| bobj | quit | 00:27 |
|---|---|---|
| === Ursinha is now known as Ursinha-afk | ||
| === fta_ is now known as fta | ||
| === dendrobates is now known as dendro-afk | ||
| === fta_ is now known as fta | ||
| === fta_ is now known as fta | ||
| === fta_ is now known as fta | ||
| === dendro-afk is now known as dendrobates | ||
| === fta_ is now known as fta | ||
| === dendrobates is now known as dendro-afk | ||
| === dendro-afk is now known as dendrobates | ||
| === dendrobates is now known as dendro-afk | ||
| === fta_ is now known as fta | ||
| === fta_ is now known as fta | ||
| === fta_ is now known as fta | ||
| === fta_ is now known as fta | ||
| === yofel_ is now known as yofel | ||
| === fta_ is now known as fta | ||
| === fta_ is now known as fta | ||
| === dendro-afk is now known as dendrobates | ||
| === fta_ is now known as fta | ||
| === fta_ is now known as fta | ||
| === dendrobates is now known as dendro-afk | ||
| === fta_ is now known as fta | ||
| === fta_ is now known as fta | ||
| === fta_ is now known as fta | ||
| === lag is now known as Guest5966 | ||
| === Guest5966 is now known as lag_ | ||
| === fta_ is now known as fta | ||
| === oubiwann-away is now known as oubiwann | ||
| === lag_ is now known as lag | ||
| === lag is now known as Guest14059 | ||
| === Guest14059 is now known as _lag | ||
| === Paraselene__ is now known as Paraselene_ | ||
| === fta_ is now known as fta | ||
| === mrevell is now known as mrevell-lunch | ||
| echtler | hi everyone, i've got a problem with my ppa build... | 12:43 |
| echtler | it uses mono/gmcs which fails with the following error: | 12:43 |
| wgrant | echtler: What's the issue? | 12:43 |
| echtler | make[3]: Entering directory `/build/buildd/libtisch-1.1/libs/wrappers/csharp' | 12:44 |
| echtler | g++ -g -O2 -Wall -ggdb -fPIC -DTISCH_SHARED="" -DTIXML_USE_STL -I. -DTISCH_PREFIX='"/usr/share/libtisch/"' -O2 -fno-strict-aliasing -Wall -ggdb -fPIC -L/build/buildd/libtisch-1.1/debian/tmp/lib/ -shared -Wl,-soname,libtisch_csharp.so -I../../../ -I../../..//libs/tools -I../../..//gestured -I../../..//widgets -I../../..//libs/simplecv -I../../..//libs/simplegl -I../../..//gestured/features -L../../..//libs/tools -L../../..//gesture | 12:44 |
| echtler | d -L../../..//widgets -L../../..//libs/simplecv -L../../..//libs/simplegl -ltools -lgestures -lsimplecv -lsimplegl -lwidgets libtisch_wrap.cxx -o libtisch_csharp.so | 12:44 |
| echtler | cd libtisch; gmcs -t:library -out:../libtisch.dll -warn:4 *.cs | 12:44 |
| echtler | ** (/usr/lib/mono/2.0/gmcs.exe:19349): CRITICAL **: _wapi_shm_file_open: shared file [/root/.wapi/shared_data-promethium-Linux-x86_64-328-12-0] open error: Permission denied | 12:44 |
| echtler | ** (/usr/lib/mono/2.0/gmcs.exe:19349): CRITICAL **: _wapi_shm_attach: shared file [/root/.wapi/shared_data-promethium-Linux-x86_64-328-12-0] open error | 12:44 |
| echtler | ** | 12:44 |
| echtler | ERROR:shared.c:401:shm_semaphores_init: assertion failed: (tmp_shared != NULL) | 12:44 |
| echtler | ** (/usr/lib/mono/2.0/gmcs.exe:19349): WARNING **: Thread (nil) may have been prematurely finalized | 12:44 |
| echtler | Stacktrace: | 12:44 |
| bigjools | argh, please use pastebin | 12:44 |
| echtler | ** (/usr/lib/mono/2.0/gmcs.exe:19349): WARNING **: Thread (nil) may have been prematurely finalized | 12:44 |
| echtler | ** (/usr/lib/mono/2.0/gmcs.exe:19349): WARNING **: Thread (nil) may have been prematurely finalized | 12:44 |
| echtler | ** (/usr/lib/mono/2.0/gmcs.exe:19349): WARNING **: Thread (nil) may have been prematurely finalized | 12:44 |
| echtler | Segmentation fault | 12:44 |
| bigjools | echtler: , please use pastebin | 12:44 |
| echtler | uargs. sorry for the spam.... | 12:44 |
| echtler | @wgrant: the problem is that the /root/.wapi directory seems to be corrupt | 12:44 |
