=== uru-afk is now known as urulama [08:52] morning, frankban [08:52] hi urulama [08:53] regarding my comment on support for start/end in /meta/stats, i guess i mixed it with /stats/counter [08:53] https://docs.google.com/a/canonical.com/document/d/1TgRA7jW_mmXoKH3JiwBbtPvQu7WiM6XMrz1wSrhTMXw/edit#bookmark=id.9k72szsuel9z [08:54] which is probably the endpoint for the results for web (charm broswer), like downlaods in the last 7 days, 30 days, half year, all time ... [09:02] urulama: yeah, so the stats counter already works for some of that [09:06] frankban: yes, great side-effect :) [11:48] urulama: do we really want to allow deletion of partial ids? [11:48] frankban: deletion? [11:48] urulama: yes, DELETE archive [11:49] This deletes the given charm or bundle with the given id. If the id does not mention a specific series or revision, all the series and revisions of the given id are deleted. [11:49] frankban: good point. [11:50] frankban: it does make sense from the API point of view. But maybe it is best to delete only concrete api and request a user to do /expand-id and iterate on all [11:51] s/concrete api/concrete charm [11:52] frankban: maybe we could do two phase process. delete on charm id just removes a charm from being available, but not from the store. and one needs to "confirm" later to remove all "deleted" charms from the store permanently? [11:52] urulama: +1 [11:54] urulama: uhm, I'd be not inclined to add state to the server. [11:55] urulama: +1 on having to specify a fully qualified url, but I'd be -1 on the two steps process: it complicates the server and also the client [11:55] frankban: ok. then let's just do fully qualified url part (gonna change the docs) [11:55] urulama: thank you [12:15] frankban: working on DEL part of archive? [12:17] urulama: yes [12:17] urulama: creating a card [12:23] frankban: tnx [12:31] * urulama cooks lunch [12:58] jujugui: the internet man is on his way. i may be going on and off line a bit in the next hour as he attempts to fix the problem. [13:08] bac: good luck [13:09] bac: i found out yesterday, that 100/10 line here costs only 5eur more then my current 10/10 ... will go upgrade soon [13:09] (per month) [13:42] frankban: updated https://code.launchpad.net/~evarlast/juju-quickstart/which-juju/+merge/227238 and https://codereview.appspot.com/132770043 [13:42] jrwren_: cool [13:43] jrwren_: why there is no Delta from patch set in reitveld? [13:43] jrwren_: how do you update the mp on reitveld? [13:43] * jrwren_ looks. [13:43] frankban: lbox propse [13:43] jrwren_: uhm, that's what we are supposed to do, weird... [13:44] jrwren_: ok, I'll take a look ASAP [13:47] morning all [13:47] morning hatch and all [13:47] jrwren_: so I watched some of those starcraft games this weekend, those guys are a little nuts [13:47] haha [13:48] I'd probably die in their scouting runs :P [13:48] hatch: you watched live stream? did you see me in the crowd? :) [13:48] I did, but no I didn't see you, the crowd was really dark and the feed was actually quite poor quality [13:49] hatch: Awesome. I hope you enjoyed it. It was very fun to be there live. [13:49] yeah I bet, I couldn't stop laughing at the start "aww they think they are real sports people" [13:49] lol [13:50] hatch: They have higher revenue than the NHL. [13:50] well, maybe the league but not the players :) [13:50] hatch: truth! [13:50] watching some talk was like bringing be back to highschool haha [13:51] hatch: how so? [13:51] oh, well they were like 17 [13:51] so they talked like they were 17 haha [13:51] no doubt about the skills though [13:51] hatch: ha! Yes. A lot of the players are very young. [13:52] It seems to be a very balanced game, even if one screwed up earlier in the game that wasn't a sure thing that they were going to lose [13:52] it was possible to come back sometimes [13:52] and the best-of-three was a good idea [13:54] yup. [13:55] they need to have an old-person league [13:55] haha [13:56] or maybe a minimum salary so that old people with