| *** jamesmcarthur has joined #zuul | 00:41 | |
| *** jamesmcarthur has quit IRC | 00:48 | |
| *** jamesmcarthur has joined #zuul | 00:49 | |
| *** jamesmcarthur has quit IRC | 00:54 | |
| *** jamesmcarthur has joined #zuul | 00:55 | |
| *** jamesmcarthur has quit IRC | 01:10 | |
| *** jamesmcarthur has joined #zuul | 01:11 | |
| *** jamesmcarthur has quit IRC | 01:16 | |
| *** jamesmcarthur has joined #zuul | 01:21 | |
| *** swest has quit IRC | 01:23 | |
| *** swest has joined #zuul | 01:38 | |
| *** jamesmcarthur has quit IRC | 02:02 | |
| *** bhavikdbavishi has joined #zuul | 02:46 | |
| *** bhavikdbavishi1 has joined #zuul | 02:53 | |
| *** bhavikdbavishi has quit IRC | 02:54 | |
| *** bhavikdbavishi1 is now known as bhavikdbavishi | 02:54 | |
| *** jamesmcarthur has joined #zuul | 03:45 | |
| *** toabctl has joined #zuul | 04:07 | |
| *** jamesmcarthur has quit IRC | 04:14 | |
| *** swest has quit IRC | 04:24 | |
| *** bolg has joined #zuul | 04:29 | |
| *** jlvillal has joined #zuul | 05:21 | |
| *** swest has joined #zuul | 05:21 | |
| *** pcaruana has joined #zuul | 05:24 | |
| *** pcaruana has quit IRC | 05:30 | |
| *** raukadah is now known as chandankumar | 05:33 | |
| *** jamesmcarthur has joined #zuul | 06:16 | |
| *** jamesmcarthur has quit IRC | 06:20 | |
| *** bhavikdbavishi has quit IRC | 06:22 | |
| *** bhavikdbavishi has joined #zuul | 06:29 | |
| *** pcaruana has joined #zuul | 06:41 | |
| openstackgerrit | Simon Westphahl proposed zuul/zuul master: Add optional support for circular dependencies https://review.opendev.org/685354 | 07:05 |
|---|---|---|
| *** tosky has joined #zuul | 07:11 | |
| openstackgerrit | Simon Westphahl proposed zuul/zuul master: Add optional support for circular dependencies https://review.opendev.org/685354 | 07:19 |
| *** arxcruz|zzz is now known as arxcruz|ruck | 07:25 | |
| *** tosky has quit IRC | 07:26 | |
| *** tosky has joined #zuul | 07:27 | |
| *** jpena|off is now known as jpena | 07:29 | |
| *** mhu has joined #zuul | 07:39 | |
| *** hashar has joined #zuul | 07:57 | |
| *** themroc has joined #zuul | 08:05 | |
| *** themroc has quit IRC | 08:17 | |
| *** themroc has joined #zuul | 08:24 | |
| *** hashar has quit IRC | 08:31 | |
| *** themr0c has joined #zuul | 08:43 | |
| *** themroc has quit IRC | 08:44 | |
| *** gtema_ has joined #zuul | 08:48 | |
| *** themr0c has quit IRC | 08:49 | |
| openstackgerrit | Simon Westphahl proposed zuul/zuul master: Add optional support for circular dependencies https://review.opendev.org/685354 | 08:53 |
| *** jangutter has joined #zuul | 09:13 | |
| *** themroc has joined #zuul | 09:14 | |
| *** smcginnis has quit IRC | 09:14 | |
| openstackgerrit | Merged zuul/zuul-jobs master: Use iptables for openSUSE https://review.opendev.org/685491 | 09:21 |
| *** zbr is now known as zbr|ruck | 09:24 | |
| *** tobiash has quit IRC | 09:27 | |
| *** tobiash has joined #zuul | 09:29 | |
| *** gtema_ has quit IRC | 10:14 | |
| *** tosky has quit IRC | 10:17 | |
| *** tosky has joined #zuul | 10:18 | |
| openstackgerrit | Simon Westphahl proposed zuul/zuul master: Add optional support for circular dependencies https://review.opendev.org/685354 | 10:32 |
| *** bolg has quit IRC | 10:32 | |
| *** bolg has joined #zuul | 10:34 | |
| *** hashar has joined #zuul | 10:35 | |
| *** bolg has quit IRC | 10:38 | |
| *** bolg has joined #zuul | 10:38 | |
| *** bhavikdbavishi has quit IRC | 10:46 | |
| openstackgerrit | Fabien Boucher proposed zuul/zuul master: WIP - gitlab - bootstrap the driver structure + Webhook support https://review.opendev.org/685682 | 11:10 |
| *** chandankumar is now known as raukadah | 11:11 | |
| *** jpena is now known as jpena|lunch | 11:31 | |
| *** wxy-xiyuan has quit IRC | 11:37 | |
