[09:47] fwereade: morning [09:47] frankban, heyhey [09:48] fwereade: I am using 1.18.1-trusty-amd64 and I see a failure when bootstrapping a local env without default series (so trusty tools are uploaded) and the trying to deploy a precise charm: http://pastebin.ubuntu.com/7274274/ [09:49] fwereade: that's the output of "tailf ~/.juju/local/log/all-machines.log" [09:49] frankban, arrgh, I know someone in australia/NZ has been working on related issues, I am not really up to date on them though [09:50] fwereade: so, do you know if this is going to be fixed in trusty? this breaks quickstart [09:51] frankban, I will try to find out; but doesn't quickstart create the environment itself? so it *could* set default-series explicitly..? [09:52] frankban, (this is not to say that core shouldn't fix it, we surely should) [09:55] fwereade: yes, I think, now that we ported the GUI charm to trusty, quickstart should be changed to be smart enough and deploy the gui version based on the bootstrap node series, and also add a "trusty" default-series when it automatically generates the first local env. But we also allow users to change the environments, and moreover that bug would hit us again if e.g. the user then uses the GUI to deploy a saucy c [09:55] harm [10:02] fwereade: I also have an unrelated issue: when setting constraints via SetServiceConstraints API, the default behavior is that if a key/value pair is not passed (i.e. arch or mem) the corresponding constraint is removed (or set to its zero value). The only exception seems to be "tags" (defined as *[]string): if they exist and you set constraints without specifying them, they are not removed. For consistency, I expe [10:02] cted them to be unset. [10:02] rogpeppe: FYI ^^^ [10:02] frankban, good catch, that tags thing is definitely a bug [10:03] frankban, they should be removed [10:03] frankban, if there's a bug for that, would you please assign it to wallyworld? he's working on constraints at the moment [10:03] fwereade: should I file a bug? [10:03] frankban, yes please [10:03] fwereade: sure [10:04] fwereade: is there a bug for the local provider/series issue? [10:05] frankban, https://bugs.launchpad.net/juju-core/+bug/1306537 [10:05] <_mup_> Bug #1306537: LXC provider fails to provision precise instances from a trusty host

[10:05] frankban, (so indeed it looks like there isn't anybody on it at the moment) [10:05] fwereade: great, thank you! [10:12] fwereade: filed https://bugs.launchpad.net/juju-core/+bug/1309449 [10:12] <_mup_> Bug #1309449: SetServiceConstraints API does not remove tags [10:32] fwereade, rick_h_ also filed quickstart bug https://bugs.launchpad.net/juju-quickstart/+bug/1309455 [10:32] <_mup_> Bug #1309455: The auto-generated local env does not work in trusty [10:34] rick_h_: for when you are available, please read the above ^^^ I suspect we need to start the process of patching quickstart in trusty/universe [11:17] frankban: looking [11:18] bah, that's not good [11:20] rick_h_: I created a card in urgent for quickstart, if you agree that would be the first think I'll tackle next week (tue) [11:20] frankban: thanks, I should have thought about trusty lxc and trusty gui charm. [11:20] frankban: yep, sounds good to me [11:23] rick_h_: yeah, the problem is not the trusty GUI charm. having that is a good thing and allows for the improvement/workaround I described in the bug. So, the problem will go away if the juju-core fix lands, but we can alleviate it by updating quickstart to use trusty defaults (which IMHO is a good idea even without that bug) [11:23] frankban: right, but I think having quickstart create a trusty environment by default is a sane default for the life of trusty [11:24] frankban: kind of smack in the head silly, but lack of trusty charms and working with trusty based juju environments hid it a bit from being so obvious [11:24] rick_h_: yes, even after trusty, the same way we defaulted to precise till now [11:25] rick_h_: we were not aware of bug 1306537. since this is a bug I guess we should not need an exception to update quickstart/trusty [11:25] <_mup_> Bug #1306537: LXC provider fails to provision precise instances from a trusty host

[11:26] rick_h_: the bug was filed seven days ago [11:26] :-/ [11:27] * rick_h_ goes to subscribe to bugs, missed I wasn't [11:29] anyway, updated this morning... trusty is beautiful! [11:29] jujugui dr appt this morning. Back in a bit [11:45] * frankban lunches [12:00] hey rick_h_ [12:11] bbiab [12:35] jujugui: in comingsoon I am not able to close the inspector: filed bug 1309503 [12:35] <_mup_> Bug #1309503: Unable to close the inspector [12:52] guihelp: I need two reviews for https://github.com/juju/juju-gui/pull/246 [12:54] fwereade: juju add-machine in trusty local envs is incredibly fast, that's amazing! [13:04] jujugui: behold the speed! http://pastebin.ubuntu.com/7275468/ [13:05] back here [13:05] bac: howdy [13:07] frankban: nice! :) [13:10] rick_h_: interview is in 50 mins, right? [13:11] frankban: yes, I'm working on the doc for it now [13:11] rick_h_: cool thanks [13:12] and man, having an MRI done is kind of draining, almost fall alseep if it wasn't for the giant machine screaming at you [13:12] Makyo: around yet? [13:13] frankban: bac Makyo interview doc shared [13:20] * frankban gives initial +1 for quoting "The Ramones" in the screenshots [13:21] lol [13:21] frankban: if you've got a sec, can you reset the code test question on the collab edit thing that Makyo setup? [13:22] rick_h_: looking [13:27] rick_h_: done [13:27] frankban: <3 [13:56] Around [13:57] Makyo: cool, frankban updated the collab thing. [13:57] wonder if bac will be back in time [13:57] Alright, cool. [13:57] oh email, you are so mighty /me digs