=== Makyo is now known as Makyo|out [12:45] hmm, a CSS library that replaces the JS in bootstrap; maybe we can use more of bootstrap this way: http://labs.thecssninja.com/bootleg/ [12:47] * bac likes ninjas [13:08] benji: have a moment for a quick chat? [13:08] bac: sure, juju-ui? [13:08] yep [13:08] it's about python-shelltoolbox 'run' [13:53] teknico: did you see my note about clean and clean-all targets? [14:07] * teknico is back [14:07] benji, yes, sorry, I forgot to reply [14:13] benji, replied, I hope you weren't counting on getting *everything* you wanted ;-) [14:13] heh [14:13] as mentioned, I'm also interested in what our colleagues prefer [14:15] I'm mostly interested in not violating expectations. There are many people who will work on this project who are not our current collegues. [14:17] that's also true [14:17] downloading all those node modules takes a while [14:18] having to type something less reflexive than "make clean" to desstroy them might save people some headache [14:19] indeed, that is the raison d'ĂȘtre of the "mostlyclean" target [14:20] you are almost certainly right, unfortuntately the sub-optimal solution has taken hold; I value consistency over slight optimizations; others may differ in their value assignments [14:21] "make clean" gets 1.38G Google hits, "make mostlyclean" 4.17M [14:21] I rest my case :-) [14:29] To all my Makefile-reviewing friends, https://codereview.appspot.com/6909046/ is available for your consumption. [14:35] benji, I'll look at it after reviewing Makyo|out's branch === Makyo|out is now known as Makyo [15:28] bac bcsaller benji goodspud hazmat jovan2 Makyo teknico: call in 2 [15:32] bac: you're a bit windy today :) [15:33] * benji buys push-to-talk buttons for everyone for Christmas. [15:39] gary_poster, in other call [15:40] will join for weekly [15:57] i'm running the jitsu charm test on ec2. it bootstrapped and is now on test_api_agent but has been there a really long time. to be expected? [16:31] nm, it eventually moved along by itself [16:32] interesting thoughts around ux http://blog.stephenwolfram.com/2012/12/what-are-you-going-to-do-next-introducing-the-predictive-interface/ [16:33] bac: right, the tests take a very, very long time [16:33] where we want to go with charm panel / service graph interaction is actually along those lines [16:33] jovan2, goodspud ^ [16:33] except we keep it simple ;-) [16:35] hazmat: looks interesting. [16:39] oh that vpn/proxy ssh tool.. https://github.com/apenwarr/sshuttle [16:40] benji: yeah, but as it was the first run it was hard to tell slow vs hung. [16:40] yep, I had the same problem [16:43] "It looks like you are writing a letter. Would you like help?" -- Clippy. [16:44] (just joking, it looks like an interesting idea) [16:44] Found problem with make. PWD is nolonger defined, leading to: ln -sf "/node_modules/yui" app/assets/javascripts/ [16:44] Can make a microbranch and fix real quick, if that's okay. [16:46] Hmm..unless I'm crazy. [16:47] Makyo: I belive it would be very odd for PWD not to be defined. [16:51] benji, Didn't realize it was an internal, I guess, sorry. Why would it be undefined/empty? [16:56] Makyo: that's a good question; maybe the cwd has restrictive permissions? [16:56] benji, Hmm.. only seems to happen if I run make as root. [17:00] Makyo: I am going to take lunch now, but I will look into it after that if you want. Note, that it might not be a make problem but a shell problem, I believe make just gets PWD out of the environment [17:04] benji, alright. I'll keep poking, too [17:12] hazmat, reading... [17:13] goodspud, its a little dense, and they take a programatic approach to it. but the core to me is context sensitive changes to a user workflow's through the system/use case [17:17] hazmat, so does this mean building in some learning mechanism within Juju/GUI to help it (and us) understand how Juju is being used, outside of basic metrics? [17:19] goodspud, i think we'll need to keep it simpler than that, to encoding some of the patterns that we envision, but collecting in app metrics to help validate, ie. substitute human with data for auto learning. [17:20] hazmat, gary_poster, benji, what's the preferred make target interface for running tests in debug and production? [17:20] right now, "make test" runs tests in the debug environment [17:20] it would be nice if keep those similiar [17:21] I was thinking of "make test-debug" and "make test-prod" [17:21] but then, what will "make test" do? [17:21] benji, on the make clean/mostly clean.. i think in common practice its make clean vs make distclean [17:21] teknico, that's fine; I'd have make test be an alias for test-prod [17:22] gary_poster, which is, again, different behavior from the current one :-) [17:22] hazmat: https://www.gnu.org/software/make/manual/html_node/Standard-Targets.html describes the "standard" targets [17:22] teknico, good point. that's more compelling, I guess, than the argument I was going to make [17:22] hmm.. yeah.. i guess we can't dynamic dispatch in the makefile depending.. since it will do either. and there's no context. [17:23] teknico: I have the same impulse to mirror the "make prod", "make devel", "make debug" with "make test-prod", "make test-debug", etc. [17:23] gary_poster, I was just playing devil's advocate ;-) [17:23] teknico, :-) [17:24] benji, interesting.. i don't think i've seen most of those