=== gary_poster is now known as gary_poster|away === gary_poster|away is now known as gary_poster [12:55] * gary_poster back in a few [13:16] jujugui: i'm currently unable to get my canonical email. use fn.ln@gmail.com if you need to send me something before it is resolved. [13:16] ack [13:17] bac: rgr [13:17] bac: hi, I am assigned to the change log in GUI card, but since I'm not sitting next to you at the moment, I don't know how to contribute ;) [13:18] benji: yeah, i should take your face off and continue with it. just a matter of resolving the data differences between charms and bundles. shouldn't take long. [13:18] sounds good [13:19] I've never filed mileage to the airport, but it is over 100 miles round-trip; do we do milage? [13:19] what's the alternative? just eating it? [13:19] yes benji, pretty sure. check company handbook, or ask your manager to do so ;-) [13:19] heh [13:27] morning [13:28] Hey [13:28] hatch: morning [13:28] I also couldn't access my email :/ [13:29] Can you get to it through the gmail interface? [13:29] it appears to be back up now [13:29] it wouldn't accept the pw for a bit there [13:29] gary_poster: expenses filed. can you confirm receipt of email? [13:30] hatch: that's what i'm seeing. mine i think is self-inflicted due to my inability not to fiddle with shiny new bits, i.e. icloud keychain. [13:30] ahh [13:30] bac I got receipts, yes. 509KB and 435 KB. Need me to actually look at them? :-) [13:30] gary_poster: nope [13:31] cool [13:31] gary_poster: just confirms canonical smtp is working for me just not imap [13:31] ah, gotcha [13:32] we made the mistake of trying san juan's best chinese restaurant last night. big mistake so soon after SFO. [13:34] I had so much Szechuan food out there. It was heaven. [13:52] bah, just chased down a 'bug' in my branch that fails on comingsoon. Doh! [13:58] bac why no Friday Burlingame dinner in expenses? [13:59] gary_poster: i was invited to be jc for the day for the orange team dinner. curtis paid. [13:59] bac oh cool [13:59] lol [13:59] An honor! [13:59] we went to a cool beer cellar [14:01] ok, I've approved everyone's canonicaladmin things (except a couple of 2014 things from hatch because I want to make sure we get those handled correctly with HR) [14:01] kewl [14:01] rick_h_, are you already reviewing huw's bundle o' branches? [14:01] if not I can [14:01] and antdillon's? [14:01] great. gary_poster i've still got a few days of vacation that aren't set. i'll try to figure that out soonish. [14:01] I am on call now [14:02] so if anyone can take them that would be great [14:02] I'll take them all [14:02] gimme gimme gimmee [14:02] gary_poster: no, I've been tryingto wrap up my 'quick' make deploy work from quicksearch atm [14:02] cool [14:02] my ubuntu vm being broken is really irritating :/ [14:05] hatch: you cheating punk... [14:05] what did I do now? [14:05] I've spent the last 2 hours trying to figure out how your bundle details view event freaking works [14:06] because it seems like it shouldn't based on everything in bundle.js, quicksearch, etc. [14:06] which is why you can't click a bundle in quicksearch and get it to open the details view [14:06] finally figure out how you got it to work from the results page [14:07] oh yeah? how was that? [14:08] hard coding a 'charmid' in the bundle handlebars to be '/bundle/{{id}}' [14:08] so as long as you're dealing with something handling the token click event it 'works' by pretending to be a charm [14:09] so...You're saying I did it properly but that's not the way it's designed to work? :P [14:09] I'm saying you cheated [14:09] antdillon: you will have to run the tests on yoru branch [14:10] there are quite a number of failures [14:10] rather than make things support the idea of a 'bundle' you pretended to be a charm in one case you needed to work :P [14:10] ohh right right - I actually have a note here that says 'change browser charmid to entityid' [14:11] heh, I just added a very similiar note [14:12] see this is really your fault....you should have KNOWN we were going to introduce bundles 6mo into the future and designed it to support that [14:12] * hatch runs [14:13] hatch: come back here, I still can't get a bundle deploy to work [14:13] hatch: let me know if you have a sec to chat [14:13] hatch, Huh, fails? OK will check it out [14:13] lol [14:13] rick_h_: I meant I was running away from you swinging at me