rick_h_benji: I'm trying to see via kanban if someone picked up hatch's review. How is that indicated?12:18
benjirick_h_: we add a "tag" with our name on the "Advanced Settings" page12:18
benji(double-click on the card)12:19
rick_h_ah, gotcha12:19
gary_posterbenji, rick_h_ fwiw you can also just hover over the tag icon for a quick look12:50
rick_h_gary_poster: yea, I didn't think to look at the tag12:50
rick_h_I was expecting to see people put their faces on the card but none had them12:50
rick_h_so was trying to think maybe notes, but nadda there and then decided to just ask 12:50
rick_h_since benji's email just came by and I knew he was around :P12:51
rick_h_luca_: can you make sure to run through the bugs marked 'incomplete' for the stand up today? I'd like to make sure we go through any questions and such.12:51
gary_posterjujugui, today is my daughter's birthday so I'm using some of my accumulated extra hours today to spend some time with her.  Going with her to gymnastics in 10 or 15 and back in an hour or so.  probably long early lunch.  which is a long way of saying...anything I can do for people right now, quickly?12:52
rick_h_luca_: https://bugs.launchpad.net/juju-gui/+bugs?field.searchtext=&orderby=-importance&field.status%3Alist=INCOMPLETE_WITH_RESPONSE&field.status%3Alist=INCOMPLETE_WITHOUT_RESPONSE&assignee_option=any&field.assignee=&field.bug_reporter=&field.bug_commenter=&field.subscriber=&field.structural_subscriber=&field.tag=charmbrowser+&field.tags_combinator=ANY&field.has_cve.used=&field.omit_dupes.used=&field.omit_dupes=on&field.affects_m12:52
luca_rick_h_: will do12:52
rick_h_oh, that was bigger than it looked12:52
rick_h_gary_poster: have fun!12:52
luca_rick_h_: also, is there a known issue of the GUI being down?12:52
gary_posterthanks :-)12:52
rick_h_luca_: ah, was going to ask about that this morning. Figured it was to do with everything going down last night12:53
rick_h_jujugui anyone know if we need to reset the charm/etc? The gui hangs on connecting to env12:53
bacgary_poster: get pictures of her on the parallel bars12:53
gary_posterrick_h_, will investigate.  bac, heh ok  will do 12:53
gary_posterhey hazmat, the readonly stuff has broken improv. looks trivial.  could you make a quick fix?  not really here.  http://pastebin.ubuntu.com/5619066/12:59
gary_posterI mean, I'm not really here. :-) daughter's bday, taking a few hours away in morning12:59
gary_posterrick_h_, luca_ temporarily up.13:06
gary_posterhazmat, we just need improv to have line 21 of http://pastebin.ubuntu.com/5619098/13:07
gary_posterhazmat maybe we should put rapi-rollup in a shared place13:07
gary_posterrather than ~hazmat :-)13:07
gary_posteralso, could you ping bac and me when this is fixed so we can make sure uistage is OK?\13:08
* gary_poster runs13:08
luca_rick_h_: thanks :)13:10
rick_h_luca_: heh, thanks gary_poster 13:10
luca_rick_h_: did you want to screen share to see this bug while its up? https://bugs.launchpad.net/juju-gui/+bug/117439213:13
hazmatgary_poster, noted, fix in progress. will need to coordinate  a charm fix as well13:14
hazmatfor the branch move13:14
* hazmat notes improv should have a unit test13:15
bacso in my raring VM, if i resize a terminal window it jumps back to its original size when i give focus to another window.  anyone seeing that on raring on metal?13:16
rick_h_luca_: sinzui had that and was looking at it. Check with him at the stand up if he's found it and fixed it. Maybe just not released yet.13:17
luca_rick_h_: kk, will do13:17
hazmatrick_h_, fixed btw 13:27
rick_h_hazmat: awesome thanks13:28
rick_h_hazmat: do we need to get the charm updated them as well or is the uistage setup all set 100% ?13:28
hazmatrick_h_, uistage auto refreshes, its sans charm13:31
hazmatit does pulls from both branches on a reset13:32
hatchwe had a slush blizzard last night :/ I was driving home down the freeway and it was so icy that the wind was pushing the car into the other lane lol13:43
hatchand this morning they look like a curling rink13:45
hatchgood thing my commute isn't that far13:45
rick_h_hazmat: hmm, seems it updated and is down again. 14:03
hazmatrick_h_, conflict error.. fixing14:04
rick_h_thanks hazmat 14:05
hazmatwtf.. bzr diff -> http://pastebin.ubuntu.com/5619274/14:07
hatchtime to switch to git? ;)14:08
rick_h_hazmat: sssh, you'll get my hopes up14:08