families can compete :D [13:56] kids don't have any money, if they want to get real money intot he sport they need adult competitors :) [13:59] frankban: i'll take a look at the PR right after the call [13:59] urulama: cool, I am updating it with another test [14:03] urulama, jrwren_ : https://github.com/juju/charmstore/pull/80 is now ready for review [14:03] hatch: I am pretty sure they know what they are doing, but I do wish for an old person league so I can play ;) [14:11] jcsackett: you in yet? Can you kick ci and then kill the node server? [14:11] hatch: as in restart the jenkins service? [14:11] sure. [14:12] hatch: though i know nothing of a node server. [14:14] jcsackett: when you kill the running tests in CI you then need to ssh into the machine and do a ps for the node server [14:14] kill it [14:14] then the tests will run again [14:14] if you don't it'll just give an ERRADDRINUSE error [14:15] hatch: it would appear restarting the jenkins process handles that; as there's no node in ps output. [14:15] oh interesting [14:16] hatch: i'm starting a build manually to make sure things can pass (or fail for the right reasons) [14:16] I triggered another build of huws branch [14:16] oh hahah [14:16] nm, i see your build. :0 [14:16] * hatch is speedy this am [14:17] it must have been kiting yesterday - 9C 35KM winds and pooooooouuuring rain [14:17] set up was bruuuuuuutal [14:18] ...that's interesting, your build died almost instantly, hatch. i've started another--i think it might have been a bad sha. [14:19] ohh maybe [14:19] hrm. [14:19] and this one died another way. fun. [14:19] jcsackett: looks like npm issues [14:20] to the twitters!!!! [14:20] yup. [14:21] hmm no reports of it being down [14:27] i suspect it's a problem with CI. [14:27] largely b/c i don't like jenkins. [14:29] or, more correctly, largely b/c jenkins doesn't like me. :p [14:29] lol is it REALLY jenkins fault? Or is it that Jenkins is so magical you have no idea what broke? :) [14:29] jenkins isn't really magical; it's just awkward. [14:29] but it's so awkward i have no idea what's broken. [14:31] haha true [14:31] jrwren_: highschool doesn't start at 8:30 there? [14:32] morning party people [14:32] hatch: quite often 7:30ish [14:32] mornin mr vacation man [14:33] jrwren_: that's just nuts! elementary is 9 and hs is 8:30 here [14:33] rick_h_: :) how's it going? [14:33] hatch: Canada is better than the US in almost every aspect. [14:33] haha, well education is handled at the provincial level......buuuuut I'll take it!!! [14:34] hatch: howdy, want to do me a favor? I didn't get the luca bugs turned to cards before I left [14:34] hatch: can you create those, with gui class of service, etc today please? [14:34] rick_h_: sure [14:34] hatch: ty much [14:34] now go vacate [14:34] vacation? [14:34] vacationation? [14:34] not sure what the act of vacationing is [14:34] hatch: sore, very sore. [14:34] rough plane ride? [14:34] the whole walk bus carry a 4yr thing is getting long [14:34] naw, just putzing around SF [14:34] google and the bus drivers seem to like to disagree on how to get somewhere [14:34] oh haha [14:35] yeah 4y/o are getting big [14:35] :) [14:35] yea, he gets tired by the EOD makes carrying a 40# lump a bit exhausting to my own day lol [14:36] but having fun, beautiful weather. Did the bay boat tour thing yesterday [14:36] oh nice - get some good pics? [14:36] we'll see, took a ton. [14:37] awesome [14:37] where to today? [14:37] jrwren_: review done [14:38] frankban: thanks. [14:38] hatch: Cal acadamey of science, japanese tea gardens, and hoping to hit up a fishing pier down by the golden gate for some sunset pics [14:39] more walkin! [14:39] maybe you can get the kit some rollerskates :) [14:39] kid* [14:39] lol [14:39] hopefully he grows some strong leg muscles :P [14:40] haha that's also an option [14:40] although probably not nearly as entertaining [14:40] friends of ours are teaching their daughter how to rollerblade [14:41] holy is that funny :) of course