| *** ianychoi has quit IRC | 11:39 | |
| *** hashar has quit IRC | 11:39 | |
| *** ianychoi has joined #zuul | 11:40 | |
| *** bhavikdbavishi has joined #zuul | 11:53 | |
| *** hashar has joined #zuul | 11:55 | |
| *** bhavikdbavishi1 has joined #zuul | 11:56 | |
| *** bhavikdbavishi has quit IRC | 11:57 | |
| *** bhavikdbavishi1 is now known as bhavikdbavishi | 11:58 | |
| *** mariowood has joined #zuul | 12:07 | |
| *** jamesmcarthur has joined #zuul | 12:26 | |
| *** jpena|lunch is now known as jpena | 12:30 | |
| *** rfolco has joined #zuul | 12:33 | |
| *** jamesmcarthur has quit IRC | 12:35 | |
| *** bolg has quit IRC | 13:09 | |
| *** ironfoot has joined #zuul | 13:34 | |
| *** bolg has joined #zuul | 13:34 | |
| pabelanger | mordred: is your gerrit talk about zuul online some place? | 13:44 |
| *** jangutter has quit IRC | 13:53 | |
| *** mariowood has quit IRC | 13:54 | |
| *** brendangalloway has joined #zuul | 13:56 | |
| Shrews | pabelanger: i think he keeps all of his talks on inaugust.com | 13:57 |
| Shrews | zuul-maint: Umm, is anyone else seeing strange things on the zuul-ci.org website? Both FF and Safari are presenting the top header of the page oddly | 14:00 |
| clarkb | Shrews: can confirm. I think that implies broken css | 14:01 |
| Shrews | wheee | 14:01 |
| Shrews | guessing https://review.opendev.org/680465 | 14:02 |
| clarkb | https://storage.bhs1.cloud.ovh.net/v1/AUTH_dcaab5e32b234d56b626f72581e3644c/zuul_opendev_logs_e02/680465/4/gate/zuul-website-build/e026ca8/html/ seems to do the same (the build for that change) | 14:03 |
| clarkb | if you push a revert we can check the build to see if it fixes | 14:03 |
| Shrews | yeah, just looked at that too | 14:04 |
| Shrews | revert coming | 14:04 |
| openstackgerrit | David Shrewsbury proposed zuul/zuul-website master: Revert "CSS fix for ul/li in FAQ" https://review.opendev.org/685717 | 14:04 |
| *** jangutter has joined #zuul | 14:08 | |
| Shrews | clarkb: seems to fix it: https://openstack.fortnebula.com:13808/v1/AUTH_e8fd161dc34c421a979a9e6421f823e9/zuul_opendev_logs_875/685717/1/check/zuul-website-build/8754ced/html/ | 14:11 |
| clarkb | Shrews: fast approved | 14:13 |
| openstackgerrit | Merged zuul/zuul-website master: Revert "CSS fix for ul/li in FAQ" https://review.opendev.org/685717 | 14:17 |
| *** pcaruana has quit IRC | 14:18 | |
| Shrews | zuul-ci.org seems back to normal now | 14:27 |
| *** themroc has quit IRC | 14:37 | |
| *** hashar has quit IRC | 14:39 | |
| fungi | thanks for noticing/fixing Shrews! | 14:39 |
| openstackgerrit | Tristan Cacqueray proposed zuul/zuul master: Replace tox-py36 by tox-py37 https://review.opendev.org/682158 | 15:03 |
| clarkb | tristanC: why replace? | 15:04 |
| clarkb | centos and rhel 7/8 have 3.6. Seems like that is an important version to test? | 15:05 |
| corvus | clarkb: oh i was thinking we only needed to test oldest/newest | 15:06 |
| clarkb | maybe? certainly oldest and newest should avoid adding in syntax that won't work in the middle | 15:07 |
| tristanC | clarkb: i think it's fine to not test 3.6, i've never seen something that works in 3.5 and 3.7 to not work in 3.6 | 15:07 |
| corvus | tristanC said he hasn't seen anything caught by 3.6 that isn't caught by 3.7 (which would be my expectation too) | 15:07 |
| clarkb | fwiw I use 3.7 myself and haven't had any problems going that direction at least. So ya probably is fine | 15:11 |
| *** armstrongs has joined #zuul | 15:12 | |
| *** armstrongs has quit IRC | 15:16 | |
| openstackgerrit | Merged zuul/zuul master: Support nodes setting 'auto' python-path https://review.opendev.org/682275 | 15:18 |
| *** bolg has quit IRC | 15:23 | |
| openstackgerrit | Merged zuul/zuul master: Annotate cancelJobInQueue with build id https://review.opendev.org/684373 | 15:25 |