rick_h_oops, meant hatch 14:08
rick_h_damn autocomplete with people matching first two cahrs14:08
hazmatoh. somebody else logged in to hotfix the code hence the conflict14:09
rick_h_hazmat: ah, sorry right. Gary did a cowboy before he ran. 14:10
hatchrick_h_: lol - I'm sure the functionality is pretty close, but it's almost impossible to find out how to do things in bzr compared to the 1000s of git 'how do I do X' posts14:11
hazmatrick_h_, working now14:12
rick_h_hazmat: thanks14:12
rick_h_luca_: alejandraobregon http://uistage.jujucharms.com:8080/bws/sidebar is back up and should be ok. 14:13
luca_rick_h_: cheers14:13
hatchrick_h_: has your phone updated to the new google play?14:14
rick_h_hatch: yea, noticed it today. Well my N10 did. Haven't chceked my phone14:15
hatchahh - my n7 didn't but my phone did, it's almost unusable on my phone14:15
rick_h_works well on the tablet, but that green on the app store side is fugly14:16
hatchyeah - and the icon/title at the top when viewing an app takes up the top 1/3 of my phone screen14:16
hatchgoogle seems to have a thing for taking up usable space with chrome for no reason14:17
hatch' see google docs on a small screen' heh14:17
hatchthe old store had these fancy banner images and whatnot...this one is just a ton of ungodly slow loading icons14:17
hatchI have a 3.75" screen though so maybe they didn't test it on anything smaller than a 4.5" screen haha14:19
hatchrick_h_: FYI - that comment is confusing to me also....oops :)14:21
bacmorning hatch.  when you're free i'd like to pair with you for a bit to get this plugin properly moved into the tree14:30
hatchalright couple minutes and I should be gtg14:30
hatchbac: ok - lets do it14:34
bachatch: cool.  let me get arranged.  guichat?14:38
teknicogrunt, can't make charm deploy tests work, jitsu watch always failing to connect :-/14:49
hatchdo we know if gary will be back for the call?15:36
rick_h_hatch: not sure. Sounded like it but maybe not15:39
Makyojujugui - call in 10, update kanban.  We'll run with it if gary_poster isn't back for the call.15:50
hatchMakyo: what's been up with you rinternet lately? someone cut a fiber line during construction or something?15:51
Makyohatch, New router.  Will be switching back to the old before the call, if I have time.15:51
hatchahh - yeah that happened to me before - funny how such a 'simple' thing can cause so many issues15:52
hatchbenji: how did you end up with a 10G memory leak? lol my computer would have long since crashed by then :D15:53
benjiI wasn't the original reporter, but I assume swap had something to do with it15:54
gary_posterthank you Makyo! :-) I'm here, though I was thinking I asked for volunteers to run the meeting.  You up for it today Makyo ?15:54
Makyogary_poster, Sure.15:54
gary_posterThanks Makyo :-)15:54
MakyoSwapping router real quick.  Fingers crossed...15:54
=== makyo_ is now known as Makyo
hatchwhat was the old one so I know never to buy one :)15:57
Makyojujugui call in 215:59
Makyohatch, Old one was a WRT54g, which works great, unless I put my phone on my desk, in which case it drops.  Also, it was time to switch to N.  New one was a Cisco E3200 which seems to have some problems running dd-wrt. 16:00
hatchhaha wow wrt54g that's a classic16:00
bacgary_poster: remind me the pw16:03
gary_posterbac password is "password"16:03
gary_posterbac that will go away16:03
bacbenji you just need the hat.  http://assets.rollingstone.com/assets/images/artists/304x304/zz-top.jpg16:07
* gary_poster wants to get benji a beard extension16:08
Makyogary_poster, lost you :/16:20
* benji gets back to the hangout too late.16:40
benjidarn firefox crashed my computer16:40
rick_h_gary_poster: will we have apache-like log tracking for metrics on the juju-gui to help us determine what *real* use patterns end up being?16:55
gary_posterrick_h_, great question.  apache is our front end in the prodstack configuration, I know.  I'll ask news17:00
rick_h_gary_poster: cool thanks. 17:01
rick_h_jcsackett: hatch Makyo review request if anyone has time please. https://codereview.appspot.com/905204317:01
Makyorick_h_, I'll take a look17:02
rick_h_thanks Makyo 17:03
gary_posterrick_h_, confirmed that we can get access to apache logs17:04
rick_h_gary_poster: cool. Maybe an interesting oakland discussion might be analytics/metrics on things. Given xx charms in the intereste/featured which charms end up used more/less and such. What path people tend to use, etc. 17:18
