16:59:32 <hellais> #startmeeting OONI Weekly Gathering 2015-08-03
16:59:32 <MeetBot> Meeting started Mon Aug  3 16:59:32 2015 UTC.  The chair is hellais. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:59:32 <MeetBot> Useful Commands: #action #agreed #help #info #idea #link #topic.
16:59:46 <hellais> long live the meetbot!
17:00:08 <hellais> so who is here? (repeat it for the meetbot)
17:00:11 <poly> hello everyone :D
17:00:18 <poly> hello meetbot
17:02:23 * meejah is idling, if you need anything from txtorcon
17:03:40 * anadahz is here
17:03:56 <hellais> sbs cannot make it, but sent me his report back via email
17:04:17 <hellais> so let's start with report backs
17:04:25 <hellais> I will go with sbs' report back
17:05:11 <hellais> simone has been helping me out with the development of the ADINA registration interface
17:05:26 <hellais> working on the client component of it and submitted a PR for it here: https://github.com/TheTorProject/ooni-adina15/pull/1/files
17:05:50 <hellais> he wrote the code to generate the client side library for speaking to the backend and implemented the basic user registration frontend
17:06:27 <hellais> EOF
17:08:54 <poly> I'll go next
17:09:24 <poly> This week I've worked with sbs to do some code review on network meter and implemented some new features
17:10:05 <poly> sbs has reviewed the argument/run menu pull request and approved it
17:10:22 <poly> (he found a few bugs which we since fixed)
17:10:51 <poly> I've added plugin management to network-meter in the meanwhile, so now NM can fully install and uninstall plugins
17:11:02 <poly> screenshots coming shortly
17:11:03 <poly> EOF
17:11:51 <anadahz> OK my turn
17:12:48 <anadahz> Last week I was working mainly to finish the lepiodpter image that includes obfs4proxy
17:13:43 <anadahz> While working on the install.sh script I tested the fedora installation and add support for arm* architecutes
17:14:26 <anadahz> However I bumped into this error: https://trac.torproject.org/projects/tor/ticket/16713 with PYTHONPATH
17:15:11 <hellais> has anybody in here ever encountered it ^^
17:15:14 <hellais> ?
17:16:03 <anadahz> Apart from thtat I have setup a Travis script to test install tor packages in armel and (hopefully) including weasel's proposed-updates repo
17:16:49 <anadahz> I'm working on that here: https://github.com/TheTorProject/lepidopter/tree/test/travis-build-test-wheezy
17:17:35 <anadahz> It seems that Travis uses very old Ubuntu version 12.04 LTS!!
17:18:09 <anadahz> so i need to work-around a lot..
17:18:51 <poly> (find screenshots of the new NM features here -> http://postimg.org/gallery/2fy9tjxi8/)
17:19:19 <hellais> poly: that is so cool!
17:19:36 <poly> colorful background not included :P
17:22:21 <anadahz> I can't remember of anything else atm
17:22:42 <anadahz> I was hoping to take some feedback on that though..
17:23:08 <anadahz> Building RasPi images is a very lengthy process..
17:23:14 <hellais> anadahz: what sort of workarounds do you need to do?
17:23:49 <anadahz> where virtualization cannot help and even fails
17:24:17 <hellais> anadahz: what made you choose to use travis vs using some VPS?
17:25:05 <anadahz> hellais: the debootstrap version of Travis does not offer support for jessie
17:25:22 <anadahz> meaning that we can't test jessie packages
17:26:24 <anadahz> many of the VPSes fail to do debootstrap and qemu emulation
17:26:40 <hellais> anadahz: does AWS work?
17:27:10 <hellais> did you discover what virtualization system creates issues for qemu?
17:27:32 <anadahz> hellais: not sure but AWS is not free and I'm not willing to rebuild (almost) everything for AWS
17:28:15 <anadahz> hellais: Not really
17:28:19 <hellais> anadahz: what did you build that is travis specific?
17:28:56 <hellais> shouldn't it be possible to just run the same shell scripts on another more performant/more modern machine?