| *** mattw4 has joined #zuul | 15:33 | |
| *** jamesmcarthur has joined #zuul | 15:41 | |
| *** pcaruana has joined #zuul | 15:46 | |
| *** brendangalloway has quit IRC | 16:30 | |
| SpamapS | The difference between 3.6 and 3.7 is all additional features. I'd definitely trust that 3.7 testing covered 3.6. The only corner cases would be things like that old bug we found forever ago in the Ubuntu python packages... it's nice to test on a real build of the interpreter. | 16:33 |
| mordred | our docker images are also built on top of latest python base image, which is 3.7.3 - and based on upstream python not distro python | 16:34 |
| openstackgerrit | Fabien Boucher proposed zuul/zuul master: Gitlab - bootstrap the driver structure + Webhook support https://review.opendev.org/685682 | 16:35 |
| clarkb | that doesn't help everyone deploying on bionic, centos and rhel 7/8 though | 16:35 |
| mordred | so I'm also on board with doing 35 + 37 | 16:35 |
| clarkb | I think it is enough of a corner case to not worry about it until it becomes a problem though | 16:35 |
| mordred | clarkb: yah | 16:35 |
| mordred | I think we'd want to replace 3.5 with 3.6 at whatever point we decide to stop supporting 3.5 | 16:36 |
| mordred | but while we have 3.5 support - just testing 5 and 7 seems sufficient | 16:36 |
| Shrews | jamesmcarthur: hi! not sure if you saw, but we had to revert your css change this morning: https://review.opendev.org/685717 | 16:37 |
| mordred | tristanC: I think you need a different base image for that 3.7 patch to work | 16:37 |
| clarkb | mordred: bionic actually has 3.7 packages but you have to install them more explicitly | 16:41 |
| openstackgerrit | Tristan Cacqueray proposed zuul/zuul master: Replace tox-py36 by tox-py37 https://review.opendev.org/682158 | 16:41 |
| jamesmcarthur | Shrews: I did. My apologies. There is a missing } | 16:43 |
| jamesmcarthur | I'll update the patch after I'm done with my current sprint of meetings | 16:43 |
| *** jpena is now known as jpena|off | 16:45 | |
| fungi | clarkb: bionic has packages for 3.7, so in theory users could deploy with it there fairly easily | 16:48 |
| tristanC | i picked fedora-latest as python3.7 is the default there | 16:49 |
| clarkb | fungi: that is true. | 16:49 |
| clarkb | mostly centos and rhel 7/8 then | 16:49 |
| *** zbr|ruck is now known as zbr | 16:52 | |
| *** igordc has joined #zuul | 17:08 | |
| pabelanger | we are doing python36 for zuul.a.c, so would be nice to keep testing on that | 17:10 |
| pabelanger | but so far, haven't really seen python36 specific issues | 17:11 |
| Shrews | Does anyone plan to do any sort of writeup about zuul related things at last week's AnsibleFest? | 17:27 |
| clarkb | in the last 7 days that ansible iptables var not being defined issue in persist-firewall role has hit us 208 times according to logstash. This is definitely not happening on every job which makes it extra weird. Also we retry these jobs because the failure happens in pre-run so not end of the world but it would be good to track that down | 17:28 |
| clarkb | pabelanger: ^ I think you maybe had ideas to try? | 17:28 |
| clarkb | pabelanger: it happens on different clouds and on different distros and in different jobs | 17:28 |
| pabelanger | clarkb: was that the -13 return code? | 17:30 |
| clarkb | pabelanger: ya rc -13 of the forked process not of ansible itself 2019-09-30 14:37:25,375 DEBUG zuul.AnsibleJob.output: [e: 53c73870b5b343b6a6c883a388b0070d] [build: bafbb86f7b924f35a595edd2d4aa7196] Ansible output: b'ok: [primary] => {"changed": false, "failed_when_result": false, "module_stderr": "", "module_stdout": "", "msg": "MODULE FAILURE\\nSee stdout/stderr for the exact error", "rc": -13}' | 17:32 |