* rick_h_ wonders if there's overlap in things done like in the software center17:18
ahasenackgary_poster: hi, I just tried to install juju-gui using juju-core, basically just juju bootstrap + juju deploy juju-gui, and got this backtrace in the install hook, wondering if you saw that, before I start debugging: 17:18
ahasenackgary_poster: http://pastebin.ubuntu.com/5619768/17:19
gary_posterrick_h_, cool.  those analytics are going to come from the manage.jujucharms.com and charmstore logs, yeah?17:19
rick_h_gary_poster: some, I imagined some would be the urls hit on the gui as well. Kind of correlated together. 17:19
gary_posterbenji, looking at ahasenack's report.  might be tied to your branch, or Kapil's.  ahasenack looking and will report back17:20
* benji looks17:21
ahasenackgary_poster: ok17:21
benjigary_poster: that's me17:21
gary_posterbenji, you on it?17:21
benjiLet me look at fixing it.17:21
ahasenackwhat is the branch, I tried lp:~charmers/charms/precise/juju-gui/trunk but it doesn't even have that file17:21
gary_posterthanks benji.  if it is slow in the slightest please revert.17:21
gary_posterahasenack, lp:~juju-gui/charms/precise/juju-gui/trunk is true trunk.  We are the first experiment towards giving autonomy to blessed charm authors.  We need to not break the charm to have the experiment work out. :-/17:23
ahasenackah, right17:23
ahasenackthe default from the config17:23
ahasenackno, wait17:23
ahasenackthat's the charm branch17:23
ahasenackgary_poster: so for some charms it's taken from somewhere else? I thought all charms were in ~charmers/charms/17:23
benjigary_poster: I'm testing the fix now by deploying the charm to EC217:26
gary_posterahasenack, right, that's the way it has been, and that's the way jujucharms.com reports it, but it's not true for the charm store, and this not true for juju itself.  Like I said, this is an experiment to give trusted charm owners who (appear to? :-/) have a good quality landing process the ability to be masters of the official charm.17:26
gary_posterthanks benji17:26
* gary_poster wants tarmac17:28
* benji too.17:31
=== matsubara is now known as matsubara-lunch
* bac <3 tarmac17:40
benjigary_poster: working charm verified, trunk has fix17:45
ahasenackgary_poster: got something else now after i switched to juju-gui from ~charmers:17:45
benjiahasenack: feel free to give it another tr17:45
ahasenackgary_poster: http://pastebin.ubuntu.com/5619848/17:46
gary_posterthanks benji.  ahasenack. should work now.  thank you for the report.  don't use the charmers one.17:46
ahasenackgary_poster: bzr is not installed, is it normally installed when using pyjuju?17:46
gary_posterahasenack, we will be pushing the ~juju-gui charm to charmers while we are in this transition period. In fact I'll ask m+3 to do so now.17:46
ahasenackgary_poster: is that a juju-core incompatibility? Just curious, will try the one from the charm store now17:47
gary_posterhey m_3.  When you get a chance, could you push the ~juju-gui juju-gui charm to the ~charmers?  I'll make an MP if you need it.  My concern is that people are going to be confused until jujucharms.com points to the right charm, which may not be for awhile.17:48
gary_posterahasenack, I don't know why bzr is not around in that traceback, but what it is trying to do will absolutely not work in juju-core.  The ~charmers branch doesn't support juju-core (until we push the ~juju-gui one over)17:49
ahasenackbzr is not installed in the instance, fwiw17:49
ahasenackat least not in PATH, I didn't actually check dpkg, and I took it down now17:50
gary_posterahasenack, understood.  Thank you, and sorry you ran into this.  We'll improve our process some more.17:50
gary_posterrick_h_, gave you a quick LGTM fwiw17:57
rick_h_gary_poster: thanks17:57
ahasenackgary_poster: deploying juju-gui with juju-core worked now18:10
gary_posteryay, ahasenack!  It should, we worked hard enough for that to happen. :-)18:10
ahasenackgary_poster: the gui user is now called user-admin? Or was that just a ui hint?18:15
gary_posterahasenack, it's a juju core change.  I think we will probably want to hide the user- prefix in the future.18:16
gary_posterit will matter when juju grows authentication18:17
gary_posterreal authentication :-)18:17
