/srv/irclogs.ubuntu.com/2013/08/16/#juju-gui.txt

hatchhmph I found a databinding test that was passing by accident00:06
hatchhehe oops00:06
hatchonly 99 more test failures to go!00:12
hatch99 test failures on the wal....99 test failures00:12
hatchyou knock one down, re-run the tests, there's 98 test failures on the wall00:12
hatch9800:21
hatch:)00:21
hatch86 aww yeah00:26
rick_hlol, what did you do to break things so?00:30
hatch70!!00:32
hatchaww man I'm killin it00:32
hatchrick_h: haha too many flybys00:33
hatchI'm not even finished this branch but I have to land it because of the flyby bug fixes/enhancements00:33
hatchyup so many failing tests because of small driveby fixes00:42
hatchbcsaller: still around by chance?00:42
bcsalleryeah00:43
hatchso I think the unit count driveby broke the google analytics code00:43
hatchbecause we no longer use ServiceUnitList correct?00:43
hatchthe test that's failing is test_model.js:15800:44
bcsallerlooking00:44
hatchmy mind is fried I should probably take a break00:45
hatchafter this test!00:45
bcsallerhatch: iirc we didn't remove the global list (I wasn't sure what was using it) so the delta should update both00:45
bcsallerand the update_agg... method does get called, we saw that yesterday00:45
hatchyup but I think that the || 0 is causing the test to fail00:46
bcsallerahh, because it was changed to take the current unit list size so prev never differs from sum00:48
hatchwell sort of00:48
hatchin investigating sum is 0 instead of 100:48
bcsallerI think its an ordering issue, it was using the cached value of unit_count, but taking the size of that unit_list directly depends on the order of the update in the delta code 00:49
hatchjust grabbing the old code to look at it00:49
hatchahh00:51
hatchaggregate00:51
hatchObject {pending: 1}00:51
hatchso yep you're right...00:51
hatchhmm00:51
hatchso we still need a 'total' sum for analytics00:52
hatchick that doesn't feel right00:52
bcsallerwe have total, its the prev value thats getting lost00:52
hatchwell no at the time of the check there are 0 units in the service00:53
hatchbut there is 1 in pending in the aggregate00:53
bcsallerservice.units.size() is total, but the old size of that list is the thing in question00:53
hatchprev_total_units is undefined in the test00:53
bcsallerahh, sounds like the order is opposite from what I thought it would be then00:53
hatchyeah so it should probably be reverted then?00:54
hatchI'm thinking that to properly fix this is going to be more than a quick driveby00:54
bcsallerthat is simplest or move the _gaq code into the handlers module and do it on a serviceunitslist after change 00:54
bcsalleryeah, I'm fine with a rollback00:54
hatchheh gaq00:54
hatchI think that's the proper place to move it but this branch has so many drivebys already00:55
hatchhow about I create a ticket for this change?00:55
hatchhttps://bugs.launchpad.net/juju-gui/+bug/121289700:59
_mup_Bug #1212897: Update service unit aggregate function to use new service unit location  <juju-gui:New> <https://launchpad.net/bugs/1212897>00:59
hatchonly 1 failure now!01:05
hatchAssertionError: expected '2w' to equal '3'01:05
hatchuhh wut?01:05
hatchlol01:05
hatchw00t 0!01:09
hatchnow to make sure the app still works after I fixed all the tests lol01:09
hatch*sigh*01:11
hatchbcsaller: I think there is a bug in your fix01:18
hatchthe prevValue stuff is now reporting the incorrect key sometimes01:18
bcsallerhatch: unexpected or wrong, if you queue two changes to a key in a single window what do you expect to see, the first value or the second?01:19
hatchdo you have a second to look at the code? I can push it up01:21
hatchbzr branch lp:~hatch/juju-gui/scale-up-ui01:22
hatchdatabinding.js:172 the bindingName and the e.changed object sometimes don't match01:22
bcsallerhatch: looking now01:27
bcsallerhatch: you set it on the binding, which is the object describing the relationship between the model and the DOM, it is not a per-change datastructure01:29
bcsallerso the issue in the preVal stuff01:30
