benjehi how we cancel an attachement before save change to s"end only message ?11:51
benjether no way must i post bug about launchpad under launchpad ?11:54
bigjoolsbenje: this is the meeting channel, please ask in #launchpad11:55
benjeok sorry11:55
bigjoolsthat's ok11:55
=== mrevell is now known as mrevell-lunch
=== salgado-afk is now known as salgado
=== salgado is now known as salgado-lunch
=== salgado-lunch is now known as salgado
=== mrevell is now known as mrevell-dinner
=== jt1 is now known as jtv
=== mrevell-dinner is now known as mrevell
RinchenWelcome to this week's Launchpad development meeting. For the next 45 minutes or so, we'll be coordinating Launchpad development.19:00
MootBotMeeting started at 13:05. The chair is Rinchen.19:00
MootBotCommands Available: [TOPIC], [IDEA], [ACTION], [AGREED], [LINK], [VOTE]19:00
Rinchen[TOPIC] Roll Call19:00
MootBotNew Topic:  Roll Call19:00
mrevellwelcome beuno!19:01
Rinchenapologies from diogo19:01
RinchenReleases Team is here19:01
BjornT_bugs team is here19:01
mthaddon(will be seeing diogo later today)19:01
barryEdwinGrubbs: ping19:01
barrymars: ping19:02
barryRinchen: flacoste will not be here19:02
Rinchenthanks barry19:02
Rinchenbugs team here? foundations?19:02
barryleonardr: ping19:02
jtvRinchen: danilos sends apologies19:03
Rinchenthanks jtv - translations here19:03
barryRinchen: seems like a few of our guys are missing19:03
bigjoolsRinchen: cprov @ debconf, al-maisan is MIA19:03
=== thumper_laptop is now known as thumper
Rinchenok, code is here19:03
Rinchensoyuz ok.19:03
Rinchenalrighty then19:04
Rinchen[TOPIC] Agenda19:04
MootBotNew Topic:  Agenda19:04
Rinchen * Next meeting19:04
Rinchen * Actions from last meeting19:04
Rinchen * Oops report (matsubara/ursinha)19:04
Rinchen * Critical Bugs (Rinchen)19:04
Rinchen * Bug tags19:04
Rinchen * Operations report (mthaddon/herb/spm)19:04
Rinchen * DBA report (stub)19:04
Rinchen * Sysadmin requests (Rinchen)19:04
Rinchen * New packages required (salgado)19:04
Rinchen[TOPIC] Next meeting19:04
MootBotNew Topic:  Next meeting19:04
Rinchensame time next week. Anyone not available?19:04
abentleyI am not available.19:05
Rinchenok, thanks abentley19:05
barryRinchen: i may not be, not sure yet19:05
Rinchenbarry, ok, update the mtg agenda please when you know19:05
Rinchen[TOPIC] Actions from last meeting19:05
MootBotNew Topic:  Actions from last meeting19:05
barryRinchen: will do19:05
Rinchenthere were none.19:06
Rinchen[TOPIC] Oops report (matsubara/ursinha)19:06
MootBotNew Topic:  Oops report (matsubara/ursinha)19:06
* Ursinha waves19:06
UrsinhaBugs for today: 252674 and 174480, two timeouts, two for bugs team19:06
UrsinhaBjornT, did you have the time to look at bug 252674?19:06
ubottuBug 252674 on http://launchpad.net/bugs/252674 is private19:06
Ursinhabugs/+index timeout problem19:06
intellectronicaUrsinha: 174480 is in pqm19:06
Ursinhaintellectronica, cool, thanks19:07
BjornT_Ursinha: sorry, no. is it still happening?19:07
UrsinhaBjornT_, yes19:07
BjornT_Ursinha: can you please add some recent oops to the bug report?19:07
UrsinhaBjornT_, ok, i'll append some more19:08
Ursinhathat's all for now, thanks guys19:08
mthaddonUrsinha, how did the oopses look yesterday given the high DB load?19:08
Rinchenthanks Ursinha19:08
Ursinhamthaddon, pretty much the same19:09
mthaddonUrsinha, interesting - ok, thx19:09
Ursinhamthaddon, np19:09
Rinchen[TOPIC] Critical Bugs19:09
MootBotNew Topic:  Critical Bugs19:09
UrsinhaOnly one critical bug not Fix Commited yet, yay!19:09
Ursinha[LINK] https://bugs.edge.launchpad.net/launchpad-bazaar/+bug/25588619:09
Ursinhabut it's now playing on pqm, so soon will be Fix Commited, thanks mwhudson19:09
MootBotLINK received:  https://bugs.edge.launchpad.net/launchpad-bazaar/+bug/25588619:09
ubottuError: This bug is private19:09
ubottuError: This bug is private19:09
Ursinhathat's all19:10
Rinchenthanks Ursinha19:10
Rinchen[TOPIC] Bug tags19:10
MootBotNew Topic:  Bug tags19:10
Rinchenwe have 2 requests19:10
mwhudson(something to mention later, pqm's queue is moving reaaaaally slowly atm)19:10
