[00:02] huwshimi: hey. https://plus.google.com/hangouts/_/calendar/Z2FyeS5wb3N0ZXJAY2Fub25pY2FsLmNvbQ.87augjtfc2pm1q1ilfuhencae0 when you are ready [00:03] gary_poster: I'm guessing you can't hear me? [00:04] huwshimi: I can't even see that you have joined [00:04] I will reload [00:04] gary_poster: Hah, I just realised it had an error [00:04] heh ok [12:05] frankban: mornuing [12:05] errr morning [12:05] good morning rick_h_ [12:05] frankban: wanted to check in on something jcastro hit yesterday. quickstart would only work with the bundle file named bundle.yaml? [12:05] "Welcome to Moleman in the Morning, I'm Moleman and you're listening to my mourning." [12:06] frankban: we export the file as export.yaml, which was making him rename to get things tested [12:06] frankban: is there any history/reason we don't accept any filename? [12:06] Makyo: that sounds like it's from something? [12:07] Simpsons. [12:07] rick_h_: IIRC quickstart accepts both a dir path (in which case it looks for a bundles.yaml file) and a file path. If you use a path to a file, the file can be arbitrary named [12:08] frankban: hmm, ok maybe he was using it wrong then or something. He made it sound like he was trying to give it the export.yaml file path and it would not let him [12:11] rick_h_: this is the pertinent code: http://pastebin.ubuntu.com/6726230/ [12:15] frankban: thanks, I'll try it out. Sorry to bug you. What I get for not taking the time to set it back up locally and test it myself :) [12:19] np :-) [13:12] rick_h_: did you see https://codereview.appspot.com/44540043/ ? [13:13] and morning btw ;-) [13:14] gary_poster: missed that one in the emails this morning. Thanks! [13:14] welcome [13:16] hey dimitern. do I understand correctly that you are proposing a backwards-incompatible client API change that would affect the GUI? [13:22] uh-oh. looks like we have some regressions before we can release. http://comingsoon.jujucharms.com/sidebar/search/precise/mongodb-21/?text=mongodb#bws-code does not render, for one [13:22] (or any similar) [13:24] wtf, the hash does it. Take that off and it loads [13:25] * rick_h_ is is mind-boggled at the hash in the search string doing that but ok [13:25] isn't that link in the wrong order? [13:25] should be hash first, then query string? looking up [13:25] possibly, but it works on jujucharms [13:25] getting them out of order is bad for other reasons though [13:26] https://jujucharms.com/sidebar/search/precise/mongodb-21/?text=mongodb#bws-code works and window.location.hash works and such [13:26] no, hash at end is correct [13:26] so I think that's the right order [13:26] yeah, sorry agreeing [13:26] "no" was to self :-) [13:27] we did the mess with removing fullscreen this release, but not sure off the top of my head would case that no-render issue [13:30] I'm diagnosing other unrelated issues reported by Maarten. So far known, and something I'll punt back to him. [13:32] gary_poster: was this via email somewhere? Or bugs? /me isn't seeing this and starting to feel like he's email blind today [13:33] rick_h_: the items from Maarten are private to me. First problem is that he is using dots in his option names, which neither Juju nor GUI support. [13:33] that came up in irc yesterday. hazmat implied only the gui had an issue with it and they were discussing proof adding a check for it [13:34] gary_poster: I meant to ask you about if that was something we should file as a bug/fix if the gui is the only straggler with the . issue [13:34] rick_h_: afaik juju still has the problem [13:44] guihelp: I need two reviews for https://codereview.appspot.com/50430043 (enable the env management functionality in quickstart). anyone available? thanks! [13:45] I encourage other people to do the review, but otherwise I will probably be available in 30min or so [13:45] gary_poster: Kapil says try bribing next time, violence won't fly. [13:45] Makyo: ok, I'll spring for a candy bar :-) [13:46] lol [14:10] gary_poster, sorry, afk, back now [14:10] dimitern: np. [14:10] gary_poster, my proposal is about the final look of the API for 2.0 [14:11] gary_poster, we'll support the current API up until the last moment and drop it before releasing, thus cleaning up the code and giving you guys some time to adapt [14:11] gary_poster, at least that's the general idea [14:11] :-) [14:13] gary_poster, and the reason is we're not really comfortable with supporting the current messy api for 5y [14:13] gary_poster, bluntly put :) [14:15] OK dimitern, we'll want a more detailed migration plan and consideration of pros and cons of such. Migrating for us will not be trivial, and we have a *huge* list of things that people want from us. In terms of design, I'll also ask GUI folks to review your proposal as it stands. It would help me a lot if you were to include the current API that you are replacing. Are you willing to do that? [14:16] dimitern: I'm really quite concerned that you are proposing losing the current API by April [14:17] dimitern: I'd be more comfortable with "we have a new API; we promise to support it for 5 years; and we have a legacy API in this release that will be dropped sometime over the next five years" [14:17] gary_poster, the current API is included btw [14:17] gary_poster, nothing is set in stone yet, but we're discussing this migration informally from some time now [14:18] gary_poster, we should've had a discussion with you guys (and still can organize it) [14:18] dimitern: current API is included: I'm afraid the way to read the doc is not clear to me then. which is which? [14:18] gary_poster, the first part is the current API [14:19] so, for