bcsallerneed a hand with a fix?01:30
hatchhaha I guess I thought that this was the place to modify the delta01:31
hatchI essentially wanted to augment the detla object with it's previous value01:31
bcsalleryeah, there isn't really an object like that because the updateDOM handles that relationship in a fixed way. Getting the preVal and saying its the first value of a buffered update window to change on the model makes sense to add, but it would have to be managed in the event layer, not the in bindings01:36
hatchso is it looking like adding the prev value stuff now is going to be an issue?01:37
hatchas in not just a few lines?01:37
huwshimiTurns out what I thought I'd broken behaves the same in trunk...01:52
rick_hhuwshimi: yay! / booo!01:53
huwshimirick_h: Yeah, what a waste of time! :)01:54
rick_hhuwshimi: make sure you keep up the current work on the cards so jcsackett and I don't dupe IE work 01:54
huwshimirick_h: Sure01:55
rick_hhuwshimi: thanks01:56
huwshimirick_h: Although apparently we can only have two active cards :(01:56
bcsallerhatch: I think it might still be a few lines, just different lines, we'd keep a mapping like we do with the keys that exists between updateDOM intervals, when we get a change for a key not in the map we record its preVal in that map and then feed that as an additional argument to update calls, then clear it at the end of updateDOM01:56
rick_hhuwshimi: I've been overriding with the exception "Die IE Die"01:56
huwshimirick_h: haha01:57
rogpeppedoes anyone know how to use the new charm browser to list all available charms?08:50
rick_hrogpeppe: hit enter in the search box11:34
rick_hrogpeppe: and if you want *all* uncheck the reviewed/series default checkboxes in the filters11:34
rick_hrogpeppe: https://jujucharms.com/fullscreen/search/?text= is the url that you end up with. 11:35
rogpepperick_h: thanks. i was sure i tried that...11:42
rogpepperick_h: i actually ended up using the charm tool, ("charm list" followed by a fair amount of manual text mangling)12:00
rick_hrogpeppe: k12:05
rogpepperick_h: i'd still like to be able to fetch a definitive list of all the charm urls in the charm store12:06
rogpepperick_h: i've just realised i've probably missed a load, as i only fetched charms in ~charmers or in lp:charm/xxx12:06
rick_hrogpeppe: http://manage.jujucharms.com/api/2/charms?text=12:07
rick_hrogpeppe: you can use the api like the gui does. Sec, let me get the docs12:07
rick_hrogpeppe: http://bazaar.launchpad.net/~juju-jitsu/charmworld/trunk/view/head:/docs/api.rst12:08
rogpepperick_h: awesome! that's just what i want, i think12:08
rogpepperick_h: thanks a lot12:11
rick_hrogpeppe: get what you need? we do have this bug out there, but it sounds like you wanted the urls directly so not sure it would have helped. #120230612:12
_mup_Bug #1202306: We need an "all" category <juju-gui:Triaged by lucapaulina> <https://launchpad.net/bugs/1202306>12:12
rogpepperick_h: yeah, it looks exactly what i need, and a nice regular format, trivial to parse in go.12:13
rick_hrogpeppe: cool, the only ones missing would be bad charms we can't import. Just a disclaimer to the final list12:13
rogpepperick_h: ah, well, i don't mind too much about them.12:14
rogpepperick_h: FWIW, here's my little script that downloads all known charms from the store: http://paste.ubuntu.com/5992556/12:37
rick_hrogpeppe: cool, apis ftw12:38
rogpepperick_h: they weren't all valid12:38
rogpepperick_h: but i successfully got 66312:38
rogpepperick_h: yeah!12:38
sinzuibac, abentley. Do either of you have time to review https://code.launchpad.net/~sinzui/charmworld/modern-py/+merge/18056213:42
abentleysinzui: Sure.13:42
abentleysinzui: r=me.13:47
hatchmorning13:50
rick_hmorning party people13:50
jcsackettmorning all.13:51
hatchyay for IE fixes you guys rock13:52
rick_h:P13:52
rick_hhatch: can you do the second review on huw's this morning please?13:53
hatchI can yep13:53
rick_hI think with that, Makyo's icon fixes, and jcsackett's in progrss branch that closes up the ones we had cards for at least13:53