mthaddonmwhudson, herb is doing a CP19:10
mwhudsonmthaddon: i mean more generally19:10
Rinchenfirst one is for jml for the tag of Stacking19:10
Rinchenthumper, you have this one?19:11
mwhudsonmthaddon: it was only 8th in the queue when i submitted ~20 hours ago...19:11
mthaddonmwhudson, https://devpad.canonical.com/~joey/pqm.html isn't much slower than normal19:11
thumperok, this is to help us identify any bugs to do with the new way we are dealing with branches19:11
Rinchen >> we'll do PQM in a sec19:11
kiko-fudai, I hate my timezone19:12
=== kiko-fud is now known as kiko
RinchenI'm +1 on the stacking tag.  I can see a need for this.19:12
kikothat sounds fine19:12
BjornT_stacking seems a bit too general, though19:12
BjornT_how about branch-stacking?19:13
thumperthat's ok with me19:13
kikosounds good19:13
intellectronicayes, i like that better too19:13
mwhudsonsounds good, as i'm going to make the same complaint about the next proposed tag :)19:13
gmbSo am I19:13
gmbAnd I'm proposing it...19:13
Rinchenoh mwhudson ! Hi, didn't see you19:13
Rinchen[AGREED] branch-stacking tag approved.  mwhudson to update wiki page19:14
MootBotAGREED received:  branch-stacking tag approved.  mwhudson to update wiki page19:14
mwhudsonRinchen: er19:14
mwhudsonRinchen: it was jml who proposed the tag, but ok19:14
Rinchennext up, gmb with the upstream tag.  gmb can you explain how this is different from bugwatch?19:14
gmbSo, the name is far, far too general and needs to be revised.19:14
RinchenI stand corrected mwhudson19:14
gmbSkipping past that...19:14
Rinchenmwhudson, I'll do it then19:15
intellectronicagmb: how about 'upstream-support'?19:16
gmbThis tag would be for organising bugs that relate to our upstream relations work (so maybe upstream-relations would work as a name). The example I've given is of a bug with the upstream report, but I'm intending for it to be used by the community team or anyone who needs us to implement a feature specifically geared towards making our (and by extension Ubuntu's) work with upstream more streamlined.19:16
gmbintellectronica: Yes, that would work.19:16
intellectronicaor 'upstream-requirement'. more accurate but less catchy19:17
mwhudsoncan you give some example bugs?19:17
gmbLet me dig some up. They're all upstream report related ATM.19:17
intellectronicabug #255050 on the wiki19:18
ubottuLaunchpad bug 255050 in launchpad "Statistical columns on the upstream report should turn green for values of 90% and above" [Medium,Fix committed] https://launchpad.net/bugs/25505019:18
gmbbugs 249517 255050 249814 256969 would be examples19:18
ubottuLaunchpad bug 249517 in launchpad "The upstream report should be tracking all open tasks" [High,Fix committed] https://launchpad.net/bugs/24951719:18
kikointellectronica, hmmm, upstream-requirement is a bit too general19:18
Rinchengmb, is this simply for the report or plugins as well19:18
kikogmb, why not 'ubuntu-upstream-relations' or something like that?19:18
intellectronicakiko: well, but this tag is quite general19:18
gmbkiko: That would work well, yes.19:18
kikoI mean, actually say what it is19:19
kikoor upstream-relations-requirement19:19
kikosomething like that19:19
gmbRinchen: I'm intending it to be used organisationally for all ubuntu upstream relations work.19:19
intellectronicai don't like -relations. nobody outside the project will understand what it's about19:19
kikointellectronica, it's okay -- that's why I said ubuntu-relations-requirement19:19
intellectronicayeah, that makes sense19:19
kikoit makes it very obvious that it's a tag to be used by people who know what that means. :)19:19
Rinchenkiko, too long. Something shorter?19:19
kikoRinchen, why?19:19
kikoI think long is fine in this case.19:20
kikoit's a very special tag to be used in a very specific circumstance19:20
abentleykiko: A relations requirement sounds like I have to take it to dinner before I get to have relations with it.19:20
Rinchenso is oops-tools ;-)19:20
BjornT_+1 on ubuntu-upstream-relations19:20
Rinchenok, ok, I'll bend19:20
gmbubuntu-upstream-relations works for me.19:20