for that comment :) [14:13] antdillon: it's ok, hatch's machine just probably timed out and had to take a breather from all the work [14:13] lol [14:14] rick_h_, He's getting a new one soon [14:14] hatch: yea, I know. now come back and tell me how the ##$@# bundle deploy is supposed to work [14:14] rick_h_: so bundle deploy doesn't work? [14:14] unos momentos capitanos [14:14] hatch: I'm trying to call it from a different place and it's failing on me in a confusing way [14:15] rick_h_: the subapps/browser/views/bundle.js listens for a click on .bundle .add and then calls _deployBundle [14:15] hatch: yea, ignore me. Stupid mistake I think. /me reloads [14:16] hatch: yea, got it called, now it's a real bundle failed error. All good [14:17] * rick_h_ has a lot of freaknig tests to write around this [14:18] is this the deploy from the autocomplete? [14:18] hatch: yea [14:19] you sure there is a lot? A bunch of it is tested already [14:19] hatch: but yea, this is why the quicksearch results didn't close for you, there was an error [14:19] ohh [14:19] does this also fix the timeout issue? [14:19] where the ac pops open after a search if it's slow to return results? [14:19] hatch: yea, I've got to do a bunch of conditional checks in there for bundle vs charm, firing the right deploy event, the views listening for that and handling a deploy correctly [14:20] hatch: no, where when you hit enter or something it fails to close. It's trying to process the selection as a 'charm' but fails [14:20] ahh [14:20] odd there was no console error [14:21] #1237457 I think [14:21] <_mup_> Bug #1237457: Hitting enter quickly in autocomplete causing result list to appear over real results [14:21] yeah that's the one [14:21] bah, or not. It's the results coming back too long afterwards, nvm [14:22] ok, well anyway, carry on. Thanks [14:22] :) [14:26] antdillon: fyi - now that I'm back on my big monitor the white space has returned on the footer of the fullscreen view - is there anyone there with a big high res monitor you can test this out on? [14:26] if not one of us can I'm sure [14:28] Switching computers, back in a sec. [14:29] here's an article on the rebase/merge decision when using git: http://blogs.atlassian.com/2013/10/git-team-workflows-merge-or-rebase/ [14:29] I can't say I fully appreciate all of the points, but if we are going to start using git we'll need to figure this stuff out. [14:29] hatch, Yep, I'll find the biggest I can find ... will have a fix in soon [14:30] benji: hatch, rick_h_ and I started this discussion last week. We're leaning toward rebase. [14:30] haha thanks :) [14:30] +1 for rebase (hoping there is some type of 'auto-rebase') we can use [14:30] They're leaning toward rebase. I hate rebase, but given that it's basically just a personal preference, meh, will go with whatever :) [14:31] benji: +1 it's a long debated topic. [14:31] yeah - it needs the collapse functionality of bzr :) [14:31] not knowing much about it, rebase sounds like a reasonable default [14:31] benji: imo the rebase is good, but that's part of the win for the master vs develop branch. master is very clean, develop has more noise. [14:32] benji: the real goal is to keep git biset a useful and meaningful tool [14:33] right - bisect friggen rocks [14:34] although as long as someone rebases 'checkpoint' commits and only commits logical ones I'd be happy with that too [14:34] just don't want broken commits in the shared branch [14:35] hatch: yea, won't have that. You rebase in the branch merged to usually. Though what that means for landing branches is still tbd [14:35] Well, no one does :P [14:35] rick_h_: sorry I have no idea what you were trying to say there :D [14:36] hatch: so typically the process is "do work in my feature branch...get things working, qa'd, etc" [14:36] hatch: then you git co develop, git merge feature, git rebase -i all that feature crap [14:36] hatch: then git push origin develop [14:36] so your feature branch is still a non-rebased mess, but develop is cleaned before pushing [14:36] so no broken commits anywhere [14:37] git co ? [14:37] hatch: but if the lander is doing the merge into develop, will have to figure that out [14:37] git checkout [14:37] ahh [14:37] don't you use aliases for things or do you just love typing? [14:37] love typing [14:37] lol [14:37] Me too :) [14:37] I only alias long statements, not commands [14:37] http://paste.mitechie.com/show/1051/ [14:38] alias the $#$@ out of everything. makes it nice to sync aliases between bzr/git as well [14:38] rick_h_: I like that - my only concern with that is that the dev could accidentilly rebase away others commits [14:39] We could sync with master if that happens, right? [14:39] as long as you don't use -f [14:39] -f: not even once [14:39] we'll see. if merging to develop is automated it shouldn't be an issue [14:39] ohh ok I see what you're doing [14:40] just repeat after me, nothing is ever lost, the reflog is my friend :) [14:40] error: Pre-requisite lp:~huwshimi/juju-gui/remove-old-inspector-styles not yet merged. [14:40] hehe oh lbox u so smrt [14:41] not sure how it knows that [14:42] https://code.launchpad.net/~huwshimi/juju-gui/more-css-removal/+merge/193002 " [14:42] hatch: see the prerequisite field [14:43] oh cool!! [14:43] I like the new loading dialog [14:43] (on the gui) [14:45] shadowrun is 33% off today http://store.steampowered.com/app/234650 [14:50] jujugui call in 10 [14:51] 2013/10/29 08:51:09 RIETVELD Failed to process response: server returned "403 Forbidden" [14:52] best message to see after spending 5minutes submitting [14:52] You can still merge on Launchpad, though, right? [14:52] I could I suppose - it did pass lint and both tests [14:53] Yeah. [14:58] jujugui call in 2 [15:14] jujugui: i can not read my canonical email again. fire away. [15:14] bac, you mean s/not// ? [15:15] yes, s/not// [15:28] hatch, I have a fix for the white bar in browse: lp:~ya-bo-ng/juju-gui/white-bar-fix [15:28] hatch, Your screen must be huge! [15:29] antdillon: 31" 2560x1600 [15:29] so yes [15:29] lol [15:29] and thank you very much :D [15:30] darn, i was thinking my internet was working so much better [15:32] hatch, I think "bundle charm token click through works when it shouldn't until the events are updated" might be fixed, from me trying to dupe. Do you know anything about it? [15:33] gary_poster: it always did work [15:33] it shouldn't work [15:33] that's the bug :D [15:33] well [15:33] unless someone fixed it [15:33] which is entirely possible I guess [15:33] hatch, oh ok. could you come by https://plus.google.com/hangouts/_/calendar/Z2FyeS5wb3N0ZXJAY2Fub25pY2FsLmNvbQ.j0rk5d371ph8331ijtf48t2uj0 really quickly for higher bandwidth? [15:33] hatch: there was some cleanup during the sprint with ant's and huw's branches I helped work ok that might have fixed it [15:34] oh awesome [15:34] hatch: but didn't think to test it...so no promises [15:34] +1 [15:38] hatch, Did you manage to sort out your comp yesterday? [15:38] antdillon: nope just using my VM as if it's a server atm [15:39] hatch, Ouch [15:39] hatch, need that new comp sooner then you thought [15:40] hatch, Give me a shout if you find anything wrong with the linkify branch [15:40] sure thing - it looks good, will just need to QA [15:41] antdillon: one pedantic thing I -could- say is to use more descriptive function names [15:42] but linkify is close enough :D [15:42] hatch, I spent more time on trying to pick a name for the function then I did in the function :/ [15:42] rofl - it happens [15:43] benji, first priority for you is helping rick_h_ on charmworld proof code. If rick_h_ has something you can tackle, great. If not, please tackle the charmworld metrics story. We can launch without that, but if we have it, it would be fantastic. [15:44] gary_poster: sounds good (I'll clarify the details of charmworld metrics once I get there) [15:45] gary_poster: benji if benji wants to get at the relation proof bits it's open. Otherwise the metrics is greenfield fun [15:46] benji, rick_h_ I'll leave it to you two to discuss and decide if benji can take over productively. That's our highest priority remaining. If you want me to participate in a call I'm happy to. [15:46] (But more than happy to leave you two to decide) [15:47] benji, for metrics, ack on details. cards exist at least, thanks to sprint discussions [15:47] gary_poster: this charmworld bug will be quick and then Rick and I (and maybe you) will get on proof stuff [15:48] gary_poster: benji one other thing that might come up is there's a charm someone ping'd about in irc yesterday not getting updated. It had a proof error, but had another partial update looking case. [15:48] perfect thanks [15:48] k [15:51] jcsackett: jcastro I'd like to get the Ghost charm in the store today/tomorrow - what's the git > lp story? [15:52] antdillon: did you say you updated your linkify branch to fix the test failures? [15:52] I don't see any new revisions [15:55] hatch, Running tests now, is it not due to the bundle stuff? [15:55] no idea, I just saw the failures and moved on, sorry :) [15:56] hatch, No prob, on it [15:56] from now on because you can't lbox you can just run this command `make lint && make test-debug && make test-prod` [15:56] before you push [15:56] it'll take a long time but avoid this in the future [15:57] hatch, Cool, I've been testing and lint before pushing but must have missed something sorry [15:57] no problem [15:58] I'm still baffled by your lbox issues [16:03] gary_poster: I haven't read it yet but here is the review for the new mbp from engadget http://www.engadget.com/2013/10/29/macbook-pro-with-retina-display-review-13-inch-2013 [16:09] hatch, cool. I want anandtech to do it :-) [16:09] oh yeah I always forget about them [16:10] the interesting bit about this review from just glancing at it is that it's not that much more performant considering it has almost double the clock speed [16:11] than the air [16:11] haswell != perf increase. When they released it haswell asctually tested 3-5% slower in most benchmarks. You just get a really nice battery usage/temp decrease [16:12] sure but he was comparing to air [16:12] apples to apples, as har as haswell is concerned [16:12] far [16:13] oh, doh [16:16] yeah - I mean, there is an improvement, just not the jump I was expecting from such a dramatic clock speed difference [16:17] yea, I wonder if the pro got the pcie upgrade? I mean that number is based off a mix of IO/etc I'm sure. Curious if something else held back the new pro vs the air [16:18] yeah - it'll be interesting to see the 15" because I remember some mumbling about the io pipeline being awesome on the 15s [16:22] benji, fwiw +1 to curtis suggestion that this returm the newest, not puke [16:22] gary_poster: k [16:25] hatch, rick_h_ are one of you already handling https://code.launchpad.net/~ya-bo-ng/juju-gui/linkify-charm-descriptions/+merge/192852 ? [16:25] gary_poster: I am [16:25] tests fail so he is working on that [16:25] cool thanks [16:25] oh k [16:31] jujugui does anyone know if a bug has been filed with core about being unable to deploy the gui in lxc? [16:32] hatch, what's the issue with that? [16:32] apt-get throws an error [16:32] hatch: does it not work? /me did some core stuff this weekend ok in lxc [16:32] hatch, pastebin? [16:32] I'll have to spin up a new instance [16:32] I'll get that going in the bg and report back [16:33] hatch, cool, i'm doing the same to compare notes with [16:33] my core testing box is 12.04 so I'll make sure to do an update first [16:35] hatch: working here in my lxc [16:35] rick_h_: what versions? [16:36] http://paste.mitechie.com/show/1057/ hatch [16:37] hatch: so this is on my desktop still on raring [16:37] cool thanks, just running deploy gui now [16:38] hatch: I had that issue pre-sprints but after updates all was fine this weekend. Figured it was updated [16:38] ahh ok cool maybe the updates fixed it [16:38] I hope so, I need to demo this at a conf next week haha [16:44] rick_h_: looks like it works [16:44] now [16:44] hazmat: ^ [16:44] hatch cool [16:55] hey abentley [16:55] how about tarmac as a config option in our jenkins charms? [16:55] or is that dumb? [16:59] * Makyo steps out [17:01] it would be awesome if there was a `juju ssh-tunnel ` command :) [17:02] hatch, cool [17:02] hazmat: what do you think of that command idea? [17:03] hatch, well.. we can already do juju ssh unit|machine command [17:03] hatch, ideally the tunnel is just done automatically for you where client doesn't have direct access [17:03] hatch, worth filing a bug that juju ssh should tunnel automatically to reach destination [17:03] right, so I have a extra box which I run my juju-core stuff but from my desktop I cannot access the ip of the machines directly so that I can view thing slike the gui without creating a tunnel [17:04] hatch, oh.. you mean construct a tunnel.. sorry [17:04] thought you mean tunnel through a bastion to ssh run a command on unit [17:04] ohh no [17:04] :) [17:05] so now with that in mind....what do you think of the idea? :D [17:05] hatch, so that should already be possible. ala juju ssh juju-gui/0 -L8081:127.0.0.1:443 [17:05] additional params