she can't see us laugh :D [14:41] there you go, he sees kids with a scooter and wants one [14:41] I need to see if a store around here has one of those [14:41] oh yeah, those are STILL big [14:41] then again I might end up carrying a scooter and a boy back to the apt [14:41] who would have thought... [14:41] hahaha [14:46] hatch: so, i ran manually via ssh, and got the npm erradinuse thing this time, but ps show's no node--is it a different process? [14:47] ^ rick_h_ CI is borked again.....wasn't it the node process that was still running? [14:48] I need one more OK: https://codereview.appspot.com/132770043/ [14:48] hatch: it'll either be a python or .sh process that starts the node one [14:48] jcsackett: ^ [14:49] dig. [14:49] thanks. [14:49] jenkins 50559 0.0 0.1 41348 8164 ? S Aug22 0:00 python ../bin/http_server.py 8888 [14:49] jcsackett: hatch often it's more clean to | grep jenkins [14:51] rick_h_: yeah, found it, killed it. thanks. [14:53] cool thanks [14:53] whoo! looks like tests might be running. i'll try restarting a build in a moment. [14:56] jujugui call in 5 kanban it up! [14:56] jcsackett: Makyo I see you guys have both interacted with this bug - I think it's resolved? https://bugs.launchpad.net/juju-gui/+bug/1340666 [14:56] Bug #1340666: Clicking the destroy icon does not destroy the machine or container [14:57] hatch: no idea if it's fix released, but Makyo was definitely working on it weeks ago, and it sure seems like the trash icon deletes things on develop. [14:58] hatch: pretty sure its safe to mark as fix committed [14:58] thanks [14:59] jujugui call now [15:01] hey huwshimi [15:08] i forgot to point out my new glasses! you are all crisper today. [15:08] bac: nice. :) [15:09] bac: are they also warby parker? [15:09] hey hatch how was my audio? internet fixer came today [15:09] jujugui sorry I forgot to mention I created cards for all of lucas bugs last week so there are a lot of cards in project 1 - some can be combined, lemme know if you have any questions about them [15:09] jcsackett: no, another boutique i found: costco [15:09] bac: it was actually pretty good [15:09] jcsackett: they refused to use my old frames. :( [15:09] urulama: do we have a card for logging in the charmstore? [15:10] jcsackett: and i had the "dr nick" of opticians do the exam. pony tail and everything. [15:10] frankban: no [15:11] urulama: ok do you want me to create one? [15:11] frankban: we discussed it a bit with Roger, as we want to log the events in such a way, that they can be easily processed, even map-reduce style [15:11] hatch: they replaced the ethernet connector that plugs into the antenna. it is a) PoE and 2) exposed to the elements. they should just proactively replace them every six months. [15:11] frankban: so, there's no card for that, yet [15:11] bac: heh yeah maybe you could grab a few and do it yourself :) or pre-schedule the replacements [15:11] frankban: on what level did you mean to do it? [15:12] urulama: I am not tackling that now, but my branch includes a TODO (add logs), so I must ensure a card at least exists for logging [15:12] so, much as "bad wolf" has become the default error message for tests, "yelling goats" should be our new random string. [15:12] frankban: aaa, the new rules ;D [15:13] frankban: i'll put one on deck [15:13] urulama: thanks! [15:13] hatch: for the price of a set of crimpers that isn't a bad idea [15:15] bac: I also wonder if they have 'weather proof' ones [15:15] like maybe putting some of that heat-shrink-wrap stuff over it [15:37] lazyPower: I upgraded the Ghost version in my charm this weekend and started looking into testing - can you remind me of what tests I'd need to get it promulgated in trusty? [15:39] jujugui: hey all, sorry I missed standup today. [15:39] jrwren_: second review on your quickstart branch. you should be good to go now. [15:39] it's ok, you now owe us all coffee and doughnuts [15:39] :P [15:40] Count on it. [15:40] kadams54: :P Did you have anything to note about your