Rinchenany objections to the tag or tag name?19:20
Rinchenok then..19:21
Rinchen[AGREED] ubuntu-upstream-relations tag approved. gmb to update the wiki19:21
MootBotAGREED received:  ubuntu-upstream-relations tag approved. gmb to update the wiki19:21
Rinchenand in my best Steve Jobs voice...19:21
Rinchenthere's one more thing19:21
Rinchenkiko and I have talked and we would like to do away with the bug tag section of  the weekly meeting19:22
Rinchenand hold these discussions on the mailing list19:22
Rinchenany objections would require some rework19:22
bigjools+1 meelion19:22
* Rinchen laughs. 19:22
kikoI think that section is so nitpicky even I can't take it much longer19:22
RinchenOk, I get the hint/19:22
kikoand I'm a nitpicker19:22
kikoat heart19:23
Rinchen[AGREED] death to bug tags section. Use the mailing list instead.19:23
MootBotAGREED received:  death to bug tags section. Use the mailing list instead.19:23
Rinchen[TOPIC] Operations report (mthaddon/herb/spm)19:23
MootBotNew Topic:  Operations report (mthaddon/herb/spm)19:23
herbLate Sunday (2008-08-10) Cherry pick of r6818 to the scripts server.19:23
herbEarly yesterday (2008-08-13) Cherry picked r6836 to PPA server and FTP master, and r6842 to lpnet* and edge*19:23
herbA cherry pick for r6846 and r6850 is in queue right now and will be deployed after the meeting.19:23
herbBug 247227 (App servers dying) continues to be an issue. There have been some updates to the bug and we've been capturing debugging information. But the root cause has not yet been isolated.19:23
ubottuBug 247227 on http://launchpad.net/bugs/247227 is private19:23
herbI'm on holiday on Monday.19:23
herbThat's it from Tom, Steve and me unless there are any questions.19:23
RinchenI'm still concerned with 24722719:24
* herb is too19:24
Rinchenwhat else can we do here to get more information?19:24
Rinchenwould an strace help?19:24
Rinchenkiko ^^19:25
mwhudsonkiko's suggestion in the report would help!19:25
stubWe can trigger it. I think we need someone with the right -fu to trace down the fault point rather than boucing between admins and our suggestions.19:25
mwhudsoni am (for my sins) reasonably experienced in debugging python with gdb19:26
mthaddonthx for volunteering!19:26
mwhudsonso if we end up in a gdb prompt when i'm around, please bug me19:26
herbmwhudson: we have a backtrace that was captured yesterday19:26
herbwe don't run gdb interactively19:26
kikoherb, I think running it interactively would really help19:27
kikoherb, and, if an actual engineer or I could do the interaction, it would be MUCH better.19:27
mwhudsoni can give you some more interesting stuff to run non-interactively then :)19:27
mwhudsonlet's talk about it after the meeting19:27
bigjoolsgdb has a remote debugging thingy doesn't it?19:27
herbmwhudson: ok19:27
kikoI realize I am being a heretic at suggesting that an ENGINEER might actually be able to TOUCH the software he wrote19:27
kikobut hey, it's august the 14th19:27
* kiko winks19:27
herbkiko: it's not about the software it's about the servers. :)19:27
stubThink of the servers!19:28
kikoherb, I don't know the difference :)19:28
RinchenThinks of the service. :-D19:28
herbstub: hehe19:28
mthaddonkiko, that's why herb and I have a job19:28
thumperkiko: it is the 15th here19:28
kikomthaddon, well.. unless you can debug yourself, this is really a suboptimal arrangement for this specific crisis.19:28
rockstarthumper, you live in the future.  You are the exception19:29
kikoI hope mwhudson can help further, but I do think that there is no big deal in having engineers able to access the servers in situations that warrant it.19:29
mthaddonkiko, I think we're being pretty accommodating here - I was just meaning we have a job because you don't know the difference between software and servers :)19:29
kikoand by not allowing this, we're all plastered.19:29
Rinchensame team, same side19:29
Rinchenfight nice19:29
mthaddonkiko, I think you're over-estimating that - we're doing everything you ask of us - and thx Rinchen19:30
Rinchenthanks mwhudson for volunteering19:30