| echtler | @bigjools: yes, sorry about that. one moment please.. | 12:45 |
| wgrant | echtler: You can't write to /root. | 12:45 |
| wgrant | The build doesn't run as root. | 12:45 |
| echtler | @wgrant: i know. i don't try that in my makefile. the error seems to be quite common with mono, and the quick fix is to remove the .wapi directory for the user | 12:46 |
| echtler | maybe here the problem is in fact that the build doesn't run as root, but tries to access /root/.wapi | 12:46 |
| echtler | right - here it is in readable form: http://pastebin.org/408118 | 12:47 |
| echtler | sorry about the earlier spam. | 12:47 |
| wgrant | echtler: Your build can't expect to have access to /root. | 12:48 |
| echtler | @wgrant: i am aware of that, and it doesn't assume that. debuild runs it fine as a regular user on my machine, so i'm not sure what causes the attempted access to /root/.wapi | 12:49 |
| === fta_ is now known as fta | ||
| === mrevell-lunch is now known as mrevell | ||
| echtler | wgrant: i've just found that the error also occurs within pbuilder, so it's probably not launchpad-specific. sorry for the noise. | 13:17 |
| === Ursinha-afk is now known as Ursinha | ||
| === fta_ is now known as fta | ||
| echtler | wgrant: for the record, i've found the problem - it's related to mono trying to use the wrong home directory and can be fixed by adding "export MONO_DISABLE_SHM=1" to debian/rules. | 13:32 |
| === fta_ is now known as fta | ||
| ahasenack | launchpad is timing out quite a bit for me today, is it affecting others too? | 14:26 |
| jpds | ahasenack: On edge.lp.net ? | 14:26 |
| ahasenack | jpds: no, production | 14:28 |
| ahasenack | I'm trying to file a bug, https://launchpad.net/landscape/+filebug | 14:28 |
| ahasenack | I even forgot what it was by now :) | 14:28 |
| jpds | ahasenack: Do you have the OOPS ID? | 14:29 |
| ahasenack | 1662G1160 | 14:29 |
| ahasenack | jpds: ^^^ | 14:29 |
| ahasenack | OOPS-1662G1160 even | 14:29 |
| ubot5 | https://lp-oops.canonical.com/oops.py/?oopsid=1662G1160 | 14:29 |
| lifeless | looking | 14:39 |
| lifeless | ok fti search timeout | 14:40 |
| Ursinha | ahasenack, it's timing out after you adding the information? | 14:40 |
| Ursinha | is that correct? so it seems by the oops report | 14:40 |
| lifeless | the performance for this should be the same on both prod and edge | 14:40 |
| lifeless | and prod has a 17 second tiemout | 14:40 |
| === fta_ is now known as fta | ||
| lifeless | ahasenack: I've filed bug 607771 | 14:44 |
| ubot5 | Launchpad bug 607771 in Launchpad Bugs "+filebug FTI query timeouts (affected: 1, heat: 6)" [High,Triaged] https://launchpad.net/bugs/607771 | 14:44 |
| lifeless | deryck: I've filed https://bugs.edge.launchpad.net/malone/+bug/607771 for you - turns out its timing out on prod even though its under the limit on edge | 14:45 |
| ubot5 | Launchpad bug 607771 in Launchpad Bugs "+filebug FTI query timeouts (affected: 1, heat: 6)" [High,Triaged] | 14:45 |
| ahasenack | Ursinha: it was at the last moment, I had content in the bug description | 14:47 |
| ahasenack | Ursinha: I was now able to file it by starting over | 14:47 |
| ahasenack | once in the timeout page, hitting reload just timed out over and over again | 14:48 |
| ahasenack | and the reload did rePOST the data (maybe having a timeout here was expected) | 14:48 |
| lifeless | ahasenack: no, we don't expect timeouts | 14:49 |
| lifeless | we've made a config change last week to reduce a particular timeout value, and some things that were right on the edge are having a bit of trouble | 14:49 |
| ahasenack | ah, ok | 14:49 |
| lifeless | stats show its causing some timeouts, *mainly* to scripts | 14:50 |
| deryck | lifeless, ack. Thanks. I'll take a look here shortly. | 14:53 |