rick_hjust that drag/drop 'odd' one left in launchpad13:53
hatchrick_h: that one is fixed13:54
hatchI landed a fix for that last week13:54
hatchupdated the ticket13:55
hatchsry13:55
hatchhaving fix committed and in progress the same color makes it hard to scan...13:55
rick_hhatch: ah, gotcha13:56
rick_hoops, my bad jcsackett. I was going through reviews this morning and it was open with the old code13:57
hatchhorrible person allert13:58
jcsackettrick_h: it's up with the right code for review now.13:59
jcsacketthttps://codereview.appspot.com/12760045/13:59
rick_hjcsackett: thanks looking13:59
rick_h"You do not have permission to view this issue"?13:59
hatchthat's what you get for being a horrible person :P14:00
rick_hjcsackett: is there a note on what the span with class absolute is for?14:00
jcsackettrick_h: zuh? that's deleted.14:00
jcsacketti don't see it in the MP when i'm looking.14:00
rick_hjcsackett: looking at the LP review14:00
rick_hhttps://code.launchpad.net/~jcsackett/juju-gui/ac-moves-the-screen-down/+merge/180363 ?14:01
jcsackettrick_h: go to https://codereview.appspot.com/12760045/14:01
rick_hjcsackett: I can't, see error above ^14:01
jcsackettrick_h: nuts to this. i'm deleting it all, will ping when i've got it resubmitted.14:01
jcsacketti have no idea why LP is showing one thing and rietveld is showing another.14:02
rick_hjcsackett: ugh and ugh. Thanks14:02
benjiabentley: thanks for the good review of my branch, it is in the process of merging, but I though you might be interested in the tweaks I made to address the points you raised: http://bazaar.launchpad.net/~benji/charmworld/update-bundle-api/revision/35014:05
abentleybenji: Have you tested that non-dict values are jsonified correctly?  Because lists and strings and bools are instances of object.14:10
hatchjcsackett: I was pretty sure that the autocomplete list box was set to position absolute by default14:10
hatchare we sure we aren't missing any classes?14:10
jcsacketthatch: rick_h and i both looked at this. it is set by default in all browser but IE1014:10
benjiabentley: ooh, that's a good question; I'll add tests to that effect and see what comes14:11
hatchseriously?14:11
hatchugh14:11
hatchthat sounds like a YUI bug14:11
rick_hhatch: so somehting was setting it in the actual element with style="xxx yyy position: relative"14:12
rick_hhatch: I couldn't dupe it in defualt yui examples in IE14:12
rick_hin other browsers there was no edits to the element with a manual style=""14:12
rick_hhatch: so not sure where it was coming from tbh14:12
rick_hhatch: and since chrome/etc aren't doing it I can't set a breakpoint on dom changes 14:12
hatchaww damn that sucks14:13
rick_hhatch: yea, it's hacky, and no idea why it's mis-behaving. but it works :/14:13
hatchI'll take one review/qa on it14:13
hatchrick_h: are you going to land huw's branch?14:14
rick_hhatch: I wasn't going to no. I figured it's not a rush and he can land it as long as he gets the two reviews14:14
hatchehh that won't be until monday14:15
hatchI can do it later14:15
antdillonHi guys, I have a prototype for Luca and he would like it hosted somewhere. How do you host a fork of the juju-gui?14:15
rick_hhatch: but it's friday. I'm not going ot notice it :)14:15
jcsackettrick_h, hatch: the MP is back up. https://codereview.appspot.com/1293204414:15
hatchrick_h: haha - well if I have time I'll land it...we'll see :)14:16
rick_hjcsackett: thanks14:16
lucaMakyo: bcsaller hey, are you guys there?14:17
hatchantdillon: luca if it's just for a short time you could put it on ec214:18
hatchI'm not sure if you have your juju env's set up though14:18
rick_hantdillon: yea, normally we run them on ec2 or something and point everyone at the url14:18
hatchI used git yesterday - I had to look up the equiv for 'bzr info' heh oh boy I'm out of touch14:19
rick_hhatch: lol14:20
hatchI even know how to diff in bzr now like a boss...14:20
hatchI have leveled up14:20
antdillonrick_h, hatch Cool, how would I get it on ec2?14:20