branch? [15:40] I'll steal them from the hotel's breakfast at the next sprint. [15:40] ;-) [15:40] lol [15:40] bac: Thank you. [15:41] hatch: No, not really. I spent most of my day on Friday working to get the last piece of the unplaced units work landed. [15:41] ok cool [15:41] hatch: Which it did. And that makes me very happy. But now it's time to forge on and get this MV stuff wrapped up :-) [15:43] kadams54: if you look at the kanban I created cards for all of luca's bugs last week, so there is a ton to choose from, some can be combined as well so if you are looking for something take a peek at the red ones :) [15:44] Excellent :-) [15:50] frankban: done the PR review, had some comments, but when resolved, LGTM! Tnx. [15:50] urulama: thanks [16:06] oo mongo is getting document level locking [16:06] that'll be nice [16:07] first step to transactions :) [16:08] urulama: I am thinking about your suggestion to delete the blob first and the entity after [16:09] urulama: I made it like that because from the user perspective it's better to have a 404 (no entity) rather than an entity found and then a weird error because the blob is not there [16:11] frankban: ah ... nevermind the comment, the code is correct. i somehow mixed referece=blob name, enitty=blob in my mind :) [16:12] frankban: removing the comment [16:19] urulama: thanks [16:32] jujugui: EOD for me. have a good day [16:32] good night urulama [16:33] night urulama [16:33] btw, anyone watched Guardians of the Galaxy or Lucy? can't decide which one to go to watch today (both look good for some brain-shut-down time) :D [16:34] urulama: Guardians, for sure. [16:35] kadams54: tnx, Guardians it is :) [16:39] urulama: heard great things about gotg, let me know :) [16:39] hatch: sure thing [16:39] A fresh advice from my friend: don't carry the kid on your shoulders when he's having an ice cream :D :D [16:41] haha [16:51] What shall I do when make test runs fine for me, but lbox submit runs make test and it fails? It seems we have different source files. [16:51] jrwren_: be sure you update your local source with the remote source [16:52] that used to happen to me when the remote source and my source had diverged so the tests were no longer valid [16:52] (not sure if that's the case here) [16:53] i'll bzr pull upstream then. [16:57] thanks hatch [16:57] worked? [16:57] yes [17:00] jrwren_: that's good, because my suggestion was gonna be to throw something out the window. [17:00] jrwren_: great :) glad to help === jrwren_ is now known as jrwren [17:08] mbruzek: are you using the Ghost charm for your blog? [17:11] hatch: he sure is :) [17:11] awesome - ok now that people are actually using it, I'll put some more effort into it haha [17:11] lazyPower: did you see my previous q about the tests? [17:11] hatch: and re: your question this morning - bare minimum of unit tests, but we'd like to see amulet tests (you can choose one or the other depending on yoru skill level.) [17:12] there isn't really a ton of code, I think i'll just go for amulet tests [17:12] hatch: i'll ping you later this week about it, but i'm going to be moving my ghost blog from an ansible deployment to the juju charm, and document the process of how easy it was to get moving with juju to migrate a ghost blog. [17:12] tbh I'm not entirely sure how to unit test shell-outs heh [17:12] so expect some more heat around your charm. [17:12] you cant. [17:13] great (about using the charm) [17:13] system calls can be stubbed if you're reading response data from the call [17:13] boo (about unit testing shell-outs) [17:13] otherwise, you're just assuming it happened and gave a zero return code. that's about it. [17:13] yeah - the code wasn't written with testing in mind so I may have to do some refactoring [17:13] but i'll start with the amulet tests [17:13] integration tests are the most important [17:13] proving that it actually works [17:13] that's an excellent place to get moving. I'm looking forward to that merge landing in the queue [17:14] if you want