=== matsubara-lunch is now known as matsubara
gary_posterbenji, out of curiosity, have you been able to dupe memory leak?18:41
gary_posterAIUI the chrome tools are the best18:41
benjinot yet, I've hacked up an improv script that continuously sends huge deltas but no leaks yet18:42
benjithe chrome tools do seem to be the best18:42
gary_posterbenji how many services/units?18:42
benjiI'm starting to think that the kind of delta might be the issue18:42
benjithe deltas are annotations18:42
gary_posterbenji 5 services, 1 unit each?18:43
bachey gary_poster, in my testing i need to show that the textarea heights are being computed correctly.18:43
bacgary_poster: but i'm getting different values when running test-server vs test-debug18:43
benjifor testing my current hypothesis, the number of services doesn't matter18:43
baci assume this has been dealt with before...  any pointers?18:43
gary_posterbenji, dunno, but he had N services/2000 units18:44
gary_posterbac, um.  the closest tests I know of are the hairy ones for the help text18:44
bacpointers that is for writing a robust test that deals with varying sizes across test environments18:44
gary_posterbac, and sadly they are not the most robust.  they are not horrible, but they are one of the offenders18:45
baci can easily show that the textarea grows or shrinks as expected but fuzzily18:45
gary_posterbac, that doesn't sound too bad...?  :-D18:45
bacif its good enough for my rice cooker...18:46
ahasenackgary_poster: I have a feeling I reported this already, but I don't remember. Can easily file a bug if needed: http://pastebin.ubuntu.com/5620058/18:49
ahasenackhappened when taking down the unit18:49
ahasenackgary_poster: pyjuju doesn't run the stop hook iirc18:49
gary_posterahasenack, ah!  thank you, sure, bug would be great.  was not aware of that myself18:50
ahasenackgary_poster: ah, found the original report: https://bugs.launchpad.net/charms/+source/juju-gui/+bug/117149018:51
_mup_Bug #1171490: "stop" hook fails <juju-gui (Juju Charms Collection):New> <https://launchpad.net/bugs/1171490>18:51
gary_posterahasenack, thanks!18:52
ahasenackgary_poster: can I mark this one as fixed? The branch is merged, but bug is open: #115263118:52
_mup_Bug #1152631: Juju-gui installation fails from lp:juju-gui <juju-gui (Juju Charms Collection):Confirmed for gz> <https://launchpad.net/bugs/1152631>18:52
ahasenackI don't know how you guys handle "Released"18:52
gary_posterahasenack, it's released thank you.  I haven't been bug gardening lately18:53
gary_posterwill get to it soon18:53
ahasenacknp, I saw it by accident, I'm not stalking you :)18:53
gary_posterahasenack, lol, np18:54
hatchhav there been no proposals or anything all day?18:57
hatchI haven't had any emails in a while...18:57
hatchthere they all come....19:11
rick_h_email overload19:11
hatchhaha oh well19:13
* hatch pushes them off till later19:13
hatchdoes anyone have any tips as to how to respond to env.get_service calls from within the application?19:13
fwereadegary_poster, bac: would you consider it a good thing or a bad thing if I were to change Options on ServiceSet (hm, why not Config?) and Config on ServiceDeploy to map[string]interface{} (from map[string]string)?19:14
bacgary_poster: i've got a branch that adds the resizing plugin.  i'm going to get it reviewed and landed and then wire it up with our config ui19:15
bacgary_poster: why do you need the change to the sig?  i guess going from specific to general is ok if there is a reason19:16
gary_posterfwereade, I'm guessing this is to be able to get bools, ints and floats?  if we clearly know what to send you for each type, I think we can adjust.  we probably need a period of backward compatibility.19:16
gary_posterbac I think you meant fwereade for the second one yeah?19:16
gary_posterbac for resizing sounds great19:16
hatchoh I should pass the fakebackend stuff into my environment19:17
bacfwereade: yeah, what gary_poster said.  we can adjust if the change is needed for internal purposes19:18
m_3gary_poster: sure... sorry for the imcomplete changeover19:19
gary_posterm_3, thanks, np19:20
fwereadebac, gary_poster: well, the issue is just that untyped service config data is kinda... wrong -- it's not how we store it and it's not how we return it -- but the actual wrongness is in fact relatively low-impact: I *can* just fix state and leave the API untouched, and I should probably just do that19:24