relations [14:19] add relation and destroy relations is old [14:19] gary_poster, grouped by "entities" with some duplication as needed [14:19] no [14:19] gary_poster, yes [14:19] oh ok [14:20] dimitern: but "getannotations" and "setannotations" is the replacement? [14:20] maybe quick hangout would be faster dimitern? [14:20] * gary_poster not feeling smart ;-) [14:20] gary_poster, no, Get/SetAnnotations is there now, but it can work on more than one entity (as per the given tag) [14:20] gary_poster, sure [14:21] perhaps i can pull in fwereade as well [14:21] dimitern: https://plus.google.com/hangouts/_/7acpieh5jic99chb2thg99kcg4?hl=en [14:21] fwereade, are you around for a quick g+ ? [14:22] dimitern, heyhey [14:57] rick_h_, benji, bac hi. mectors is trying to dev a charm. he wants to deploy something from the charm store. ingestion is taking a while (still on the order of minutes, so not necessarily an indication of a problem). (1) what's the usual timeline for ingestion? (2) is there a URL to see the ingestion queue? (3) I saw a mention from bac yesterday that a bundle was stuck. is the queue stuck? [14:58] manage.jujucharms.com/heartbeat [14:58] gary_poster: 15min ingestion time and about 8ish minutes to process the queue [14:58] gary_poster: ^^ url to watch the queue fill/empty [14:58] gary_poster: but nothing that will specifically list it [14:58] gary_poster: make sure proof passes on his charm before anything else [14:58] gary_poster: normal ingestion cycle time can be 30 minutes or more (depending on how lucky you are) [14:59] gary_poster: shows about 1500 charms in the queue. they get processed at about 200/minute [14:59] also, the stuck basket doesn't impeed (sp?) other things in the queue [14:59] bac, rick_h, benji, perfect thank you [14:59] benji: i don't think so. i've been seeing it process in under 15 minutes and sleeping a bit [15:00] bac: right, but if you push your branch just after enqueueing you'll have to wait through an entire injest/enqueue/injest cycle [15:00] so I would bet worst case is around 30 minutes [15:00] benji: oh, well try to not be so unlucky! [15:01] benji: yes, that is true. === hatch_ is now known as hatch [15:01] benji: and now that we aren't cron-based we don't know when the cycle starts [15:02] yep [15:39] hi hazmat, in the juju-deployer README, the section "Stack Definitions" stops abruptly as if you were in mid-thought and someone offered you pie. [15:40] jujugui call in 20 [15:49] 20? :-) [15:50] jujugui call in 10 [15:50] lol [15:50] shoot, I joined up and sitting here bacause of the ping [15:51] I'm a trained monkey [15:52] lol [15:52] * hatch hands rick_h_ a banana [15:58] jujugui call in 2 [15:58] now there is an actionable warning [16:00] benji pingy [16:17] hey frankban [16:32] oops [16:32] gary_poster [16:32] I'll rejoin? [16:32] hatch: np, wanted to see if you wanted to chat or if everything is cool [16:32] I don't need to talk unless you do [16:32] oh nope nothing else [16:33] anything on your side? [16:33] nope. :-) thanks hatch. have a nice trip and vacation [16:33] thank yas! [16:37] hatch: https://plus.google.com/u/0/+PaulIrish/posts/EhmRC5skp6z [16:38] Yay! [16:38] * hatch does a little dance [16:38] hatch: if they get promises in that fixes the native impl issue [16:38] only with async call stacks [16:38] lol [16:38] but that answers my question about the native promises and async callstacks [16:38] curse you promise folks...might actually get it into stuff yet [16:38] it's ok, just accept that you're wrong :P [16:39] * rick_h_ goes to grumble in his "too damn much syntastic sugar will be the death of me" corner" [16:39] lol [16:39] well if it's part of the native api with async callstacks it's no longer sugar [16:39] it IS the api :D [16:40] hatch: true, but some of the new es6 stuff drives me nuts as well [16:42] arrow functions, that's what had me going bonkers [16:46] rick_h_ yeah I don't like arrow functions in any language [16:46] personal preference [16:46] yea [16:46] glad someone else feels the same way :) [16:47] don't agree with me or the world will end [16:47] Did...did you two just agree on something? :o [16:47] rofl [16:47] *boom* [16:47] :-) [16:49] hey jcastro [16:49] hey so I noticed that the gui exports export.yaml as a bundle [16:49] but quickstart expects bundles.yaml [16:49] I was wondering if you had plans to just have quickstart accept whatever.yaml or if I should file a bug [16:50] or even if you think it might be a good idea [16:50] jcastro: IIRC you can pass a full path to an arbitrary named file in quickstart [16:50] huh that didn't work for me [16:50] I'll try it again and get back to you! [16:51] jcastro: cool thanks [17:00] Alright, I'm out _o/ [17:00] jujugui looking for a review on my branch from this week https://github.com/juju/juju-gui/pull/73 [17:00] bye Makyo, safe travels [17:00] Will review at hotel [17:01] Cheers [17:01] frankban: do you have a link for your review you needed? [17:01] rick_h_: it's linked to the card [17:01] sorry, thought I looked [17:02] ah, under external link [17:02] rick_h_: np, new kanban puts the link at the bottom [17:02] jcastro: just worked as expected here: http://pastebin.ubuntu.com/6727619/ [17:04] * hatch is taking off as well, talk to you all later [17:05] bye hatch [20:49] ugh, to test that monster bundle i have to ingest all of those named charms. [20:49] great time to take the dog out [20:50] bac: can you mock out the part of the test that verifies a bundle exists to always be true? [20:50] bac: enjoy [20:50] rick_h_: could but i want to see it work for realz === gary_poster is now known as gary_poster|away