any help/pointers on amulet, feel free to reach out. mbruzek is also an amulet master [17:14] the version in my gh repo is running ghost 5.0 btw [17:14] er [17:14] 0.5.0 [17:14] i'm a -stable kinda guy. I use whats in the charm store unless i *really* need something else, and then i deploy from personal namespace. I'm getting away from doing local charm deployment as much as possible [17:14] the overhead of remembering to upgrade charms locally is something I can't be asked to do [17:15] haha np - I'll try and get on those tests quickly [17:15] probably wont be until wednesday though [17:15] no rush :) [17:15] I may wait until you land your amulet tests before id o the migration so i can highlight the fact its a tested charm [17:15] and make sure you ping whomever does your review to update the quality info on ghost, that'll give you another star on quality [17:15] we *should* be doing that anyway, but i typically forget :( [17:16] sure that's fine [17:48] Hey hatch just got back from lunch catching up on the scroll back. I AM using ghost and I CAN help you with amulet. [17:59] mbruzek: I was asking specifically about the ghost charm :) because I wrote it and wasn't going to put a ton of time into it until others were using it :) [17:59] hatch, I am using the ghost charm. [18:00] excellent - well please file any bugs or feature requests on the GH repo :) [18:00] https://github.com/hatched/ghost-charm [18:00] GH ? Not Bazzar ? [18:01] the gh repo is the source of truth, lp is just a compile-to target :) [18:01] mbruzek: so how do you have your env set up? Are you using mysql/apache? [18:01] roger that. [18:01] Actually no. [18:02] don't get mad [18:02] But I am using haproxy [18:02] I realize this is not a way to do it in production. [18:03] But I am using the haproxy charm and ghost charm to run my blog [18:03] that's fine :) [18:03] haproxy works pretty well [18:03] you should look into using mysql though [18:04] why? sqlite not going to keep up? [18:04] if you're just using the built in sqlite db, you'll have to manually transfer everything if you ever upgrade the charm [18:04] harder to do backups...etc [18:05] hatch, I had a question about the configuration. When I tried to set the "host" to the actual IP address my blog goes dark. [18:06] mbruzek: yeah typically you want to just leave it at 0.0.0.0 and let the image handle it [18:06] tbh I can't think of a time when you'd want to specify the ip [18:06] hatch OK that is what it is set to now [18:06] but maybe if you were running some fancy network set up and you wanted it to use a specific ip [18:07] are you running haproxy and ghost on the same machine? [18:07] hatch yes. [18:07] juju deploy cs:precise/haproxy --to 0 [18:08] juju deploy cs:precise/ghost --to 0 [18:08] juju add-relation ghost haproxy [18:08] It was as simple as those three commands. [18:08] hatch, I will look at adding mysql. [18:09] mbruzek: awesome - yeah just curious really :) [18:09] mbruzek: so if you switch to mysql you'll have to manually port the posts over [18:10] (sorry) that's not done in the charm yet [18:10] Oh darn, all 2 of my blog postings [18:11] haha - mbruzek I am very interested in that procedure though - I'm sure others will need to do the same [18:11] keeping the blog post id's the same (if it doesn't automatically) [18:11] etc [18:11] I'd even like to include it in the readme if it's small enough, or an a supplementary file in the charm even [18:11] OR [18:11] if it's really easy, PR's accepted :P [18:12] (the charm is in javascript) :D [18:12] * mbruzek shivers [18:14] hah yeah it was a challenge in some places but it's written in JS because Ghost is a js project [18:14] understood. [18:53] evening fabrice [18:53] evening ! [18:55] jujugui need some reviews/QA (especially QA) on clearing ECS changes https://github.com/juju/juju-gui/pull/508 [19:00] Makyo: I'll take one [19:00] got to grab some lunch first [19:00] Sure [19:11] Makyo: working on my own PR but as soon as I wrap that up I'll dive into the