| lifeless | also https://bugs.edge.launchpad.net/malone/+bug/607776 | 14:56 |
| ubot5 | Launchpad bug 607776 in Launchpad Bugs "+filebug-show-similar FTI query timeouts (affected: 1, heat: 6)" [High,Triaged] | 14:56 |
| deryck | gotcha | 14:56 |
| kamal | PPA question: I'm polishing up a wiki doc which claims that the PPA system can build binary packages for 'armel' ... But I thought the builders only build i386 and amd64 since Karmic and I've never known them to build for armel at all. Is it even possible to make the PPA system build for armel? | 15:03 |
| kamal | https://wiki.ubuntu.com/Upstream | 15:04 |
| wgrant | kamal: Normal PPAs are restricted to x86 and AMD64. Some Canonical-controlled PPAs may have armel support. | 15:06 |
| wgrant | There's no good way to virtualise armel machines at the moment, so they can't safely be made available more widely. | 15:06 |
| kamal | wgrant: ok, thanks -- I'm going to remove the reference to armel there. | 15:06 |
| bigjools | and they're dog slow | 15:06 |
| wgrant | kamal: That sounds like a good idea. | 15:07 |
| wgrant | bigjools: That too. | 15:07 |
| * bigjools boggles at the empty PPA queues | 15:07 | |
| wgrant | bigjools: You don't have a couple of Python and archive rebuilds up your sleeve? | 15:08 |
| bigjools | oddly, no | 15:08 |
| === fta_ is now known as fta | ||
| === dendro-afk is now known as dendrobates | ||
| === dendrobates is now known as dendro-afk | ||
| === mtaylor is now known as mtaylor|breakfas | ||
| === dendro-afk is now known as dendrobates | ||
| === fta_ is now known as fta | ||
| === thekorn_ is now known as thekorn | ||
| === deryck is now known as deryck[lunch] | ||
| === oubiwann is now known as oubiwann-away | ||
| hannie | ping danilos | 16:10 |
| hannie | danilos, ping | 16:37 |
| danilos | hannie, sorry, otp | 16:37 |
| hannie | What does otp mean? | 16:38 |
| geser | On The Phone | 16:43 |
| === dendrobates is now known as dendro-afk | ||
| === deryck[lunch] is now known as deryck | ||
| === dendro-afk is now known as dendrobates | ||
| === fta_ is now known as fta | ||
| === beuno is now known as beuno-lunch | ||
| Kasuko | Can anyone help with a failed PPA build. It's a python program and the build log says it failed on "/bin/sh: python2.5: not found | 17:41 |
| Kasuko | but it installs python2.6 | 17:41 |
| === mtaylor|breakfas is now known as mtaylor | ||
| ochosi | hi everyone, i have a question: my packages got rejected because i am not "permitted to upload to the RELEASE pocket in a series in the 'SUPPORTED' state". up to now i never had any probs with pushing packages to my ppa - what does this mean? (https://launchpad.net/~shimmerproject/+archive/ppa) | 17:50 |
| bigjools | ochosi: it means you're trying to upload to Ubuntu, not your PPA | 17:51 |
| ochosi | bigjools: oh damn, my bad... (now i realise the mistake) thanks a bunch! | 17:52 |
| bigjools | np | 17:52 |
| Kasuko | Any ideas why PPA would fail to find python2.5 when it's actually installing python 2.6 | 17:55 |
| bigjools | it's most likely to be a packaging problem | 17:55 |
| Kasuko | I know but I can't figure out what. It's a super simple command line tool and I followed the Python Packaging tutorial to a tee | 17:56 |
| bigjools | your best bet is to ask in #ubuntu-motu for packaging advice | 17:56 |
| Kasuko | ok will do, thank you | 17:57 |
| === beuno-lunch is now known as beuno | ||
| === fta_ is now known as fta | ||
| === dendrobates is now known as dendro-afk | ||
| Kasuko | ok I made a mistake. I uploaded a version ubuntu1 and that failed to build but instead of doing what I should have done and made a version ubuntu2 I requested deletion of ubuntu1 and was going to try to reupload it before I realized how stupid I was. How do I resolve this? | 18:51 |