fwereadegary_poster, bac: I was only really asking in the hope that you'd say "oh god it's such a mess at the moment yes please" ;p19:25
fwereadegary_poster, bac: and in the absence of that, I think it's best to keep compatible19:25
gary_posterfwereade, heh.  The only related thing we'd say  "oh god it's such a mess at the moment yes please" is to divide up text lines and text multi-lines in the schema.  The fact that "string" can be either has consumed a sad number of days of development to produce an attractive UX that can look like a text line but expand to a text area.  It's just about done, but clarifying that would still be fabulous.  The other thin19:32
gary_posterg you describe would be nice theoretically but is not causing us practical problems.19:32
m_3gary_poster: they match now... I'll automate that until jujucharms.com is updated (requires sitdown in oakland evidently)19:35
gary_postercool, thanks m_3!19:35
fwereadegary_poster, heh, I can imagine the hassle, but I can't see an obvious way to resolve it19:35
gary_posterfwereade, cool, then at least we already have almost solved it. ;-)19:36
rick_h_bac: might want to check out http://bazaar.launchpad.net/~launchpad-pqm/launchpad/devel/view/head:/lib/lp/app/javascript/formwidgets/resizing_textarea.js for any corner cases 19:36
rick_h_bac: didn't look close but worth a run through. 19:36
bacrick_h_: well, isn't that interesting!19:38
rick_h_bac: sorry, thought it was tied to the editor stuff in LP and then you guys were going RTE and such so thought you were doing html -> plain text. 19:38
rick_h_but seeing your Y.Plug looked famaliar :)19:39
gary_posterboo hoo19:39
gary_posterrick_h_, does that one have tests?19:39
bacrick_h_: it should've been packaged as a YUI utility and open sourced.  then we would've found it19:39
rick_h_gary_poster: definitely. 19:39
gary_posterrick_h_, bac, boo hoo.  talk it out among y'selves and figure out what to do?  I'll join if desired, and leave it to you to decide otherwise.19:40
rick_h_gary_poster: sorry, just maybe run through mine and see if I covered any corner cases you didn't consider. Maybe check tests for ideas of tests missing19:40
gary_posterrick_h_, well, yours is battle tested, yeah?  My default thought would be19:41
rick_h_gary_poster: well been on LP for 6mo? so not aware of any bugs filed against it. 19:41
gary_poster"Does the LP one behave the way we need it to?  Can we import it directly?  If so, let's use it."19:42
gary_posterbut bac, I'm leaving that decision to you, with me to help out or make a decision if you want me to.19:42
rick_h_quoted from?19:42
gary_postersorry rick_h_ , I meant, that would be my default thought19:42
rick_h_ah, gotcha. Thought I missed an email/convo I should have seen19:43
gary_posterI was quoting my imaginary self ;-)19:43
bacgary_poster, rick_h_: sure, i'd prefer to just use the one in production.19:43
rick_h_bac :( so sorry I didn't come forward sooner. 19:44
gary_posterbac, assuming it meets all our needs19:44
rick_h_bac: really did think you guys were doing more html stuff19:44
bacrick_h_: well, we didn't put out a group-wide call for ideas19:44
gary_posterrick_h_, if you weren't here then we wouldn't even know now19:44
bacgary_poster: yes, after verifying it works for us19:44
bacgary_poster: if we use it, do we need to port the tests over or just treat it as 3rd party code?19:45
gary_posterok bac.  I'll hold off on reviewing the rest of your branch then.  FWIW, the only comment I had so far was that AFAIK we put files like this javascripts/assets and don't need a plugins dir19:46
gary_posterbac, port tests: I would prefer so, yes.  LP is being maintained admirably, but...19:46
rick_h_so bac, first glance I don't copy over all the font attrs, but I don't see you watching for window resize events that might effect the text input. 19:46
gary_posterbac, alternatively you can release it separately :-P19:46
bacgary_poster: we discussed that and thought assets was for 3rd party stuff.  since the plugin was home grown i treated it as such19:46
bacgary_poster: if we use LP's then i'd put it in assets19:47
rick_h_bac: I'd probably have done it in widgets even though it's not technically a Y.Widget 19:47
bacyeah, well...19:47
bacdirectories are cheap19:47
rick_h_bac: true19:47