QA queue. [19:26] guihelp: my own PR for QA: https://github.com/juju/juju-gui/pull/509 [20:05] jcsackett: I found an easier way to accomplish what I needed https://github.com/hatched/juju-gui/commit/9da0a126618b4a86f94e1f9a806bcc25102ae24f does this conflict with what you're working on? [20:07] if it does....it's a race to see who lands first :P [20:37] hatch: not even a little bit. [20:37] jcsackett: great - my original version was much bigger [20:37] heh [20:38] I've also found an issue with the token rendering that will need a refactor post 1.0+ [20:38] the tokens are never cleaned up even if they don't exist any longer (the machine is destroyed) [20:38] they aren't shown in the dom, but they are still instantiated [20:42] hatch: fun times. [20:55] kadams54: spin up a juju env, I'll need a qa shortly :) [20:55] lol [20:59] jujugui lf reviews and a qa https://github.com/juju/juju-gui/pull/510 [21:00] kadams54: this is the fix to the bug you found in the last branch [21:00] * kadams54 warms up his wrecking ball [21:01] Which begs the question… how exactly do you warm up a wrecking ball? I mean, besides the Miley Cyrus way. [21:01] hatch: On it. [21:02] lol [21:53] Makyo: review done [21:53] I'll hold off on qa until the bug kadams54 found is fixed [21:53] there was one issue [21:54] Cool, thanks [22:02] kadams54: I don't understand your comment - it sounds like the issue that the previous branch fixed [22:02] the tokens should never disappear [22:02] just update [22:06] kadams54: can you check to make sure you have cleared your cache and all that [23:03] Morning [23:05] oh huwshimi I thought you were working the night shift ;) [23:05] or, I suppose the graveyard shift [23:05] :) [23:06] hatch: I replied to your comments [23:10] thanks going back through again [23:10] the test failure was a legit one [23:10] ^ huwshimi [23:11] huwshimi: +1'd with a comment request [23:12] hatch: Oh, all I can see is that the build was aborted. I'll run locally [23:14] huwshimi: http://ci.jujugui.org:8080/job/juju-gui/1678/console [23:14] CI hang so there was a bunch of messed up runs [23:14] hung* [23:15] oh I see [23:17] hatch: ugh, it runs fine locally :( [23:18] huwshimi: try running test-prod [23:18] ah ok [23:20] oops, silly mistake [23:26] kadams54: If you're around, just need a +1 on my branch if you're happy: https://github.com/juju/juju-gui/pull/507 [23:31] huwshimi: imho if he doesn't pop in in the next couple hours - jfsi [23:31] :) [23:31] :) [23:31] hatch: Well there was a QA OK and no code comments... [23:32] huwshimi: the other thing is that when you're submitting branches that are multiple cards, can you put in the pr, the cards which are in that branch, or inversely the pr in the cards [23:32] maybe he thought your code was perfecto! [23:35] so perfect in fact the +1 was not even necessary [23:39] hatch: Yep, I put the pr in both cards, I can do the opposite as well if that helps [23:41] oh I missed that heh [23:41] my workflow is usually from GH to kanban [23:45] hatch: Yeah, me too, I'll put the titles in a comment or something [23:46] just fyi the title and the body of the first post are what gets written into the git log for the merge commit [23:46] so that's why we have been putting the qa comments in an extra comment [23:46] hatch: Yeah, I thought I could do it with the QA comment [23:46] hatch: I wish cards had unique ids, would make this stuff much easier [23:46] ugh I know [23:46] well....they do [23:46] but they aren't really accessible hah [23:49] concrete just about to be poured under the house for my new office :) [23:49] you are getting a basement? [23:49] I'm not entirely sure how concrete gets poured under a house without there also being a big hole :) [23:50] well yes, the garage under the house is being converted into a room for me to work in [23:52] ohh interesting [23:52] don't you park in your garage? [23:52] :) [23:52] hatch: I just park in the driveway, it's a bit too small [23:54] ahh gotcha [23:56] well yay! in that case :)