| maxb | Kasuko: You proceed with uploading an ubuntu2 | 19:01 |
| Kasuko | maxb: but it is saying can't find .orig.tar.gz | 19:01 |
| maxb | Kasuko: Although, preferably, you take this opportunity to question whether you really ought to be using version numbers which _look_ like official ubuntu ones. | 19:01 |
| maxb | Kasuko: for that, you ensure you build your source package using the -sa option to dpkg-buildpackage (or similar) | 19:02 |
| Kasuko | maxb: yes I am being told that in #ubuntu-motu as we speak so I am changing that, my mistake | 19:02 |
| === dendro-afk is now known as dendrobates | ||
| === danilos is now known as daniloff | ||
| === leonardr is now known as leonardr-afk | ||
| === fta_ is now known as fta | ||
| Kasuko | Do I have to wait for all targets to build on an "any" architecture PPA package before I can copy the binaries to a different rtarget? | 19:48 |
| smallfoot- | you motherfucker, fix launchpad now you fkcing assholes | 20:08 |
| smallfoot- | im tired of this ubuntu halfassed piece of shit crap | 20:09 |
| smallfoot- | always bugging out | 20:09 |
| smallfoot- | Timeout error | 20:09 |
| smallfoot- | Sorry, something just went wrong in Launchpad. | 20:09 |
| === epsy is now known as \u03b5 | ||
| smallfoot- | cant even report bugs | 20:09 |
| smallfoot- | fuck this ubuntu shut, im installing microsoft fucking windows 7 bitches! | 20:09 |
| smallfoot- | shit been broken for months and nobody fix it | 20:10 |
| === dendrobates is now known as dendro-afk | ||
| Kasuko | smallfoot-: ... uh? What? | 20:21 |
| purvesh | can some one help me for filling the IRC Channel registration form ? | 20:23 |
| Pici | purvesh: This is #launchpad, if you're trying to register a channel on freenode, then try asking in #freenode | 20:24 |
| smallfoot- | Kasuko, yeah you cant report bugs | 20:25 |
| smallfoot- | cuz it comes TIMEOUT ERROR | 20:25 |
| purvesh | Pici, i had tried at there but still i didnt got any reply, and i want to open irc channel for communicating my Launchpad Group | 20:26 |
| Pici | purvesh: #freenode is still the best place to ask about that | 20:27 |
| deryck | smallfoot-, I know it's a pain. We want to fix it. In the meantime, have you tried using apport to file your bug? | 20:27 |
| === Ursinha is now known as Ursinha-nom | ||
| smallfoot- | deryck, cant use apport in this case | 20:30 |
| deryck | smallfoot-, why not? | 20:30 |
| smallfoot- | deryck, you want to fix it? its been unfixed for months! | 20:30 |
| deryck | smallfoot-, it's hard to fix. | 20:30 |
| === dendro-afk is now known as dendrobates | ||
| smallfoot- | oh | 20:31 |
| deryck | smallfoot-, so why can't you use apport? | 20:32 |
| smallfoot- | cuz it dont have a package to file against | 20:32 |
| deryck | smallfoot-, what is "it" in that sentence? What do you want to file a bug about? | 20:33 |
| smallfoot- | that there is no package 'yasca' in repo | 20:34 |
| deryck | smallfoot-, how did you get yasca on your system? | 20:35 |
| smallfoot- | its not on my system, and that is the problem | 20:35 |
| smallfoot- | i want it in my system, but its not in the repo | 20:36 |
| deryck | ah | 20:36 |
| lifeless | smallfoot-: are you still getting timeouts now? we raised the limit earlier today. | 20:37 |
| lifeless | smallfoot-: what is the url of the page you are on ? | 20:37 |
| smallfoot- | lifeless, yes still timeouts | 20:38 |
| smallfoot- | lifeless, https://bugs.launchpad.net/ubuntu/+filebug?no-redirect&field.tag=needs-packaging | 20:38 |
| lifeless | smallfoot-: what are you putting in the summary field ? | 20:39 |
| lifeless | smallfoot-: hello ? | 20:40 |
| smallfoot- | lifeless, yasca | 20:42 |
| lifeless | indeed, that times out for me too on the duplicate search, I'm trying the actual filing now. | 20:43 |