targets referenced commonly [17:24] teknico, still though, unless we had a watcher keeping prod up-to-date, devs will want make test to have the current sematics, I think [17:25] teknico, and then, when we get a gating test like in lbox, we would use make test-prod [17:25] (or both, but the prod is the most important one IMO) [17:26] gary_poster, then I'll keep "make test" as an alias to "make test-debug", ok? [17:26] teknico, +1 [17:34] uhm, in the line: "node ./test-server.js" [17:34] how do I interpolate values into test-server.js? [17:36] do I need templates? I hope not [17:41] You could pass an argument to the server. [17:41] teknico, ^^^ [17:41] node ./test-server.js debug [17:41] Or something [17:41] Makyo, yeah, I'm on stackoverflow trying to see how :-) [17:41] process.argv apparently [17:53] 78 failures, yay! [17:59] uhm, cannot find a simple reason for all those failures [18:00] but the week is over, if anyone wants to have a try, it would help: lp:~teknico/juju-gui/run-tests-in-prod-too [18:01] have a nice weekend, everyone! [18:07] Makyo: how'd the PWD investigation go? [19:14] anyone know how to get lbox to work with the charm? i get an error "Failed to load focus series for "charms" [19:20] benji: ^^? [19:21] bac: hmm, I'm sure I have used it with the charm code, but I don't remember that problem [19:22] * benji meditates on the phrase "focuse series for 'charms'" [19:23] bcsaller: check out https://codereview.appspot.com/6867080 [19:23] benji: thank you [19:23] it is just about small enough for a self review, but if you cast your eyes upon it and find it pleasing, I will land it real quick and a merge from trunk will get you rolling [19:23] bcsaller: ^^ [19:28] why does the secondary story not have a "review" lane? Is "review coding" for that? [19:34] I don't think its intentionally different [19:34] * benji offers (virtual) cookies to the next person to review https://codereview.appspot.com/6909046/ [19:37] bcsaller: ok, the branch is merged to trunk, after you merge it type "make undocumented" and "make yuidoc-lint" will be much nicer to you [19:37] benji: thank you, looking over your other branch now [19:38] cool, thanks [20:03] benji: would you have a chance to peruse my branch in review? sorry, it is an old skool MP since lbox hates me. [20:03] bac: sure [20:06] benji: thanks. bbiab [20:41] bac: r=me, with a couple of very small suggestions and a question or two [20:50] benji: cool, thanks. [21:04] benji: if you're not against -*- python -*- why do you mention it? :) [21:07] benji: i started the output with \n because otherwise the first line will have the timestamp and level marker followed by the first line of multi-line output. the subsequent lines will then start in the first column. i just thought it looked nicer, and easier to parse chunks, the way i did it [21:07] bcsaller, Makyo: would either of you have a moment to review my branch? it would be nice to get it landed today. [21:08] I have to step out soon, but I can look at it tonight [21:08] bac: :) I just mentioned it because I don't think I've seen that before in Canonical code. Also, there should be a space after the comment character. ;) [21:08] bcsaller: ok, that works. sorry it isn't in rietveld but lbox wouldn't cooperate [21:08] bac: ah! gotcha; a comment explaining that would be good [21:44] gary_poster, the ifeq statement in the makefile did it, I'll do a diff. [21:46] http://pastebin.ubuntu.com/1417690/ [21:46] gary_poster, benji ^^^ [21:46] Would like some input on this. [21:47] benji, re: https://bugs.launchpad.net/ubuntu/+source/sudo/+bug/410022 [21:47] <_mup_> Bug #410022: sudo doesn't propagate $PWD < https://launchpad.net/bugs/410022 > [21:47] Makyo, +1, I'm OK with landing it given our earlier discussion [21:47] thank you Makyo [21:47] have a great weekend [21:48] You too [21:48] Makyo--I suggest you add a comment about why it is necessary [21:48] now really running away ;-) [21:48] Will do gary_poster [21:48] thx [21:49] Makyo: interesting; I would really like to know why it is doing that [21:49] it only happens when make is run as root, right? [21:49] Well, definitely as sudo, but in any environment where PWD is not set in the environment variables. [21:53] "$PWD considered harmful" [21:53] teah, saw that also [21:53] yeah [21:54] gary_poster: would you have a moment to do a quick review? [21:54] bac if it is super dupe duper fast. not really supposed to be here :-P [21:55] i guess the speed is up to you. otherwise ben has said he'll look at it later, which means i could land in the morning, which, for all practical purposes, is just about the same. [21:55] so, go away [21:55] :-) [21:55] ok [21:55] thank you [21:55] have a great weeken [21:55] d [21:55] and so will i [21:55] cool [21:55] ta [21:55] * bac off to read scotusblog [21:59] Makyo: I found it. sudo removes non-whitelisted environment variables because they might leak important information... why PWD isn't in that whitelist by default, I do not know [21:59] regardless, your solution is good, at least we now know why it is good [22:00] benji, I'm not sure, either, and the comments on the bug were not too informative. Thanks for looking; mind if I just land that change? [22:00] Makyo: +1 [22:00] Wow: https://tech.dropbox.com/2012/12/welcome-guido/ [22:00] benji, thanks [22:05] huh, benji, that is a wow [22:14] * Makyo walkadoggy. [22:32] * bac o/