* mwhudson goes to find his scripts, which funnily enoiugh are on vostok...19:30
kikomthaddon, I disagree vehemently, but it is not to say I don't appreciate your effort19:30
Rinchen[ACTION] mwhudson to get with mthaddon to see if we can super-charge gdb to give us better diagnostic info on the staging.19:30
MootBotACTION received:  mwhudson to get with mthaddon to see if we can super-charge gdb to give us better diagnostic info on the staging.19:30
Rinchenanything else for the OSAs?19:31
Rinchenok, moving on19:31
Rinchenthanks herb and mthaddon19:31
Rinchen[TOPIC] DBA report (stub)19:32
MootBotNew Topic:  DBA report (stub)19:32
stubWe were getting a high DB load yesterday, apparently from multiple copies of some scripts running. I guess these where old ones before we had decent locking helpers. Load is back to normal now.19:32
stubThe script we need to run to open Intrepid for translations is increasing load enough to push some of our expensive pages (such as the bugs.launchpad.net front page) over the hard limit, causing timeouts.19:32
stubWe have a few options to proceed, including trying again with some tweaked tuning settings (may require bouncing PG), making the code that is being unfriendly to other processes more friendly (but slower), or scheduling some downtime while the unfriendly section of code needs to run - we think half an hour including some slack.19:32
stubMy read-only-launchpad branch is still blocked getting through PQM - no chance to look further today. As far as I have got, it seems that for my branch only in the PQM environment the Librarian reconnection code isn't working, instead causing the Librarian to hang. I'll need to instrament the librarian and run tests or do some manual tests stepping through the librarian code with pdb to get further - yay :-P19:32
stubThere are still a two post-cuttoff reviews to do - one soyuz and one from Barry that isn't in my wiki queue. The soyuz (lucille) one isn't showing up on the merge status cgi - not sure why (pushed to devpad?).19:32
barrystub: crap, i'll put it in the queue right now. apologies19:32
abentleystub: What's the status of my review?19:32
bigjoolsstub: the lucille one is just a permission change, not sure why it's not showing up19:33
stubThat would be the one I didn't see above my queue...19:33
stubbigjools: If it is just permissions, I don't need to see it.19:34
bigjoolsah ok19:34
stubabentley: I do yours too now I see it :)19:34
abentleystub: Thanks.19:34
Rinchenstub, do you need someone to help with the read-only branch?19:35
stubAnd that is my example to kiko on a place we use message_id without storing the message...19:35
kikostub, lucille?19:35
stubI'll yell for help if I need twisted hand holding19:35
kikostub, or.. something else?19:35
kikostub, I've had a similar problem with the librarian before in my local tree19:36
stubkiko: There are soyuz scripts still connecting as the 'lucille' db user.19:36
kikostub, and that's relation to message_id? :)19:36
stubkiko: https://devpad.canonical.com/~jamesh/pending-reviews/abentley/launchpad/better-threading/full-diff19:37
kikostub, oh, abentley's branch. sorry, you really tricked me there!19:37
stubkiko: I just mentioned it in my bug because I was reminded by that branch I was supposed to review.19:37
kikoI like the funky caps in that patch19:37
stubIts 1:40am tomorrow. Its confusing. Deal.19:38
kikostub, hmm. wonder why we don't store that message, actually. thumper?19:38
thumperwhy should we store the whole email?19:38
thumperit serves no purpose19:39
stubIf it is auto generated, there isn't much point.19:39
thumperit will never be looked at19:39
* thumper agrees with stub19:39
Rinchenanything else for stub before we move on?19:40
Rinchenok, thanks stub19:40
gmbintellectronica apologises - he's lost his connection and will be back when possible.19:40
Rinchen[TOPIC] Sysadmin requests (Rinchen)19:40
MootBotNew Topic:  Sysadmin requests (Rinchen)19:40
Rinchenso I have a question19:40
Rinchenfor herb and mthaddon19:41
Rinchennormally I track rt's in this section19:41
