[13:18] hey teknico, approved your holiday request [13:31] benji, I was going to try and dupe teknico's report on your branch, unless you've already verified/duped. have you? [13:32] gary_poster: nope, I have been trying things bug was just about to give up and ask for a pairing session [13:33] ok benji, I'll be ready in just a few minutes & I can at least try to dupe with you. pairing, or talking, with teknico afterwards might make sense (and I have another call in an hour) [13:33] k [13:37] Something I meant to mention earlier to everyone: http://uistage.jujucharms.com:8080/juju-ui/version.js works now, and can be useful to confirm that you are verifying on QA what you expect [13:37] I'll try to remember to mention this on the call [13:38] mm, I'll send an email [13:43] g'morning [13:47] gary_poster, where would i put the card for annotation support given our current stories? [13:48] hazmat, secondary story please (we need it there) [13:50] hazmat, I was just writing an email, but you know the answer to this and can share it in five seconds I suspect. From the email draft: [13:50] On a related note, note that our README is not rendered properly in the charm store (unlike, say, http://jujucharms.com/charms/precise/wordpress). I saw that ReST is supported (see "Creating a README File" in https://juju.ubuntu.com/docs/write-charm.html) so we're not doing something else right. Maybe we just need to move the charm's README.txt to README.rst? [13:51] Referring to http://jujucharms.com/~juju-gui/precise/juju-gui [13:51] gary_poster, only markdown is supported for rich formatting (client side) in the store atm [13:51] oh ok hazmat. So we should switch, maybe. And maybe change the charm docs? [13:51] I mean https://juju.ubuntu.com/docs/write-charm.html [13:52] gary_poster, we can add a card to the charm store for rst support, but its low priority [13:52] right [13:54] gary_poster, we should switch and change the docs.. the gui is going to have the same issue rendering readme formatting [13:54] hazmat, there is no JS ReST parser AFAICT either. [13:54] Right was going to say the same thing [13:54] ie it needs a js renderer or pre rendered format [13:55] * hazmat nods [13:55] I think charm docs should say "Markdown" [13:55] period [13:55] If charm store and GUI don't support ReST, why would you write it? [13:55] and how can we say we support it? [13:56] indeed [13:56] gary_poster, i agreed we should change it, it was originally written prior to these ecosystem of tools [13:57] gary_poster, we could support it via a static pre-rendered pipeline [13:57] that serves up the asset from the charm store [13:57] er.. browser ;-) [13:57] hazmat, sure, but what priority is that? :-) [13:57] indeed [13:58] hazmat, I'll contact jorge and see if he wants a bug for this or if I should just keep this in a card for our own plans [14:02] gary_poster, https://code.launchpad.net/~hazmat/juju/markdown-only-readme/+merge/140203 [14:03] not sure where the docs went to [14:03] they used to be a series off of.. https://launchpad.net/juju [14:06] hazmat, awesome. so I don't need to metion this to/convince jcastro then, and you'll get it merged? or do you want us to help getting it merged? [14:06] ah.. got it.. its still there. i'm just inactive in the charmers team that owns it. fixed and done. [14:06] awesome thank you hazmat [14:06] gary_poster, its merged, doc sites should update when their cron kciks in. [14:06] great thanks again [14:08] gary_poster, np.. thanks for the summary writes up... good morning reading. i pushed the login stuff into review.. [14:08] * teknico is back from lunch [14:09] great hazmat thanks [14:10] gary_poster, branch details are in the card if anyone's interested [14:11] cool. will look asap (other stuff in my queue atm) [14:14] so we're switching the charm docs to markdown? [14:14] great, maybe that'll help me making a fool of myself slightly less often :-) [14:14] hazmat, one other question while I have you, in reference to the GUI hiding the GUI charm: you mentioned the GUI filtering by charm tag. charm tags are not yet implemented, right? So, for now, we'd need to filter by charm name? [14:15] teknico, not sure what you mean about making a fool of yourself less often, but glad you are OK with it :-) [14:15] gary_poster, referring to