sure [15:58] Makyo: hatch got called away, did you get a solution hehe? [15:59] luca, Don't have anything atm, was hatch's task. [15:59] jujugui call in 1 [15:59] luca, will poke around more after the call. [16:00] Makyo: cool [16:00] luca: I just ended up going with 'Default' instead of undefined when it's not specified [16:00] hatch: nice, ok [16:15] luca: if you decide you want it changed in the future it's pretty easy to change [16:19] bac: your branch looks good. I had a question though: do we need URLs without user names for promulgated bundles? [16:19] benji: yes, the jujugui charm is an example [16:20] benji: last week we discussed that and decided that the deployer will always want a versioned basket [16:20] err, ignore me [16:20] benji: that decision is open to be revisited but that's what i went o [16:20] n [16:21] benji: btw, this branch was very easy to do based on the nice stuff you added recently [16:22] bac: right, I think we always want a version, but don't we want to be able to say "I want the official big mysql cluster bundle, regargless of who is its current manager"? [16:22] I'm glad to hear I made the code a bit better. [16:25] jujugui does anyone know the status/details of the card in Inspector 'Inspector does not work with core: cannot iterate over undefined WRT units' ? [16:26] no, sorry, sounds like that one needs testing with a real core deployment [16:27] yeah there is no ticket attached :/ [16:27] I created it in a rush, sorry. [16:27] Will try again. [16:27] Just wasn't listing units. [16:27] benji: yeah, perhaps. sinzui do you have an argument against what benji says? [16:27] Makyo: ohh, so that really needs to be fixed then before we unflag? [16:27] If it's still an issue. [16:27] benji: it'll be trivial to add... [16:28] yeah; if we aren't sure one way or the other we can land it as-is and add this later [16:28] Makyo: if you have time today/tomorrow it would be awesome if you could test it out again :) [16:29] hatch, doing it now; need a juju environment up anyway. [16:29] oh awesome thanks! [16:29] I got my first SMS spam this morning at 5am - I have finally made it! [16:43] hatch, looks good, trashing the card. [16:44] *phew* thanks for looking into that Makyo [16:47] bac: We do want to support short url. the GUI will always use the full url. But from the command line, I might want to type just:fast-wordpress [16:48] bac, remember search will only return tip and reviewed bundles have a higher score,. The GUI will always suggest the user use the latest official charm. [16:51] rick_h: I'm just going through the tickets and came across this one https://bugs.launchpad.net/juju-gui/+bug/1209016 I know you fixed it but it appears that the + sign doesnt turn orange on hover...is this a css or image issue? [16:51] <_mup_> Bug #1209016: Right hand zoom slider handles sprited improperly [16:51] bac: benji: do either of you have time to review https://code.launchpad.net/~sinzui/charmworld/api3-search/+merge/181092 ? I have an implementation question and am open to discuss net steps. [16:52] sinzui: could you have a look a this MP: https://code.launchpad.net/~adeuring/charmworld/1206659-simpler-es-mapping/+merge/181100 ? [16:53] * sinzui looks [16:53] hatch: looking [16:55] hatch: looks like a JS issue. The hover class is added, but the sprite css doesn't change the icon. If you hard edit the css the image will show [16:55] hatch: so rather than adding a css class it needs to add and remove [16:56] right looks like the old class is still taking precident [16:56] I'll make a card for this ticket [16:56] hatch: comment added with the details https://bugs.launchpad.net/juju-gui/+bug/1209016/comments/1 [16:56] <_mup_> Bug #1209016: Right hand zoom slider handles sprited improperly [16:57] cool thanks [16:57] jcsackett: did you need a second pair of eyes on the event stuff? [16:59] jujugui I'm having a hard time concentrating through the meds. Going afk. [16:59] Alright, good luck [17:06] adeuring, r=me with comments [17:06] sinzui: thanks! [17:11] sinzui: Making the two lists "constants" is fine, but I don't understand your suggestion to add an XXX: Do you mean line 230 of the diff? I don't think that an XXX is needed there [17:12] adeuring, okay. I wont press you to make the change. [17:18] jujugui. Note that https://jenkins.qa.ubuntu.com/ is not responding. So charmworld is not collecting charm test results at the moment. [17:18] hmm [17:19] sinzui, raising to webops [17:59] hazmat, sinzui: i got booted off-line. m_3 and i raised the jenkins issue on #is yesterday [17:59] bac: they're still working it too it looks like [18:00] thank you bac [18:00] m_3, even more dead than yesterday [18:02] benji: thanks for the review. i'm going to land it now and do the promulgated as another branch [18:02] cool [18:33] sinzui, benji: do we think this is the URL to get JSON for the deployer use for promulgated branches: http://manage.jujucharms.com/bundles/mysql/tiny/json [18:34] bac: I /think/ so. I don't remember exactly. [18:36] benji: i've added it to the doc and will proceed. having an optional version right in the middle is not pretty [18:36] mmm [18:36] or maybe the version is only there if there is an owner [18:39] bac, I think so. There is no version and no owner in it [18:40] bac: I shared "Deploying Charmworld and Juju-GUI to Prodstack" with jujugui