hatchantdillon: hopes and dreams.....hopes and dreams14:20
hatchbut for real....do you have a juju env set up?14:21
hatch:)14:21
antdillonhatch, Ran out of them a long time ago14:21
hatchhaha14:21
antdillonhatch, Locally I do, nothing else where14:21
rick_hantdillon: heh, you can be a docs guinea pig for marcoceppi and company :) https://juju.ubuntu.com/docs/getting-started.html14:21
hatchok so yeah you can follow those docs14:21
hatchor try and find one of us with some time to put it up for you :)14:22
antdillonhatch, I'll trail the docs and see how I get on :)14:22
Makyoluca, hey, sorry, little late.  Here now.14:22
antdillonrick_h, Thanks14:22
hatchantdillon: sure thing - let us know if you run into any issues14:23
hatchyou'll require an ec2 account of course14:23
hatchjcsackett: LGTM QA OK14:24
hatchoh Windows 8.... "Shutdown" "Update and restart" umm I think it's missing an option there14:26
hatchexcept I think shutting down also installs the updates heh14:26
hatchrick_h: at the sprint/ sometime else I'd like a hand fixing my vim/terminal setup, somehow I managed to turn http://vim.spf13.com/ into a white background with white text in vim14:28
hatchlol14:28
rick_hhatch: lol, ok. Sounds good14:29
hatchI'm sure I'm missing something but I just can't figure out where these styles get applied14:29
rick_h:colorscheme14:29
rick_h:colorscheme <tab> to get a lits of ones available14:29
lucaMakyo: can we talk scaling?14:29
rick_hand maybe that can get you unstuck for a moment14:29
Makyoluca, sure.14:29
hatchrick_h: yeah I have to do :color ir_black every time I open it14:30
hatchhttps://github.com/seebi/tmux-colors-solarized I attempted this first14:31
hatchalso with varying degress of success14:31
hatchheh14:31
hatchman I can not type today14:31
rick_hhatch: well look for 'colorscheme' in your ~/.vimrc14:33
rick_hhatch: it should be set in there, maybe twice if running different colors for cmd-line vs gui14:33
hatchI think there is a conflict between the colorschemes and the tmux/zsh themes14:33
lucaMakyo: guichat?14:33
Makyoluca, sure, on my way.14:33
hatchluca: are you talking the sclaeup/down?14:33
hatchI'm working on that right now, mind if I join?14:33
benjiabentley: I added tests for serializing non-mapping types and made them pass; thanks14:43
abentleybenji: Cool.14:44
Makyojujugui call in 1014:50
MakyoKanabaaaaaaaan!14:50
hatchthis....is.....kannnabaaaaaaaan14:51
hatchMakyo: I was really hoping I could drop this upgrade units stuff off a cliff instead of pick up another card :P14:56
Makyohatch, Well, you can defer it for another...three minutes?14:57
* Makyo helpful14:57
hatchlol14:58
Makyojujugui call in 114:59
Makyoabentley, call now15:01
=== schwuk is now known as schwuk_away
=== schwuk_away is now known as schwuk
hatchbcsaller: once I'm done with rick then we can chat15:46
benjiabentley: I need to take a quick break and then do you want to chat about handoff of your current card?15:47
sinzuiadeuring, We have a call scheduled in 75 minutes, but I think Mark R is off today15:47
sinzuiWe can show up, but I suspect we need to reschedule.15:48
adeuringsinzui: ok15:48
abentleybenji: was going to take lunch in a few.  How about 68 minutes from now?15:52
benjiabentley: sounds good15:54
bcsallerhatch: just let me know, I'll  have some coffee in the meantime15:57
hatchbcsaller:  ok in guichat16:05
=== schwuk is now known as schwuk_away
hatchso hows everyones day going?17:48
hatchthat good hey? ;)17:59
hatchyusss https://www.webkit.org/blog/2910/improved-support-for-high-resolution-displays-with-the-srcset-image-attribute/18:02
bachah, typo of the day from an email i just received:  "If I click on the undicode it brings up the glyph"18:23
hatchjujugui could I get two reviews on https://codereview.appspot.com/12784046/ pllllzzzzz18:25
benjihatch: I'll do one18:25
bacsure18:25
hatchthanks guys18:25
abentleybenji: lp:~abentley/charmworld/bundle-heads18:33
benjicool18:34