on that line should be passed through [17:05] oh cool [17:05] I didn't know that [17:05] I'll try [17:06] oh I destroyed the env already [17:06] I'll try later I guess :D [17:07] hatch, that's not really what you want though.. although it works, really it sounds like you want a iptables port forward from host to container (ie. avoid ssh tunnel overhead) [17:07] hatch, incidentally that syntax worked for me [17:08] yeah, I know that juju shoudln't be concerned about the firewall stuff [17:08] so thats why I was thinking some kind of convenience method [17:09] but your approach there sounds reasonable [17:23] jcastro: I don't know what that would mean. What would such a config option do? [17:23] "tarmacify this jenkins"? [17:26] jcastro: There's not much integration between jenkins and Tarmac AIUI. It's just a way you can configure a job. So installing the tarmac package and configuring the job would be enough, without having a specific config. [17:26] yeah I was thinking the config option would be the job configuration [17:26] * jcastro shrugs [17:28] jcastro: did you see my msg earlier about the story of getting my ghost charm from github to lp? [17:28] jcastro: It's possible, but that would only give you a single job, and potentially the config would be pretty long to specify all the other job options. [17:28] hatch, no, tell me! [17:28] jcastro: I've thought of providing jobs as subordinate charms. [17:29] jcastro: well what am I supposed to do? just check it out locally and push to lp? or do you have a tool/procedure for this? [17:29] hatch, we don't have anything automated afaict [17:29] jcastro: We're kind of hamstrung by the fact that juju doesn't support complex configuration options (i.e. dicts and lists), and subordinates is a way to work around that. [17:29] I know orange squad is/was going to work on the github workflow? [17:30] abentley, yeah, I was just tossing it out as an idea [17:30] jcastro: ok cool I'll just do it manually then tonight [17:30] jcastro, hatch: jcsackett was working on that, but he's away this week. [17:31] hatch, yeah! [17:31] hatch, let's shoot for "working well enough to get noticed on the upstream forum" [17:31] the only thing it's missing is the http interface because I haven't had enough time to do it yet [17:32] but thanks to your qa we know that the mysql interface works well :) [17:33] bac: benji do you know what the command is to deploy a bundle from the juju command line? [17:33] juju-deployer [17:34] hatch: you need to get lp:juju-deployer or 'pip install juju-deployer'. benji is the pip version up-to-date? [17:34] thanks [17:34] hmm [17:34] ok maybe I'll leave that out of the GUI then [17:35] basically I'm adding a 'deploy' tab to the bundles page [17:35] nah it should be there [17:35] hatch: you might find this useful: models.loadRecentCommits = function(revisions) { [17:35] var commits = []; [17:35] if (revisions) { [17:35] Y.Array.each(revisions, function(commit) { [17:35] commits.push({ [17:35] author: { [17:35] name: commit.authors[0].name, [17:35] email: commit.authors[0].email [17:35] }, [17:35] date: new Date(commit.date), [17:35] message: commit.message, [17:35] revno: commit.revno [17:35] }); [17:35] }); [17:35] deploy by gui, deploy by quickstart, deploy by deployer [17:35] hatch: this: https://docs.google.com/a/canonical.com/document/d/1paE70HEVgKA6LIjngzMa83OYr5jkZrOJ0hVNrYUuva0/edit#heading=h.rmx4dpm78c5i [17:36] ahh cool thanks [17:43] woot! reading source code ftw. I'm a cheating cheater [18:23] bac: sorry, I was at lunch; I don't know if the deployer is on pypi or if it is up to date [18:25] rick_h_: I'll take the review [18:26] hatch: k, doing review comments now [18:26] oh ok cool I'll wait then [18:26] huw goes and removes lines of code and you go adding them all back in :P [18:27] ying/yang :P [18:28] haha [18:32] hatch: ok, have fun with the review [18:32] will do [18:32] jujugui need another review please. I'll bug hatch to QA it. https://codereview.appspot.com/18920045/ [18:33] on it rick_h_ [18:33] gary_poster: ty much [18:46] rick_h_: do you have a second to discuss a different approach to the monkeypatch? [18:46] hatch: sure thing [18:47] https://plus.google.com/hangouts/_/calendar/Z2FyeS5wb3N0ZXJAY2Fub25pY2FsLmNvbQ.j0rk5d371ph8331ijtf48t2uj0?authuser=1 [18:54] hatch: I'm an evil evil man bwuhahahaha [18:54] lol what now [18:54] hatch: no, just that you want to