and orangesquad so that everyone knows the fastest way to get things deployed [18:40] sinzui: yay [18:42] sinzui: i have given it a gold star so i'll always have it with me. [18:52] rick_h: back yet? [19:45] hatch: what's up? [19:59] rick_h: oops sorry didn't see you replied [19:59] I'm going to be changing the template that the constraints use a little [19:59] hatch: rgr [19:59] feelin better? [20:00] hatch: little bit, took a half day sick and a nap [20:00] now nap-groggy, but wheeeee [20:21] hey hatch, since updating your jshint i'm getting complaints about the 2 line indents on all files. any idea what i need to change? [20:22] jcsackett: you merged trunk? [20:22] that's how i got your jshint update. [20:22] should there no longer be 2 space indents? [20:22] * jcsackett goes to see if there's more trunk to merge [20:23] hmm well no there should be [20:23] guichat real quick? [20:24] hatch: sure, one sec. [20:25] bac, do you have time to give me some feedback on https://code.launchpad.net/~sinzui/charmworld/api3-search/+merge/181092 [20:25] sinzui: sure [20:28] jujugui (hatch, bcsaller?) - the saveAs noop is what's causing the memory leak for me in the websocket_logging tests. Am I behind on a version or something? Have clean-all' [20:28] Have run clean-all etc. [20:29] afaik that code hasn't been touched in quite some time, this is the export saveAs you're referencing? [20:31] It's just the line "saveAs = function() {};", I think. In the browser, as hatch says, I get a save as dialog, but Phantom just dies, then I get a message that it's used all available system memory. [20:31] oh THAT's why phantom always dies for me now [20:31] hatch: threw node_modules/jshint/bin into my path and now all is well. [20:31] jcsackett: awesome :) [20:32] Makyo: looking at the test [20:32] hatch, it [20:32] It's test/test_websocket_logging.js:59 [20:32] That test. [20:35] oh em gee who wrote FileSaver.js [20:35] lol [20:35] FileSaver.js is a polyfill we pulled in [20:35] sinzui: are you asking for a full review? or a mid-imp discussion? [20:36] bcsaller: yeah I know, :) [20:36] Makyo: what if you change those to websocketLogging.prototype.saveLog = ... ? [20:37] bac. I hope the former, but the later is not unreasonable since I think the metadata issue should be solved in this branch [20:38] jujugui: can i get two reviews for https://codereview.appspot.com/13092044 [20:38] on it [20:40] thanks, hatch. [20:42] No luck, hatch [20:43] darn - ok I'll have to look into the memory profile in a few [20:43] Will poke around too [20:43] jcsackett: reviewing [20:44] jcsackett: sanity check on the test method there please. [20:44] jcsackett: but other than that ok. Heading out. [20:45] grrr, reviewboard is timing out [20:46] Makyo: hmm I can't reproduce the memory leak in the browser [20:47] does it happen in the browser for you too or just in phantom? [20:47] hatch, just phantom, the browser shows the save dialog. [20:47] ok I don't get any save dialogue [20:47] and I shouldn't because the saveLog method was monkeypatched [20:48] sinzui: is the metadata currently used? i think your suggestion is a good one. [20:48] it.only('responds to the saveWebsocketLog event', function(done) { [20:48] hatch, the memory bug seems to be a phantom thing, closing uncleanly because of a crash with the file saving. [20:48] that's the proper test right? [20:48] That appears to be it. [20:48] ok trying phantom [20:48] rick_h: ack. [20:48] no issues [20:48] Makyo: maybe I need your branch [20:49] That's why I was wondering about versions. [20:49] ok well trunk as far as this morning was A O K [20:49] if you push up your branch I can take a look [20:50] bac: it is not used for bundles. Aaron and rick introduced it to charms to describe things we add to the charm, like related_charms. I think adding to metadata is keeping with the grand plan, but you or benji might know of changes to the plan while I was in IoM [20:50] hatch, I haven't touched that test, nor anything that test touches. I'm currently checking trunk, so give me a sec. if it is my branch, I'll push it and send it your way. [20:51] alrighty [20:51] sinzui: i don't know anything about it. it looks good to me. [20:51] maybe it's just in your head ;) [20:51] * hatch waves his hand "this is not the test you are looking for" [20:51] Nnnnnnno. [20:52] hatch, dies in trunk, too. [20:52] hmm [20:52] that's very odd [20:53] phantomjs --version ? [20:53] I'm showing 1.9.1 [20:53] Okay. then I will add doctype to meta data to charm and bundle in API3. bac I intended to look at the index_client.api_search() I discovered. Do you agree? Or should I work on the interesting endpoint? [20:54] Makyo: the thing that's intersting is that you said you get a save dialog - I do not... [20:54] 1.8.0 - let me try updating. [20:55] yeah that version is pretty old [20:55] Dec 2012 [20:55] * hatch crosses fingers [20:55] sinzui: api_search [20:55] That's what I was thinking, s'why I asked. Hope it works! [20:55] bac, thank you for the direction. [20:56] * sinzui cleans up metadata rules [20:57] hatch, It crashes faster this time :) [20:57] lol [20:58] and it works perfectly fine here... [20:58] 11MB of ram usage [20:58] Fffff. [20:58] It's no longer memory-leaking, thankfully. [20:59] Just crashing. [20:59] ohh ok [20:59] phantomjs always crashes on me for no reason [20:59]