https://codereview.appspot.com/6945058/ [14:17] teknico, lol, gotcha. np. Other topic: I have not yet tried to dupe the problems you saw in benji's branch. I think I can try to do that in just a few minutes. Maybe you and he can have a chat soon to investigate that and also discuss the Makefile issue you raised face to face...or however benji wants to resolve that. [14:18] gary_poster, yep, I was wondering whether I was the only one to see those prod test errors [14:18] teknico: do you have a moment to work on this? [14:18] right. I did not follow the process and did not run the tests. bad reviewer :-( [14:19] gary_poster, I reordered the cards in the primary ready to code lane, I wonder whether it helps or hinders :-) [14:19] benji, sure [14:19] teknico, cool. unfortunately I find that orderings sometimes don't stay long in the kanban board, but I like what you did on the face of it [14:20] thank you [14:20] teknico: cool; please run a "make clean && make prod" and then open chrome, clear the cache, and then open the front page of the app and see if there are any errors in the console (404s in particular) [14:21] gary_poster, I know I'm prone to tweak things too much sometimes, and people might be thrown off by it :-) [14:21] :-) np [14:21] benji, ok [14:24] benji, done, no 404 in sight [14:25] benji, that's in your branch, not trunk, right? [14:25] teknico: great; now please stop the prod server and run "make test-prod" [14:25] teknico: yep, my branch [14:26] benji, I see 404s on /juju-ui/assets/event-simulate.js and /juju-ui/assets/node-event-simulate.js [14:26] a clue! [14:27] which is weird, because they are indeed included in test/index.html [14:28] uhm, I'm being silly [14:30] they are indeed not in the place where they are requested from, in build-prod [14:30] yep, I think I know what is going on; one second [14:31] if I am right, this is further evidence that we really need to get our caching story straight [14:31] Whoever is working on the login story, do you have the assets you need from our designers here? [14:32] if you want to try something for me while I do this you can stop the test server, start a devel server, reload the test page (it is ok if there are test errors for this excersize), stop the debug server, do a "make test-prod" [14:33] I bet the tests will pass. [14:33] benji, trying [14:33] thanks [14:34] goodspud, good question. that will be benji, but he is looking at something else this second. I was thinking that we would first try to do the corners in CSS (if designers could give the CSS to us it would save bunches of turn-around time but AFIAK that's not happening) [14:34] If that doesn't work then we will need images for the corners [14:35] godspud, otherwise we shoudl have some dimensions of the box [14:35] ooh, godspud, now you have more spudly powers [14:36] gary_poster, I've been working on my super powers but spudly is all I can come up with for the moment [14:36] These things take time [14:36] :-) [14:36] goodspud, dimensions and font sizes of the header and so on would be convenient too...and info on the gradient and shadow [14:37] lol [14:37] gary_poster, I'll chase up Matt and Greg to get them to provide assets and dimensions, if not CSS if at all possible [14:37] thank you very much goodspud [14:37] A pleasure [14:39] benji, what you asked has two different problems. I tried reading between the words, but nothing changed :-) [14:40] I'll wait for you to complete your intestigation [14:40] teknico: what were the problems? [14:41] benji, 1) mentioning "devel" first and then "debug; 2) different ports on "proper" and test servers, so reloading the test page does not work [14:43] (I guess I meant "reading between the lines" :-) ) [14:45] teknico: mmm, indeed; thanks for trying. Meanwhile my test validated the idea so I'll have a fix in the branch soon for you to try again [14:46] benji, great [14:58] benji, you fixed it, great job! :-) [15:01] thanks teknico, benji. have you two also resolved the Makefile discussion? [15:01] gary_poster, not yet [15:01] teknico: ok, the branch should work for you now [15:01] benji, ^^ :-) [15:01] benji, now, about that other small issue... :-) [15:01] cool [15:01] * benji looks back at the review to remember "the Makefile discussion" [15:02] benji, about copying versus linking