| clarkb | ansible exits 2 | 17:32 |
| pabelanger | yah, let me look at my notes again. I think I had an idea what it was, but post conference brain is hitting me | 17:32 |
| clarkb | and it only happens on the second run on the persistent-firewals role | 17:32 |
| pabelanger | right, i think something to do with including same role more then once might be the key | 17:33 |
| fungi | Shrews: i have some notes i still need to digest but could probably integrate into an ml post. it was an interesting experience, exciting to talk to so many unknown current and potential zuul users, and also interesting to contrast ansible community development against our own | 17:33 |
| clarkb | pabelanger: perhaps we can use import_role so that the role is less dynamic? | 17:35 |
| pabelanger | yah, I also think the usage of our variables might be related. eg: them crossing playbook bounds, over passing into role | 17:36 |
| pabelanger | I haven't testing import_role much, but worth a try | 17:36 |
| clarkb | other ideas: we can use shell instead of comamnd. maybe iptables-save needs to run in a more fully fledged environment? | 17:36 |
| fungi | Shrews: i was particularly surprised that the community is told at the contributor summit what the secretly designed development roadmap holds for them in the future | 17:36 |
| clarkb | pabelanger: we don't pass anything into the role | 17:36 |
| clarkb | pabelanger: the role takes the current iptables ruleset and persists it to disk | 17:36 |
| pabelanger | clarkb: if we could figure out why SIGPIPE would be called, that would be helpful | 17:36 |
| clarkb | there is no input | 17:36 |
| pabelanger | clarkb: need to look again, but I think we register some variables in the top level playbook, then the role references them | 17:37 |
| Shrews | fungi: i'm sure a ML post about your thoughts/experiences would be well received :) | 17:37 |
| clarkb | pabelanger: not that I see. The role appears to be fully self contained | 17:38 |
| clarkb | pabelanger: I'm going to push up a change to use import_role to see if it works at all | 17:38 |
| pabelanger | ack | 17:39 |
| openstackgerrit | Clark Boylan proposed zuul/zuul-jobs master: Use import_role on persistent-firewall https://review.opendev.org/685765 | 17:40 |
| clarkb | pabelanger: filename:"job-output.txt" AND message:"The task includes an option with an undefined variable." is my logstash query if you want to see more hits too | 17:45 |
| clarkb | import_role seems to work in testing https://b3b0b103e21a776f5487-60046e015843f13b22f3609583ac17fc.ssl.cf1.rackcdn.com/685765/1/check/zuul-jobs-test-multinode-roles-centos-7/85017e7/job-output.txt 'List current ipv4 rules' is executed three times in a single play there | 17:55 |
| clarkb | at the very least it doesn't seem to regress | 17:55 |
| clarkb | maybe we can merge that then check logstash to see if the behavior changes any, and if it doesn't then replace command with shell next? | 17:55 |
| openstackgerrit | Tristan Cacqueray proposed zuul/zuul master: WIP: Replace tox-py36 by tox-py37 https://review.opendev.org/682158 | 18:19 |
| *** bhavikdbavishi has quit IRC | 18:33 | |
| *** hashar has joined #zuul | 18:36 | |
| *** jamesmcarthur has quit IRC | 19:01 | |
| *** jamesmcarthur has joined #zuul | 19:06 | |
| openstackgerrit | Tristan Cacqueray proposed zuul/zuul master: WIP: Replace tox-py36 by tox-py37 https://review.opendev.org/682158 | 19:21 |
| *** jamesmcarthur has quit IRC | 19:34 | |
| clarkb | any other zuul jobs reviewers want to review https://review.opendev.org/#/c/685765/ we can track that with logstash to see if it helps once it lands | 19:35 |
| * clarkb finds lunch | 19:35 | |
| *** jamesmcarthur has joined #zuul | 19:41 | |