| lifeless | smallfoot-: https://bugs.edge.launchpad.net/ubuntu/+bug/607937 | 20:43 |
| ubot5 | Launchpad bug 607937 in Ubuntu "yasca (affected: 1, heat: 6)" [Undecided,New] | 20:43 |
| lifeless | smallfoot-: edit to your hearts content | 20:43 |
| smallfoot- | thanks | 20:44 |
| smallfoot- | how you made that? i couldnt do that | 20:44 |
| lifeless | I just tried | 20:44 |
| lifeless | and it worked | 20:45 |
| lifeless | I put 'yasca' in the summary, clicked next. | 20:45 |
| lifeless | it timed out but gave me the form to fill in, so I put in a description and clicked file bug ... | 20:45 |
| smallfoot- | oh, never worked for me, i tried 10000 times | 20:46 |
| smallfoot- | maybe cuz you used edge | 20:46 |
| lifeless | edge has an even lower timeout | 20:46 |
| lifeless | but perhaps | 20:47 |
| lifeless | we're working to reduce the difference between prod and edge. | 20:47 |
| lifeless | to make things like this easier to diagnose and fix | 20:47 |
| smallfoot- | oh | 20:47 |
| smallfoot- | dont konw how you did it first try | 20:47 |
| smallfoot- | i tried 9999 times, just got timeout | 20:48 |
| lifeless | neither do I | 20:48 |
| lifeless | back in a bit | 20:48 |
| smallfoot- | alright | 20:48 |
| smallfoot- | thanks <3 | 20:48 |
| === purvesh is now known as purvesh_ | ||
| === purvesh_ is now known as purvesh | ||
| gmb | lifeless, Usually, when people see an actual timeout and no form when filing bugs, there's a Javascript problem. | 21:05 |
| lifeless | gmb: how does that work | 21:06 |
| gmb | The +filebug page degrades to the non JS version if the JS fails to load. And if it takes a long time for the JS to load you'd see the same effect. | 21:06 |
| lifeless | the js is loaded from apache though, right ? | 21:06 |
| gmb | lifeless, Yes. But on a high latency connection or with a browser plugin that does something to JS... | 21:07 |
| lifeless | ok | 21:07 |
| lifeless | what could I have done better in analyzing it | 21:08 |
| gmb | lifeless, Not much, really, except to ask the user to try another browser. | 21:08 |
| gmb | lifeless, Also, in timeout cases, entering a really simple title for the bug (e.g. 'aaa') and then changing it at the next step is usually sufficient to sidestep the timeout. | 21:09 |
| lifeless | gmb: we don't really have the headroom for that to last more than a couple of days | 21:09 |
| lifeless | :) | 21:09 |
| gmb | But that the bugs dupe search is problematic is understating it by an order of magnitude. | 21:09 |
| lifeless | to the ^2 | 21:10 |
| lifeless | heres another one for your enjoyment | 21:10 |
| lifeless | https://bugs.edge.launchpad.net/malone/+bug/607948 | 21:10 |
| gmb | lifeless, Well, the dupe search times out many, many times every day, I expect. But the users don't notice because the JS version of +filebug handles it gracefully. | 21:10 |
| ubot5 | Launchpad bug 607948 in Launchpad Bugs "timeout on NullBugTask:+index (affected: 1, heat: 6)" [High,Triaged] | 21:10 |
| lifeless | gmb: oh, I think they notice | 21:10 |
| gmb | WHY ARE WE STILL USING NULLBUGTASK. AAAAAAAAAAAAAAA | 21:10 |
| gmb | *ahem* | 21:10 |
| lifeless | they say 'wow that took 20 seconds' | 21:10 |
| gmb | I mean: | 21:10 |
| gmb | "Oh" | 21:10 |
| lifeless | gmb: how should that be fixed. | 21:10 |
| gmb | lifeless, Hand-wavy version that glosses over the fact that we've tried and apparently failed to do it before: we should *always* refer to a bug in a useful context. | 21:11 |
| gmb | lifeless, I can't remember the root cause of the problem off the top of my head. | 21:12 |
| lifeless | gmb: by which you mean, I presume, that when its not in a useful context we should redirect and not render ? | 21:12 |
| gmb | lifeless, Yes, pretty much. I was under the impression that intellectronica had done that work already, though. deryck may know... | 21:12 |