Rinchenask for ones that need some attention19:41
RinchenDo you guys feel that you could do this section?19:41
Rinchenas in, we come to the OSAs with RT prioritization19:41
Rinchenand for help in RT scheduling?19:42
mthaddonRinchen, actually, it's very useful to know which RTs you feel need prioritization19:42
Rinchenmthaddon, right. Right now I hope on the IS channel and poke elmo19:42
Rinchenmthaddon, I'm curious if you think the OSAs would be a better choice.19:42
Rinchenso this would be your section and we could poke you19:42
Rinchenvs me poking on the IS channel19:42
Rinchenit has some drawbacks and some positives.19:42
mthaddonRinchen, it's better to not use us as RT proxies, but to work directly with IS - we can certainly help priortise RTs that need it, but otherwise we end up as middle men which isn't efficient for anyone19:43
Rinchenok, thanks mthaddon19:43
RinchenIs anyone blocked on an RT or have any that are becoming urgent?19:43
* beuno would like RT 3142019:43
bigjoolsRinchen: the two I emailed you about last week19:43
Rinchenok beuno I'll have a look at it19:43
beunothanks Rinchen19:43
Rinchenhmm, bigjools. Ok, I'll touch base again with IS on those.19:44
kikoRinchen, is the xlstproc stuff sorted?19:44
kikoour documentation is/was pretty out of date the last time I looked into it19:44
salgadoAFAIK we're generating the docs manually19:44
Rinchennot that I've seen. :-(19:44
salgadook, you know about that already19:45
RinchenI'll go through the priority list again today and make sure our tickets are correctly set19:45
Rinchen[ACTION] Joey to review RT ticket Priorities in RT to ensure they are accurate.19:45
MootBotACTION received:  Joey to review RT ticket Priorities in RT to ensure they are accurate.19:45
Rinchen[TOPIC] New packages required (salgado)19:46
MootBotNew Topic:  New packages required (salgado)19:46
salgadoanything for me?19:46
salgadoha ha19:46
* stub points at the clock19:46
salgadothanks Rinchen.  please move on before kiko remembers of something. ;)19:46
Rinchenok, thanks salgado. I've been pinging folks as RT's come in and ensuring they have submitted bugs for you19:47
Rinchen100% compliance so that was great19:47
Rinchenwhich leads me to one last topic for today that was not on the agenda19:47
kikosalgado, hmm, I just remembered19:47
* Rinchen pauses for kiko 19:47
kikosalgado, I should thank you for the quick updates to meta-lp-deps the past week :)19:47
salgadokiko, you're welcome!19:47
Rinchen[TOPIC] Weekly meeting agenda19:48
MootBotNew Topic:  Weekly meeting agenda19:48
Rinchenkiko and I have chatted briefly19:48
Rinchenand we both agree we'd like to make some progressive changes to the agenda of this meeting19:49
Rinchenwhat I'd like to know at this moment is if there are any suggested topics that you feel can be done on email and not during the meeting?19:49
Rinchenthe topic list again is:19:49
Rinchen * Next meeting19:49
Rinchen * Actions from last meeting19:49
Rinchen * Oops report (matsubara/ursinha)19:49
Rinchen * Critical Bugs (Rinchen)19:49
Rinchen * Bug tags19:49
Rinchen * Operations report (mthaddon/herb/spm)19:49
Rinchen * DBA report (stub)19:49
Rinchen * Sysadmin requests (Rinchen)19:49
Rinchen * New packages required (salgado)19:49
Rinchenbug tags is gone19:49
mwhudsoni guess in a perfect world, the 'report' sections could be done on list and then just talked about in here19:50
RinchenI'm going to combine oops and critical bugs into one section19:50
UrsinhaRinchen, nice, i was going to suggest that19:50
Rinchenmwhudson, good idea19:50
mwhudsonbut that may require unlikely amounts of discipline19:50
Rinchenit is nice though to have the reports during this meeting19:51
sinzuiShouldn't "´╗┐New packages required" be handled elsewhere?19:51
Rinchensinzui, I am inclined to agree19:51
kikosinzui, the only reason we do it here is to remind people19:51
bigjoolsyep, new packages could go offline19:51
mwhudsonyeah, and pasting in the reports to the channel doesn't take long at all19:51
kikoin case they have forgotten19:51