yui assets [15:03] benji, I see that you added a disclaimer in comments [15:03] ah! I'm not against linking, it is just that the links previously created were not right; the prod and debug sites produced 404s [15:03] right, Gary and I decided that it wasn't worth the work right now to figure out how to get links instead of copies [15:05] benji, btw did you make a bug for the prod CSS issues? would like to have it on the board [15:06] gary_poster: I don't think so. I will do so now. [15:06] benji, gary_poster, that decision doesn't feel right, given the history of changes [15:06] ty benji [15:07] which does not mean it *is* not right, just that it does not *feel* so :-) [15:09] benji, anyway, approved [15:10] teknico: the tweaked branch worked for you? [15:11] benji, it did, as mentioned above [15:11] and in the review [15:12] cool, thanks [15:13] I need to file another bug I discovered during this debug session. The devel server generates 404s for /juju-ui/assets/combined-css/rail-x.png on the trunk. [15:17] teknico (benji), symlinks vs. copies: I like the idea of symlinks for this. however, benji found that the symlink approach we had was not working for these files. I don't think that copies in this case are horrible--the Makefile is limited in what it can do in this case anyway, since we do not specify what files we actually care about--and I cared more about getting the prod tests working, so I was fine with a pra [15:17] gmatic solution of reverting the change rather than investigating it. I'd be fine with a bug to convert the copies to symlinks in the future, and I would thank whoever filed it, but I would triage it as low at this time. [15:17] OTOH, since we don't have any build tests, changes like this are more expensive than they might appear [15:18] s/OTOH, // [15:23] gary_poster, when can we talk about build tests? [15:23] teknico, there's a card for weekly retrospective [15:23] gary_poster, great [15:24] speaking of build tests, while reading the make manual (I am easily entertained) I happened upon the --just-print, --question, and --what-if make options that might be useful for writing tests [15:26] benji, I'm more easily entertained by searching for alternative build systems :-) (there actually went part of my weekend) [15:27] and yes, I have read gary's blog posts, so I'm *not* raising the issue [15:27] (at least not right now ;-) ) [15:28] bac bcsaller benji goodspud hazmat Makyo teknico call in 2 [15:40] Makyo: after you get off your next call ping me [15:40] bcsaller, alright [15:40] goodspud, ping [15:41] Whoops... missed the meeting [15:41] is it over? [15:41] Yes, but you and i were going to talk about fancy relations. [15:41] goodspud, yes, but Makyo wants to talk to you now :-) [15:41] https://plus.google.com/hangouts/_/3bf1056afd2f9bc2b1f136097c5af5369a200312 [15:42] Sorry guys. Chatting with Alejandra and we got distracted [15:42] np [15:42] Makyo, joining [15:47] Makyo, cool re recess, looks nice [15:48] hazmat, apologies if I missed it but did not see reply to following. in reference to the GUI hiding the GUI charm: you mentioned the GUI filtering by charm tag. charm tags are not yet implemented, right? So, for now, we'd need to filter by charm name? [15:49] Also, you made a card titled "Error Collection" with description "Aggregate errors into collection." I dragged that into "Needs Specification" because I didn't know what you meant [15:49] gary_poster, charm name is what i meant [15:49] cool hazmat thansk [15:50] gary_poster, that's the javascript error collector [15:50] hazmat, oh! the OOPS thing [15:50] gary_poster, yu [15:50] yup [15:50] got it thanks [15:54] hazmat, I still thank that needs to be specced--either directly by you or with goals from you and a proposal by us--but I changed the card to hopefully clarify the goal better--it helps me anyway. [15:54] let me know how you want to proceed on specifying that [15:58] * gary_poster has to step away from IRC for a phone call [15:58] back in no later than an hour [16:01] bcsaller, ping [16:01] Makyo: gimme a sec, have to reboot, unity is acting out [16:02] bcsaller, sure [16:17] Makyo: let me open a hangout === teknico_ is now known as teknico [17:32] gary_poster, I won't be able to attend the daily stand up tomorrow. Have to take the good woman to the