gary_posterbac, ack. FBOW, ns-routing-extension, reconnecting-web-socket, d3-components, and app-subapp-extension are all precedent for our general-purpose code going in assets/javascripts19:48
rick_h_ok, well let me know if I can be of any help bac. The tests and such are in that dir. To check it's use file a bug and on the bug entry it auto resizes. 19:48
bacrick_h_: thanks!19:49
gary_posterrick_h_, does it start at essentially a single line, or is it easy to make it do that?  bac, does yours work that way?19:49
rick_h_gary_poster: it takes a min-height setting19:49
rick_h_gary_poster: so filing bugs it starts out at some 50 lines ish19:49
gary_postercool thanks19:49
bacwth FBOW?19:49
gary_posterfor better or worse :-P19:49
rick_h_https://bugs.launchpad.net/juju-gui/+bug/1157138 and click edit icon on the bug description box19:49
_mup_Bug #1157138: deploying a new service does not update service name in environment view <juju-gui:Fix Committed by gary> <https://launchpad.net/bugs/1157138>19:49
rick_h_bac: FBOW?19:49
* gary_poster used it19:50
gary_postermy fault19:50
* rick_h_ is slow EOD :P19:50
gary_posterI was overcome with a lack of desire to type "for better or worse"19:50
gary_posterin punishment, I have now typed it twice :-P19:51
gary_posterrick_h_, I'm going to try updating the demo site...19:51
rick_h_gary_poster: k19:51
bacrick_h_: yeah, that seems to be what we want, at least from playing with the bug entry box19:56
rick_h_bac: :/ cool. 19:56
bacrick_h_: i may bug you later.  let me pull it in and see what happens19:56
rick_h_bac: definitely. anything I can do to help19:56
benjiok, I have reproduced one delta-related leak: service annotations cause memory usage growth over time19:56
hatchu go girl!19:59
hatchgota love debugging those memory leaks20:01
hatchusually when you find the cause you're like 'oh duh...no wonder this is causing a leak' heh20:01
gary_posterrick_h_, OPTIONS from manage.jujucharms.com...interesting is not working from the GUI, at least over https20:06
gary_posterrick_h_, see tinyurl.com/jujugui20:06
rick_h_gary_poster: 'not working'? as in no response?20:06
gary_posterrick_h_, I meant tinyurl.com/jujucharms sorry20:07
rick_h_gary_poster: yea, fighting the app cache or something atm20:07
gary_posterrick_h_, it loaded for me on one machine now20:08
gary_postertrying first one again.  rick_h_ the https aspect appears to be a big issue20:08
gary_posterbecause a lot faster over http20:08
rick_h_gary_poster: k, the bug fix for the OPTIONS loading is on staging. but not manage. yet20:08
gary_posterrick_h_, ah ok.  you want me to switch the jujucharms thing to staging?  Maybe just for a second, to see the difference?20:09
rick_h_gary_poster: sure, but it'll be non-https so won't show the https overhead as well I guess20:09
gary_posterrick_h_, oh right then it won't work20:10
rick_h_gary_poster: mthaddon is past EOD but can see if mew can set it?20:10
rick_h_bah, why won't the app cache load? but going straight to it laods20:10
gary_posterrick_h_, you mean change it on manage.jujucharms.com?20:10
rick_h_gary_poster: can see if mew can upgrade the rev on manage.jujucharms.com20:10
gary_posterrick_h_, that would be cool20:10
rick_h_we didn't update because the apache side isn't ready and mthaddon is EOD20:10
gary_posterah, gotcha20:11
rick_h_coming empty for mew in #is 20:12
rick_h_coming up empty that is20:12
rick_h_gary_poster: so yea, OPTIONS tooks 12s for me 11.5 of that receiving the 1.1MB payload20:12
gary_posterrick_h_, ack20:12
rick_h_gary_poster: that's fixed, so the first one will turn into a 200ms request20:13
rick_h_the second one will still suck. I did some research today and it's the svg in the charms killing it. So have a card to re-thjink that. a single charm json dump goes frmo 3k to 48k with the icon20:13
rick_h_since interesting is all the charms we've picked to have pretty svg icons it's all the charms that are giant and just makes it all worse20:14
gary_posterrick_h_, ack, svg makes sense.20:14
rick_h_gary_poster: ok, will hopefully have apache update tomorrow, will get a manage. deploy, and should help make a noticeable differnce with the long term plan to redo how svg/icons work.20:15
gary_posterrick_h_, cool.  if we can svgs separately--and maybe only get the ones we see?--then that sounds good to me to keep the rest as is, at least for now20:16