| lifeless | if only he was still here | 21:13 |
| lifeless | we could ask him :P | 21:13 |
| lifeless | https://bugs.edge.launchpad.net/ubuntu/+source/libv4l/+bug/260918 appears to be the actual bug place | 21:13 |
| ubot5 | Error: Could not parse data returned by Launchpad: list.index(x): x not in list (https://launchpad.net/bugs/260918) | 21:13 |
| lifeless | and that too falls down goes boom | 21:13 |
| lifeless | ubot5: har har har | 21:13 |
| deryck | I think intellectronica meant to do the work and other things got in the way. | 21:14 |
| deryck | other work priorities, I mean | 21:14 |
| gmb | Bother. | 21:14 |
| === dendrobates is now known as dendro-afk | ||
| === fta_ is now known as fta | ||
| gmb | lifeless, Of course, the real irony is that ubot5 has given us a useable URL. | 21:15 |
| gmb | Well, useful... except that it times out on edge. | 21:15 |
| lifeless | tada | 21:15 |
| deryck | lifeless, what is the timeout set to on edge/lpnet now? | 21:16 |
| lifeless | gmb: https://lp-oops.canonical.com/oops.py/?oopsid=OOPS-1662EA1548 is the oops for it timing out on edge | 21:16 |
| ubot5 | https://lp-oops.canonical.com/oops.py/?oopsid=1662EA1548 | 21:16 |
| lifeless | deryck: 14/17 | 21:16 |
| lifeless | deryck: we tried 12 but it was a disaster | 21:16 |
| lifeless | gmb: 844 queries before it gets killed | 21:17 |
| deryck | lifeless, right. Was just curious about the current setting. | 21:17 |
| * beuno has not been able to file bugs on edge at all today | 21:17 | |
| lifeless | beuno: its working for me | 21:17 |
| lifeless | beuno: please try | 21:17 |
| deryck | bdmurray considers this a feature | 21:17 |
| lifeless | gmb: 153 repeats of 'is someone subscribed' | 21:18 |
| lifeless | gmb: with limit 1. This seems insane. | 21:18 |
| gmb | lifeless, That's a privacy check, I'd guess, but htat's just a guess. | 21:18 |
| beuno | lifeless, I tried re-submitting it 22 times until I gave up and used firebug to inspect the POST and file it on production | 21:18 |
| lifeless | beuno: you know, there is this 'do not redirect me' button. | 21:18 |
| lifeless | beuno: but separately, really - is it working for you now. Please check. | 21:19 |
| lifeless | beuno: it may time out on the fti for dupes, but you should still get a form. | 21:19 |
| lifeless | and be able to complete the filing | 21:19 |
| lifeless | gmb: I'm sure its something like that, but we need to pull that up out of per-person and make it set based. | 21:20 |
| lifeless | gmb: otherwise it will suck forever. | 21:20 |
| beuno | lifeless, yes, I used that button after 22 times, but I loose my bug description, so I *really* tried | 21:20 |
| beuno | I will try in a bit | 21:20 |
| beuno | now | 21:20 |
| * beuno -> yoga | 21:20 | |
| gmb | lifeless, Couldn't agree more. | 21:20 |
| lifeless | hmm | 21:23 |
| lifeless | 233 requests on a changelog page for library file aliases | 21:23 |
| lifeless | >< | 21:23 |
| lifeless | bigjools is going to get up, and order a hit on me, I suspect | 21:23 |
| geser | wgrant: did you say, soyuz doesn't accept uploads with "Architecture: linux-any"? bdrung managed to get an upload through: https://edge.launchpad.net/ubuntu/+source/acetoneiso/2.2.2-2 => no build records | 21:30 |
| === dendro-afk is now known as dendrobates | ||
| === ubuntu_ is now known as purvesh | ||
| === fta_ is now known as fta | ||
| === leonardr-afk is now known as leonardr | ||
| === flacoste is now known as flacoste_afk | ||
| === verterok_ is now known as verterok | ||
| wgrant | geser: Ah, good point. We don't check for buildable architectures in the primary archive, so it will be accepted there anyway. | 23:20 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!