| openstackgerrit | Tristan Cacqueray proposed zuul/zuul master: WIP: Replace tox-py36 by tox-py37 https://review.opendev.org/682158 | 19:51 |
| openstackgerrit | Tristan Cacqueray proposed zuul/zuul master: WIP: Replace tox-py36 by tox-py37 https://review.opendev.org/682158 | 20:04 |
| openstackgerrit | Clark Boylan proposed zuul/zuul-jobs master: Escape special filename characters in urls https://review.opendev.org/685785 | 20:27 |
| openstackgerrit | Tristan Cacqueray proposed zuul/zuul master: WIP: Replace tox-py36 by tox-py37 https://review.opendev.org/682158 | 20:27 |
| openstackgerrit | James E. Blair proposed zuul/zuul-registry master: Initial implementation https://review.opendev.org/683958 | 20:36 |
| corvus | clarkb, tristanC: ^ i think that's the complete first stage -- it implements an authenticated+ssl registry with local and swift storage backends. that's enough to run an intermediate registry. next stage is building on that to add the pass-through support to make it a buildset registry. | 20:38 |
| corvus | but before we do that, i think we can add tests, docker builds, and put it into production as the opendev intermediate registry. hopefully that will fix our zero-byte upload issue (or at least make it possible to debug). | 20:39 |
| corvus | Shrews: i just noticed your comments on the previous ps of that; i'll take a look now. | 20:39 |
| corvus | Shrews: ayup. i'll take care of adding docstrings in a followup, and squash it if i finish it before there are further reviews. | 20:43 |
| pabelanger | cool, was just looking into docker registery stuff too! So may also be able to give some feedback of zuul-registry soon if needed too | 20:48 |
| tristanC | corvus: that LGTM, the root dispatcher is very instructing | 20:48 |
| *** pcaruana has quit IRC | 20:50 | |
| Shrews | corvus: the biggest issue for me was the lack of an abstract protocol class for the storage stuff | 20:51 |
| *** irclogbot_2 has quit IRC | 20:54 | |
| *** irclogbot_0 has joined #zuul | 20:57 | |
| *** irclogbot_0 has quit IRC | 21:01 | |
| *** irclogbot_3 has joined #zuul | 21:03 | |
| clarkb | I've pushed https://review.opendev.org/685792 to track the perstistent-firewall issue with e-r | 21:05 |
| *** irclogbot_3 has quit IRC | 21:07 | |
| *** irclogbot_2 has joined #zuul | 21:09 | |
| *** jamesmcarthur has quit IRC | 21:11 | |
| openstackgerrit | Clark Boylan proposed zuul/zuul-jobs master: Escape special filename characters in urls https://review.opendev.org/685785 | 21:12 |
| *** irclogbot_2 has quit IRC | 21:13 | |
| *** jamesmcarthur has joined #zuul | 21:17 | |
| corvus | i caught a couple cleanups i missed, so i'll definitely do this as a new revision | 21:21 |
| *** irclogbot_0 has joined #zuul | 21:22 | |
| openstackgerrit | Tristan Cacqueray proposed zuul/zuul master: Replace tox-py36 by tox-py37 https://review.opendev.org/682158 | 21:22 |
| *** irclogbot_0 has quit IRC | 21:25 | |
| openstackgerrit | James E. Blair proposed zuul/zuul-registry master: Initial implementation https://review.opendev.org/683958 | 21:25 |
| corvus | tristanC, clarkb, Shrews: ^ for realz this time :) | 21:26 |
| *** irclogbot_0 has joined #zuul | 21:27 | |
| *** irclogbot_0 has quit IRC | 21:33 | |
| openstackgerrit | Jeremy Stanley proposed zuul/zuul-website master: Link to instructions on vulnerability reporting https://review.opendev.org/685799 | 21:35 |
| clarkb | fungi: (and others) https://review.opendev.org/#/c/685785/ passes tests now and should fix the problem with :'s in filenames | 21:36 |
| fungi | thanks! | 21:37 |
| *** irclogbot_0 has joined #zuul | 21:37 | |
| clarkb | sr | 21:40 |
| clarkb | derp | 21:40 |
| *** irclogbot_0 has quit IRC | 21:41 | |
| corvus | i'm working on the docker image for the registry now | 21:45 |