rick_h_gary_poster: yea, I'm thinking if we ship the visible ones, say 5 by default on sidebar, but do links servied from manage.jujucharms.com for the rest, they'll only load in small batches but be immediate for the first load20:17
rick_h_gary_poster: but that's off the top of my head, will see about a real fix/get more opinions. 20:17
gary_posterrick_h_, +1, though why don't we go for simplicity first and just load svgs as individual assets and see how that fares?20:17
gary_posterless code, better than what we have now, might be good enough20:18
rick_h_gary_poster: true, go all the way the other direction and see if we need to meet in the middle at all20:18
gary_posteryeah, at least with the images20:19
rick_h_gary_poster: right, agree20:19
gary_postertrunk now hides the help text but I'm doing something wrong with a manual charm update...20:20
hatchLFR https://codereview.appspot.com/9035045/ plz :D20:22
rick_h_gary_poster: ok, abentley_ and I will work together to get a test of using links for svgs tomorrow and maybe we can revisit the staging site with some changes EOD tomorrow and re-evaluate20:22
gary_postercool thanks rick_h_ .  any word on how stability of manage is going?20:23
rick_h_gary_poster: so we've got one *known* blocker where updates to the fulltext schema will break it. We have a card to get it into the data migration process to prevent it20:24
rick_h_gary_poster: so should be stable except when we do the ont thing we know that breaks it :)20:24
gary_posterhave there been thoughts on how to make it more robust in general--isolation of some sort?20:24
rick_h_gary_poster: but we also need a migration story for when upstream changes how they present test data which has bit us before. They're going to file a bug from now on, but we don't have a good way to sync the changes20:24
rick_h_gary_poster: honestly, not a ton. Fire-fighting atm. 20:25
gary_postergotcha.  I wonder if that kind of change is what we really need before we say we are ready. :-/20:25
rick_h_gary_poster: yea, but also we pull so much data from other places, bzr, juju gui stats, jenkins testing, we don't really know how fluid/etc those are yet because it's not been used. 20:26
rick_h_gary_poster: we're learning, and catching up. 20:26
rick_h_plus new tech to us like elasticsearch/mongo makes for fun learning and learning how to make it robust as well20:27
gary_posterrick_h_, ack.  the last one (new tech) is just hard, no question.  I was thinking more of the first kind of thing, but granted that you want to trust data sources because it is a lot harder not to.20:29
rick_h_ok, time to run away. have a good night all20:31
gary_posteryou too, night20:32
hatchok now I can finally get onto what I was working on last week wrt this service/charm stuff haha20:34
hatchtalk about a refactor20:34
benjiI hate consuming fundimentally non-visual technical information via video.  Write things down people.20:41
hatch^ that!20:41
jcsackettrick_h_, hatch: can you think of any reason why calling .empty() would work in debugger, but not in regular execution? is "empty" timing sensitive?20:57
hatchjcsackett: no it's syncronous20:57
hatchthe only thing I can think of is that the element is being re-rendered into the hole20:58
jcsacketthatch: well, something should be rendered into the whole right after i call .empty. but with debugger right after .empty() everything renders as i expect afterwards. without, not so much.20:59
hatchhmm without any more context I can't really say21:02
hatchempty as far as I can tell will either error or work because all it's doing is calling remove and destroy on each child element21:02
jcsacketthatch: so, what i need is to completely reset a node; like it's overflow etc, because just calling setHTML doesn't then have the node scrolled at the top.21:03
jcsackette.g. we render something that has overflow, we get scrollbars, but at render we're a few steps down.21:03
hatchnode.set('scrollTo', 0);21:04
gary_posterhatch LGTM with small comments for test branch.  nice.21:04
jcsacketthatch: nope; it already shows scrollTo at 0, and setting it doesn't change it.21:05
jcsackettlike i said, this worked in tests, but the actual ui not so much.21:05
hatchgary_poster: the reason why those clobberings aren't done in the beforeEach is because in the tests that don't clobber - if the code screws up it will cause the test to fail - which is what we want.21:05