abentleybenji: It has a bunch of refactoring  because I wanted to be able to test the bundle docs, to ensure they had the both the mongo id and the elasticsearch id.  But I'm not wedded to that approach.18:34
benjiok18:35
hatchhmm something just dinged me but I don't show any messages :/18:37
abentleysinzui: I doubt that we will ever be able to implement  #1086551 since updating a readme means pushing to launchpad.  (Doing it locally would lead to conflicts when the branch's copy of the readme changed.)18:47
_mup_Bug #1086551: Allow inline editing of a charm's README <charmworld:Triaged> <https://launchpad.net/bugs/1086551>18:47
sinzuiabentley, indeed18:48
hatchso funny story jshint now has indentation support and it thinks it should be done differently then gjshint18:53
hatchlol18:53
hatchthe new jshint is crazy fast though18:55
hatcheven if all it's doing now is throwing 15027 errors18:55
sinzuiabentley, I think the bug implies ownership. I think an oauth proc would be required. It would branch, patch, and push in the name of the oauth18:55
sinzuiabentley, I don't see anyone building that when charmers review every charm before reading the webui18:56
Makyojujugui 1 review for trivial: https://codereview.appspot.com/1306904318:56
abentleysinzui: In order for it to do that, it would have to be part of Launchpad, because the only way Launchpad currently supports updating a branch is via SSH.18:56
abentleysinzui: Or we'd need to update Launchpad's SSH to support OAUTH.18:57
hatchjujugui is there a reason why we jshint our json mocks?19:01
hatchI'd like to remove them as json now follows the same config as js19:01
bcsallerI can't think of a good reason19:01
Makyohatch, +1  If the mocks are broken, it will be reflected in the tests.19:02
bachatch: so to qa your branch i'll need to load it up in ec2?19:02
hatchbac: no I can give you a script19:03
hatchone sec19:03
hatchsorry19:03
hatchbac: here is the code that will throttle the add unit creation https://gist.github.com/hatched/118a91ba19be71bde4f219:04
sinzuiabentley, I thought there was a comparable bug for Lp or loggerhead, but I cannot find one. I wanted to dup the issue19:04
hatchit's a hack so you'll have to do it manually19:04
hatchbcsaller: Makyo ok thanks for the sanity check, removing them from JSFILES in the makefile19:05
hatchoh boy this is one heck of a command19:06
hatchhaha19:06
hatchhmm19:06
hatchit 'should' be ignoring them now19:06
hatchI think...19:06
benjiabentley: I'm going away now, but I don't see anything in your branch that raises any questions -- instead I'm sure that'll happen when you're not here to answer them ;)19:07
abentleybenji: okay, have fun.  If you do have questions, sinzui's mental model is close to mine.19:08
benjicool19:08
hatchbenji: lol fyi `e` is the convention in YUI for event object :)19:14
benjiI'll be sure to donate some of our lottery winnings to the YUI project then.19:15
hatchhaha - belive me I'm definitely a fan of the real variables19:15
hatchmaybe it's just habit19:15
hatchso funny thing - after fixing the config and the makefile for the new jshint - the new one actually caught more errors19:17
hatchso my guess is that it would have caught the one nicola found19:17
abentleyorangesquad or bac: Could you please review https://code.launchpad.net/~abentley/charmworld/remove-redudant-id/+merge/180642 ?19:31
sinzuiI can19:31
sinzuiabentley, \o/ I can remove an xxx from by branch when your lands19:34
abentleysinzui: :-)19:34
sinzuiabentley, r=me19:37
abentleysinzui: Thanks.19:38
hatchwow...there were quite a few errors that this version is making me fix19:55
hatchjujugui looking for two reviews on the upgrade jshint branch https://codereview.appspot.com/13072043/ Thanks20:20
jcsacketthatch: you've got oen.20:25
jcsacketts/oen/one/20:25
hatchthanks sir20:26
hatchbac: I agree that we should pluralize that but right now I reaaaaaly want to land this branch so in a follow-up :)20:27
bachatch: a-ok20:27
hatchI'd also like a QA of that jshint upgrade branch......just to be safe :)20:27
hatchbasically make sure it can still 'make' and 'make lint'20:28