| *** irclogbot_3 has joined #zuul | 21:50 | |
| openstackgerrit | Merged zuul/zuul-jobs master: Escape special filename characters in urls https://review.opendev.org/685785 | 21:52 |
| *** irclogbot_3 has quit IRC | 21:53 | |
| *** irclogbot_3 has joined #zuul | 21:55 | |
| *** irclogbot_3 has quit IRC | 21:59 | |
| openstackgerrit | James E. Blair proposed zuul/zuul-registry master: Fix container image build https://review.opendev.org/685808 | 21:59 |
| *** irclogbot_0 has joined #zuul | 22:01 | |
| *** irclogbot_0 has quit IRC | 22:05 | |
| *** irclogbot_0 has joined #zuul | 22:07 | |
| *** irclogbot_0 has quit IRC | 22:15 | |
| *** irclogbot_0 has joined #zuul | 22:17 | |
| *** tosky has quit IRC | 22:24 | |
| clarkb | http://status.openstack.org/elastic-recheck/index.html#1846093 is now tracking the thing that https://review.opendev.org/#/c/685765/ aims to influence | 22:33 |
| clarkb | I'm able to watch that go in if I can get reviews | 22:33 |
| corvus | clarkb: do you have a link to something where i can read about the reasoning there? (it seems a little voodoo to me) | 22:34 |
| corvus | clarkb: (also, why do we think ansible telling us that "iptables returned -13" means "ansible broke" rather than "iptables broke"?) | 22:35 |
| clarkb | corvus: fwiw its a lot of voodoo to me. Basically pabelanger and I noticed that it only happens on the second invocation of that role. Its always fine on the first run. We've noticed problems with recurring and nested includes in the past so this is a shot in the dark | 22:36 |
| clarkb | I'm trying to find the IRC log from before ansiblefest | 22:37 |
| corvus | i'm reading https://docs.ansible.com/ansible/latest/user_guide/playbooks_reuse.html#dynamic-vs-static | 22:38 |
| clarkb | corvus: http://eavesdrop.openstack.org/irclogs/%23zuul/%23zuul.2019-09-18.log.html#t2019-09-18T17:23:38 that was pabelangers analysis from a couple weeks ago | 22:39 |
| corvus | clarkb: ok. i'm doubtful that this is the problem (based on the documentation, there should be no difference here), and all things being equal, i think we'd want to default to include_role as a habit, so we're not debugging statically assigned role arguments. but i also value your hunch that this may be an ansible bug, and making this change would help confirm or exclude that. so i'm +2 on this for that | 22:42 |
| corvus | purpose, but can we switch this back if we are able to exclude this theory? | 22:42 |
| clarkb | corvus: ya that is why I got the e-r bug up too. If this doesn't change the occurence of this bug we should revert | 22:43 |
| corvus | clarkb: sounds good, +3 | 22:43 |
| clarkb | corvus: comment on https://review.opendev.org/#/c/685808/1 | 22:43 |
| corvus | clarkb: and thanks for the catch-up :) | 22:44 |
| clarkb | I'm working through the registry change now | 22:51 |
| fungi | the lambdas of lambdas are a bit mind-bending | 22:56 |
| openstackgerrit | Merged zuul/zuul-jobs master: Use import_role on persistent-firewall https://review.opendev.org/685765 | 23:17 |
| *** mattw4 has quit IRC | 23:26 | |
| clarkb | corvus: comments posted on the registry change | 23:39 |
| openstackgerrit | James E. Blair proposed zuul/zuul-registry master: Fix container image build https://review.opendev.org/685808 | 23:40 |
| corvus | okay, that should have the container build working, and it also adds a functional test | 23:40 |
| *** hashar has quit IRC | 23:40 | |
| *** armstrongs has joined #zuul | 23:49 | |
| corvus | clarkb: thx, responded -- some need to be fixed so i -1d till i get to that tomorrow | 23:55 |
| *** jamesmcarthur has quit IRC | 23:57 | |
| *** jamesmcarthur has joined #zuul | 23:58 | |
| *** armstrongs has quit IRC | 23:59 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!