jcsackettunless i pause with debugger.21:05
gary_posterhatch ok cool.  then go with my second suggestion :-()21:05
gary_poster:-) I mean21:05
gary_posternot blubber lips21:06
hatchhaha - can do!21:06
* Makyo dogwalks21:08
hatchbenji: are there any docs/email about how to use the npm caching stuff you did?21:09
hatchjcsackett: does the content get deleted with empty?21:09
hatchit just doesn't scroll to the top?21:09
benjihatch: yep, docs/process.rst describes how to use it21:09
jcsacketthatch: if i pause via debugger, i see it get deleted. then on resume the new stuff renders as we would like.21:10
jcsackettif i don't pause, i don't see any evidence that empty was ever called, and the render problem is there.21:10
hatchbenji: thanks - the docs are kind of all over the place haha so I wasn't sure where to look21:10
hatchthe render problem being the lack of scrolling to the top?21:10
jcsacketthatch: correct.21:10
* hatch puts thinking cap on21:11
jcsacketthatch: if it's not a simple thing, don't worry about it. i have just been reminded i have a dinner thing to get to today. i was just thinking there might be something simple about empty i didn't know.21:11
jcsackettbut if it's just a synchronous destroy as i thought, then alas, there is no such luck. :-P21:12
hatchnope that sounds like some odd edge case thing to me21:12
hatchdoes it do the same cross browser?21:12
jcsacketthatch: yeah.21:12
jcsackettconfirmed on FF and Chrome.21:12
jcsacketti'll pick it back up tomorrow, probably bug you more then. thanks for validating my fear, at least. :-)21:13
hatchhaha np have a good night21:13
jcsackettyou too.21:13
gary_posternight all21:14
hatchhave a good night21:14
hatchanyone around for a quick review? https://codereview.appspot.com/9035045/21:29
bacrick_h_: your code went into our tree nicely and the tests seem easy to adapt21:37
hatchhe had already written the same thing?21:37
bacthanks for the review Makyo.  that branch is being abandoned, though, for a better solution.21:39
bachatch: yeah.  very nice too.21:39
bachatch: part of an obscure skunkwerks project called "launchpad" :)  http://bazaar.launchpad.net/~launchpad-pqm/launchpad/devel/view/head:/lib/lp/app/javascript/formwidgets/resizing_textarea.js21:39
hatchlol looking21:40
bacabout a zillion times better than the crap i was writing21:40
* bac walks dog to avoid face removal21:40
hatchthis looks good21:42
hatchgona take a while to convert it all to camelCase21:42
rick_h_hatch: stay away! :P22:01
rick_h_bac: cool, glad it works for what we need then. 22:01
huwshimirick_h_: I wouldn't start work on that icon sizes bug as it might be that Luca is mistaking the sizes with the featured icons which are larger (in both sidebar and fullscreen mode), unless of course things have changed. I was going to fix that up today anyway, but I might just reply to the bug with this note and see what he says...22:58
rick_h_huwshimi: not following23:09
rick_h_huwshimi: I know they said today they'll give us 3 dimensions to use for the badge later 23:09
rick_h_huwshimi: since the badge is always 40x40 and is too big on the sidebar and not big enough on the charm details23:09
rick_h_huwshimi: but my understanding is that if the url is /bws/fullscreen the charm-token icon should be 96x96 vs 64x6423:10
huwshimirick_h_: Right, that's what the bug says, but that's not what the visuals have, I just wanted clarification...23:11
rick_h_yea, that's what I was telling them. They verified that's what they want23:11
rick_h_huwshimi: so I think this one falls into the outdated visuals category :/23:12
huwshimirick_h_: I'm not convinced it's not an oversight... :)23:13
huwshimirick_h_: We need to clarify that they no longer want the sidebar "Featured charms" section at 96px now anyway, so I'll add a note to the bug23:17
rick_h_huwshimi: ok, I'm missing something. WHen was featured 96x9623:19
huwshimirick_h_: Oh, actually the new WIP visual we got in that email thread the other day has 96px for everything. So you're right, it looks like it will change23:19
rick_h_except sidebar, which is 64 right?23:20
huwshimirick_h_: There's no sidebar in that mockup, but I assume it'll be 64. No idea about the Featured Charms section though.23:23

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