kikothe problem is that they have been forgotten a lot in the past, so we tried to use this to discipline ourselves19:51
RinchenI also think that RT's can be done offline. People can just ping me.19:51
kikoif that need is no longer served, then let's evict it19:51
abentleyDo we have any criteria for things that should be discussed interactively, vs things that should be discussed in a ML?19:52
Rinchenkiko, it's been good this past week but that's the first time I recall it being done consistently.19:52
Rinchenabentley, that's the real crux there.19:52
kikoabentley, things discussed interactively should be relevant to most of the audience here19:52
kikoabentley, so changes in policy, new experiments being run, interesting anecdotes you found during the week, code bloopers and gotchas, etc19:52
kikoabentley, there's a discipline side to this meeting that I generally would like to minimize -- this is the new packages stuff, the OOPS crapola, etc19:53
RinchenIn truth, it's possible to take the entire existing agenda and do it via the mailing list.19:53
Rinchenand keep this meeting for special topics of interest.19:53
Rinchenat least, that is my feeling at the moment19:54
kikowell.. I think we don't need to trash all of it, but I agree we can minimize the filler19:54
abentleykiko: What you're saying sounds to me like a series of presentations-- very different from what we have now, but potentially quite useful.19:54
abentleyAnother option might be to have each team summarize their past week?19:55
Rinchenthumper, BjornT_, jtv?, anyone else with opinions?19:55
kikoabentley, I would like our team to drive the meeting more19:55
kikoabentley, I don't like the idea of status reports as much as I like the idea of anecdotes19:55
kikostatus reports sounds "discipline"19:55
kikowhat I want is "knowledge and experience sharing"19:55
rockstar+1 abentley19:55
barrykiko: maybe each team can drive in round robin, and present interesting things they're working on, or discovered, etc.19:56
marskiko, so a "what's new in reviews" section for the entire team?19:56
rockstarEach team will have some experience to share.19:56
Rinchenit might be nice to have a "good news" item that we do on the TL call19:56
marsRinchen, +119:56
kikomars, yeeeah, something like that19:56
kikomars, "lowlights in code reviews" etc (shocker! but mostly joking)19:57
abentleyThis is a public forum.  Is there a risk of exposing our "trade $ecret$"?19:57
Rinchenok, if you guys have any suggestions for adds and deletes to the agenda, please email the list.  By 1 Sept we'll have a new agenda running.19:58
Rinchen[ACTION] Everyone: Please consider adds and deletes to the agenda to make the weekly higher-bandwidth meetings more useful.19:58
MootBotACTION received:  Everyone: Please consider adds and deletes to the agenda to make the weekly higher-bandwidth meetings more useful.19:58
kikoabentley, I don't worry about those, and for the most part you shouldn't either19:59
RinchenAnything else on this topic before I close?19:59
abentleykiko: Cool.19:59
sinzuiWhat we really risk exposing is our love of profanity19:59
mwhudsonsinzui: we're programmers.  no-one will be surprised20:00
kikoabentley, there are certain topics to be careful with -- security and privacy-involved ones in particular -- but generally speaking it's okay to talk about code, design and the future20:00
MootBotLINK received:  http://www.techtoolblog.com/archives/wtfminute-code-reviews20:00
Rinchennew mootbot feature20:00
kikoauto-detection of LINKs20:00
Rinchenlove it20:00
kikoand you said you hated eggdrop :)20:00
Rinchenthanks guys (and girl) for hanging in with us today20:00
RinchenThank you all for attending this week's Launchpad Developer Meeting. See the channel topic for the location of the logs.20:01
MootBotMeeting finished at 14:05.20:01
abentleyRinchen: Thanks.20:01
mrevellthanks all20:01
gmbThanks Rinchen20:01
mrevellnight all20:01
jtvMootBot: what timezone are you in!?20:01
=== Moot2 is now known as MootBot
=== kiko is now known as kiko-afk
=== salgado is now known as salgado-afk

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