hatchanyone else able to take a peek/qa of that jshint branch?20:37
hatchit went up 2 major versions so could use another qa :)20:38
* sinzui thinks elasticsearch tests get slow after repeated test runs20:48
sinzui31s, 44s, 49s, 57s, 79s, 168s20:50
hatchhaha20:54
hatchthat's odd20:54
hatchrick_h: Makyo: I'm going to land huw's changes now20:54
MakyoGood luck.20:55
MakyoYer gonna need it.20:55
hatchlol am I?20:55
MakyoWell, no, probably not.20:55
hatchhaha20:55
MakyoJust trying to branch out from Airplane! to Star Wars.20:55
hatchyou should try anchorman20:56
hatchstay classy san diego20:56
MakyoHah, I have yet to see that, actually.20:56
MakyoJames won't watch it with me, though.20:56
hatchno? is he a hater?20:57
MakyoHahaha20:57
hatcha Will Ferrell hater that is20:57
hatchthey are out there20:57
MakyoI have to be honest, there's only one Will Ferrell movie I've liked so far.20:57
hatchoh so you're one of 'THEM'20:57
MakyoYep :)20:58
hatchlol I pretty much like all of his stuff20:58
MakyoI did like Stranger Than Fiction well enough.20:58
MakyoMostly because that's some of the least typecast work I've seen.20:58
hatch haha20:59
MakyoHe wasn't even against type, just not 100% goofball.20:59
hatchahh yeah true true21:03
hatchdid the autocomplete pushdown bug in IE get landed?21:03
MakyoLooks like.21:29
* Makyo dogwalks21:34
hatchbcsaller: I have a viewlet idea I'd like to bounce off of you, you around?21:41
bcsalleryeah, I'll hop on chat21:41
hatchcool I'm there21:41
hatchof course there is a new IE bug21:59
hatchhttps://bugs.launchpad.net/juju-gui/+bug/121326022:01
_mup_Bug #1213260: DDing a charm renders service icon under sidebar <ie10> <juju-gui:New> <https://launchpad.net/bugs/1213260>22:01
hatchhttps://bugs.launchpad.net/juju-gui/+bug/121326222:04
_mup_Bug #1213262: Text is too large in ghost inspector name field <ie10> <juju-gui:New> <https://launchpad.net/bugs/1213262>22:04
hatchhttps://bugs.launchpad.net/juju-gui/+bug/121326422:06
_mup_Bug #1213264: Overview Inspector has horizontal scroll bar in IE10 <ie10> <juju-gui:New> <https://launchpad.net/bugs/1213264>22:06
hatchnice fresh stream of IE bugs coming in on a Friday :)22:06
MakyoNone of these are so high priority as to freak out about, thankfully.22:07
hatchnope, functionally it's going well22:07
hatchMakyo: although the charm icons still are only 1/4 on trunk22:08
hatchI'm going to assume that hasn't landed yet?22:08
Makyohatch, That fix was in charmworld, not gui.22:08
hatchohh ok so it's still 'landing'22:08
hatchgotcha22:08
Makyohatch, According to sinzui today, once that hits production with other work, a max of 15 minutes until icons are rebuilt.22:09
hatchgreat - we'll have to keep an eye on it next week22:09
hatchhttps://bugs.launchpad.net/juju-gui/+bug/121326722:12
_mup_Bug #1213267: Unit scale up constraints dialogue wraps <ie10> <juju-gui:New> <https://launchpad.net/bugs/1213267>22:12
MakyoUgh, IE :|22:15
hatchlol I know right?22:16
hatchhey no js errors yet22:16
hatchthat's gota be something22:16
hatchlol22:16
hatchIE has this odd thing where it puts an X on the right of an input box to clear it...?22:17
MakyoYeah, for touch stuff.22:21
MakyoI switch out of metro first thing, now.22:21
hatchoh I'm not in metro22:23
hatchI'm in desktop ie22:23
hatchI just updated it though so mayvbe it's a new feature22:25
MakyoOh, yeah, updating now.22:25
MakyoMaybe I'll get them, too.22:25
hatchhttps://bugs.launchpad.net/juju-gui/+bug/1213270 here is the bug if you want to comment22:25
_mup_Bug #1213270: IE10 has native 'Clear Input' X while input is focused in inspector <ie10> <juju-gui:New> <https://launchpad.net/bugs/1213270>22:25
MakyoWill.  Dunno if we can do anything about it, but we'll see.  There are some attributes added to input elements now that might help.22:27
MakyoLike the no-autocomplete and no-spellcheck or whatever.22:27
hatchyeah I was thinking there might be a flag like no autocomplete22:27
hatchhaha22:27

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!