some more [14:00] * rick_h_ is thinking no bac [14:00] kadams54: around? able to join an quick interview? [14:01] yup [14:01] kadams54: https://plus.google.com/hangouts/_/canonical.com/ed-cashin-team?authuser=1 [14:01] kadams54: https://docs.google.com/a/canonical.com/document/d/18hkqeQBRrS8uC9DNjgz8F7kXHJgICC4eFquxovH5FRI/edit# [14:09] rick_h_: sorry. want me to join? [14:10] bac: up to you, we have kyle stand in at the moment but if you'd like to participate you're welcome to join in [14:50] jujugui call in 10 [14:59] jujugui call in 2 [15:00] jcsackett: were you joining today? [15:00] jcsackett: no planning poker so optional [15:20] frankban: so as a Go kid, interfaces sound like mixins based on that call? [15:22] rick_h_: if you mean embedding is similar to YUI mixins, then I think you are right [15:22] I was thinking of interfaces, the way he described them seemed very python mixin-like [15:23] * rick_h_ needs to find some Go time [15:24] rick_h_: interfaces allows for polymorphism: if the passed struct receives the set of method the interfaces declares, then it's ok [15:24] ok, so it's a bit of the opposite [15:24] nvm, I see now [15:25] the struct declares the interfaces and mixing the declarations and fails if it's missing any bits from them [15:25] rick_h_: e.g. I can define a function accepting a io.ReadWriter. That means I can pass any value implementing both Read(p []byte) (n int, err error) and Write(p []byte) (n int, err error) [15:25] right [15:26] rick_h_: my understanding is that he described it fairly well, he then switched to embedding and described it with a containment metaphor [15:26] right [15:26] I've not run into embedding so that was cool [15:27] sounds like a form of composition that could be useful [15:28] rick_h_: yes AFAICT it's like delegation + a dynamic dispatcher. So if b receives foo() and a embeds b I can either call a.foo() or a.b.foo() [15:29] oh, auto dispatching is interesting [15:40] yes, especially considering it can help implementing interfaces [15:54] frankban: can you toss the link for your branch in the card? /me is still sorting emails [15:55] rick_h_: it's already in the card [15:55] rick_h_: https://github.com/juju/juju-gui/pull/246 [15:55] bah, one of these days I'll get used to that external link [15:55] I keep looking in the description [15:55] heh [15:55] I always think of the link as the bug number [15:56] thanks [16:03] rick_h_: sorry i didn't respond to ping earlier--i've been on and offline all morning with bad internet. (and don't know for sure that i'm actually connected now, tbh) [16:03] jcsackett: welcome all good [16:12] frankban: review in, let me know if Makyo can't get to qa and I'll try to load it up [16:12] QAing now [16:12] * rick_h_ goes to cook up lunchables now [16:12] rick_h_, frankban ^ [16:12] thank you both! [16:57] hazmat I attempted to use bumblebee and it bricked it until I reset all of the display drivers. I am now running nvidia-prime but I think that's running the gpu all the time :/ [18:05] hey rick_h_ i was thinking if you didn't want to mess with that reservation (cc and what-not) we could just do 2x 6 pax via opentable [18:08] bac: all good, I read through the email. I thought they just needed me to call but I've got to scan some cards and such [18:08] bac: which is crazy, but will get it done [18:08] just hope jcsackett appreciates the extent we go for this :P [18:08] i suspect he does not [18:08] rick_h_: yeah, i just get nervous sending cc info like that. very non-standard and silly. [18:09] yea, true [18:33] rick_h_, bac: y'know, if public house isn't easy to do, it's totally fine with me if something else happens. i'm just saying it's a fine option. :) [18:33] jcsackett: they're making it darn hard to get a reservation. They want me to email copies of my CC and license [18:33] jcsackett: they were not as easy to deal with as they could be but i think it is almost done [18:33] rick_h_: damn. [18:33] that's...a little crazy. [18:34] rick_h_: they say why they need something like that? [18:34] because pie [18:34] :P [18:34] jcsackett: you can do up to 7 people in opentable. over that requires TSA-like scrutiny [19:41] jujugui: i'm contemplating what i can start messing with monday morning so i'm not twiddling my thumbs before the 11am call. is bug 1309503 a likely candidate, or are there dragons? [19:41] <_mup_> Bug #1309503: Unable to close the inspector [19:43] jcsackett: yep, let's chat as there's new stuff there [19:43] rick_h_: okeedokee. [19:44] otp at the moment [19:44] rick_h_: ok. ping me when you're free--i'm largely open this afternoon. [19:52] jcsackett welcome back :) [19:52] hatch: thanks. :) [20:34] rick_h_: does juju gui support deploying local charms in bundle format? [20:34] "in bundle format" ? [20:34] sorry i'm being silly with terminology. [20:34] it displays local charms now [20:34] i have a bundle, with full paths to local charms [20:35] lazyPower: no, juju core increments the version number and things break down. [20:35] so you have to use the deployer [20:35] and they'll show in the GUI [20:35] but you can't drag/drop a bundle with local charms and have it work [20:35] thats what I was lookin for. Thanks my friend. [20:35] I was using quick start and things "just worked", i tried the drag and drop method and it went funky [20:35] so, much appreciation for the clarification. [20:58] jcsackett: still around? [20:58] jcsackett: if not I'll be up and at em in the monday AM and we can catch up then [20:59] my first call isn't until 9am so we can catch up before that [21:00] * rick_h_ is out, have a good weekend all! [21:08] rick_h_ I see you're out and I'm starting to cook dinner. Monday AM works for me. [21:08] Have a good weekend, all.