eye hospital as part of a regular check up she has. [17:41] ok goodspud thanks for heads up [17:57] hazmat, I still owe you a review, however, the lp:~hazmat/juju/rapi-login is not proposed, where am I supposed to write the review? [17:58] teknico, good question.. [17:58] hazmat, also, the card says: "prior branch for diff: lp:~hazmat/juju/import-export" but bzr cannot find it [17:59] teknico, just pushed it [17:59] import-export [18:00] not there yet, it'll maybe take a few minutes? [18:01] teknico, merge proposal -> https://codereview.appspot.com/6935068 [18:02] teknico, https://code.launchpad.net/~hazmat/juju/export-import [18:02] not import-export [18:02] teknico, sorry about that. thanks [18:02] hazmat, no worries, updating the card [18:05] hazmat, about the wire protocol, point 4 includes 'result': 'success' , while points 5 and 6 use 'err': true, it looks a little weird [18:05] teknico, that's in keeping with the error handling across the other api methods [18:06] teknico, result: True would work as well [18:07] hazmat, one would think choosing one of 'result': ['success'|'failure'], or 'err': ['true'|'false'] would be preferable [18:08] I don't know whether there's any such consistency in the rest of the api [18:08] teknico, omission of err is false, we can change the result value [18:08] teknico, for error handling there is absolutely [18:09] hazmat, great, so I'm just missing the bigger picture [18:10] hazmat, so this is pyjuju code, my review will likely not be extremely authoritative :-) [18:10] teknico, error handling for other methods is an exception catch .. api context _on_error [18:11] teknico, fair enough.. given the lack of pyjuju devs, its probably sufficient unless gary_poster is chiming in/reviewing as well. [18:15] mm [18:17] hazmat, I trust teknico's review if that's all you'd like. Ben is another option, of course, if you want a pyjuju dev. But do what you think makes sense. [18:17] I'm happy to look at it, but would be in an hour or so, and I like the idea of you being able to land this and move on [18:18] hazmat, done, fwiw [18:19] biab [18:24] teknico, thanks [18:25] bcsaller, do you have time to do a review on https://codereview.appspot.com/6935068 [18:25] hazmat: I can do that yes. [18:25] Makyo: going to take that review quickly [18:26] bcsaller, alright, poking around translate. [18:31] bcsaller, thanks [18:31] * hazmat lunches [18:40] http://www.sencha.com/blog/the-making-of-fastbook-an-html5-love-story [18:54] bcsaller, responded, hopefully that elucidates the intended scope and purpose which i think is a bit smaller than your comments were referencing. [18:54] hazmat: I suspected that was the case [18:55] bcsaller, main purpose is to have a node with a restrictive acl that can only be accessed if the ws client has successfully authenticated. [18:55] a followup branch will require the ws client to before using the rapi api [18:55] hazmat: yeah, thats what I was reading, but I wasn't sure of the scope [18:56] sounds good, +1 [20:38] hazmat did you land rapi-login already? If so will move kanban card. [20:38] bac, I can pair a bit. maybe worth trying? though you only have a few more minutes before youe EoD don't you [20:39] gary_poster: juju-ui? [20:39] k [20:46] gary_poster, not yet.. monday's are meetingful.. should be by EOD [20:54] cool thanks [21:29] gary_poster: i'm going for a walk. will be biab [21:30] cool [21:55] hmm.. merging out of order in bzr pipeline.. is non trivial [21:58] gary_poster, did you want me to add the gui env api for login as well? [21:58] wasn't sure if you already had that in your branch [21:59] hazmat, not my branch--benji will be doing it--but yes, please [21:59] (IOW, please merge it) [22:00] ack [22:07] hazmat, I did Markdown on a one-off and it rendered locally, but charm store is not happy (http://jujucharms.com/~juju-gui/precise/juju-gui). Do I need to call the file README.md? Looks like that's what the wordpress charm does... [22:09] gary_poster, yeah.. it looks like it wants the .md [22:09] ok cool thanks [22:15] gary_poster, benji login merged to rapi-rollup fwiw [22:15] thank you hazmat [22:15] cool [22:30] yay http://jujucharms.com/~juju-gui/precise/juju-gui [22:56] gary_poster: MP submitted. lbox blew up so it isn't in RV. one issue outstanding. [22:57] * bac dinners