find a better way, but I'm so evil doing things you can't [18:55] haha [18:57] * hatch resists offering 'well I would do it this way' comments [18:57] :D [18:57] hatch: they're legit. I'm open to new ideas as long as I get to ignore the ones I don't like :P [18:58] haha - nah they are dumb [19:11] rick_h_, I made some suggestions but fundamentally LGTM. Thank you! [19:12] gary_poster: cool thanks for the look through [19:12] welcome [19:22] rick_h_: code review done - doing qa now [19:23] oh man I love how fast the sprite generation is now :D [19:24] rick_h_: so is the little dinky + button from design? [19:24] no [Deploy] button? :D [19:27] hatch: yes [19:27] Booo [19:27] hatch: it's straight from them, asked for one during sprints and got this [19:27] ok read my comments and lemme know if you want to discuss [19:27] hatch: yea, figure they'll update once they see it, but figured it'll do until it's on comingsoon [19:29] hatch: ok, you and gary_poster agree on the multiple so I'll back that out. Ugh for subclassing AC, but I can see the point since it's getting complex. It'll be nice because I can name it quicksearch or something as well, though it'll make the diff bigger [19:30] hatch: nothing shuts down pending AC requests. It's a pure 'hide the list of suggestions' move. The tracking/killing AC requests in-flight will have to be a feature of the new widget since I don't think AC does it on its own [19:30] ahh gotcha [19:30] and I'm glad you agree on the subclassing :) [19:30] ok I'll LGTM pending that change then...cool? [19:30] hatch: kinda :P [19:31] haha - well you agree it's the right thing to do....:P [19:31] hatch: I feel that the new widget is a follow up branch? It'll be it's own diff, splitting tests, creating new file, etc [19:31] ok sounds good - just make a card so it doesn't get lost :) [19:31] lgtm'ing [19:31] hatch: yea, I thought about it, but seemed like a lot of work for the 'three lines' duped so avoided it [19:32] lgtm'd [19:33] thanks, will garden up the small stuff. New card added. [19:33] I'm so glad I can deploy the gui in a lxc again, I was worried I woudln't be able to demo it next week [19:34] benji, did #1229179 turn out to be bigger than we had hoped? [19:34] <_mup_> Bug #1229179: Revisionless bundle requests raise ValueError [19:35] gary_poster: not really, finishing writting a comment in the test now and it'll be submitted [19:35] is there documentation for the quickstart? I can't seem to find any on the drive [19:35] awesome benji. [19:36] hatch, no, one sec. [19:37] gary_poster: thanks [19:40] hatch: note that quickstart is in beta when you write docs. Steps are as follows. 1) install quickstart: sudo add-apt-repository ppa:juju-gui/quickstart-beta && sudo apt-get update && sudo apt-get install juju-quickstart [19:41] 2) juju quickstart [the same URL as used by deployer] [19:41] hatch, step 2 currently only works with files, not URLs, but frankban will be adding tomorrow, and should be trivial [19:42] ok awesome - so any updates to the deployer/quickstart we'll have to make sure we update the bundle deploy docs page [19:43] hatch ack. I guess you could say "there is no step 3": you simply wait for the GUI to launch, and after you type in the password, you will see that the GUI shows the bundle being deployed. Also, fwiw: [19:44] hatch, I don't think you need this, but http://pastebin.ubuntu.com/6325901/ [19:44] thanks [19:44] welcome [20:03] rick_h_, does "Reject bundles with cs: urls and charm key in services section" need to be done before we are done with proof work? I don't quite understand what it i means, so if we need to do it, we should make sure I'm following along. :-) [20:14] I feel like this install guide is missing some steps at the start http://bazaar.launchpad.net/~juju-deployers/juju-deployer/trunk/view/head:/README [20:14] jujugui: any reason why we use this date format for charm/bundle change logs: 28/Aug/13 Brad Crittenden Added README and icon - REVNO2 [20:14] as opposed to 2013-08-28 [20:14] because we hate everyone? [20:14] bah [20:15] lol [20:15] gary_poster: bah means what? [20:15] gary_poster: does that mean you'll complain when i fix it? [20:16] I don't know why we do it that way, but as person familiar with computers, I prefer 2013-08-28, but I wonder if 28/Aug/13 is more friendly to others. "bah" means I don't think it is associated with hating everyone. :-P [20:18]