17:28:58 <anadahz> many of the VPSes do not provide info on the type of virtualization
17:29:43 <anadahz> anadahz: Travis build use also qemu
17:30:27 <anadahz> so no running these scripts in any VPS will not necessarilly work
17:30:49 <anadahz> Apart from that they are *very slow!
17:31:39 <anadahz> I setup a local bare metal system but it's not fun to do for every qemu run
17:32:21 <hellais> I would suggest at least trying it on another, faster, VPS or other machine during development to avoid having all the overhead given by travis
17:33:12 <anadahz> Indeed but provision another bare metal system is not so easy
17:33:17 <hellais> if it turns out to be too painful to adapt it from the travis scripts then perhaps it's good to revert back to travis, but I would image that development would be easier if you have a system where you can actually interact with the script you are running and do quick fixes etc.
17:33:27 <hellais> anadahz: do you *need* baremetal?
17:33:54 <anadahz> this is how I run lepidopter build scripts
17:34:13 <anadahz> most of the VPS that I 've tested were not working
17:34:15 <hellais> anadahz: if you do I can give you a spare 24 core machine that I planning on discontinuing, but I can renew it for another month
17:34:29 <anadahz> hellais: discard it
17:35:08 <anadahz> we shold aim for other solutions
17:35:36 <anadahz> anyways the Travis build script will work OK
17:35:47 <anadahz> but my issue is the setup of ooniprobe in armel
17:36:18 <hellais> (fun fact: if you google "travis for testing raspberry pi images", without the quotes, the first hit is lepidopter :P)
17:36:52 <anadahz> :P
17:37:11 <anadahz> EOF
17:37:14 <hellais> ok cool
17:37:17 <hellais> I shall go now
17:37:36 <hellais> so I have been working on the ADINA registration backend
17:37:57 <hellais> last week a lot of progress was made and I feel confident we can deploy something by the end of this week
17:38:48 <hellais> we now have support for creating teams and users, joining teams, listing teams and leaving teams
17:39:10 <hellais> ACL is implemented all around to ensure only users with the proper priviledges can do certain actions
17:39:59 <hellais> today I also started integrating the code written by sbs into ooni-web in this branch: https://github.com/TheTorProject/ooni-web/tree/feature/join_teams
17:40:43 <hellais> EOF
17:41:09 <hellais> so are there some particular topics that we should be discussing together?
17:42:45 <poly> I was wondering if anyone was running MacOS could help us later with testing network-meter
17:43:03 <poly> as for now, all the testing has been done on linux (both me and sbs)
17:43:19 <poly> I have windows VMs to test on in the future
17:43:25 <hellais> poly: I can and am interested in, doing that :)
17:43:46 <hellais> I have OSX 10.10
17:44:07 <poly> we'll be testing sometime in the near future hopefully
17:44:22 <poly> hellais: I'll shoot you an email when we're reading for testing
17:44:27 <poly> thanks in advance
17:44:30 <hellais> poly: ok great.
17:44:53 <hellais> poly: also since 10.10 you can run OSX inside of a virtual machine
17:45:16 <poly> hellais: does apple provide sample images like MS does?
17:45:17 <hellais> without hacky techniques
17:45:24 <hellais> at least that is what I read somewhere on the internet
17:45:51 <hellais> poly: you may need to have a Mac to download it, but yosemite is free to download from the app store
17:47:35 <hellais> anything else?
17:48:39 <hellais> if not I would say we move onto next steps in the interest of keeping this nice and tight as the gunner would say
17:48:53 <hellais> so I will be working mainly on two things:
17:49:12 <hellais> 1) Finishing the ADINA event registration system with the goal of deploying it by the end of this month
17:49:45 <hellais> 2) Looking for instances of blocking of wikipedia inside of the reports we have gathered so far.
17:50:49 <hellais> on the second point if there is somebody interested in learning more about how the data pipeline works and would like to help out in designing the API in preparation for the hackathon to ping me anytime this week
17:51:05 <hellais> EOF
17:54:43 <hellais> anybody else want to share what they will be hacking on this week?
17:57:30 <anadahz> mostly on the install and lepidopter image
17:59:09 <hellais> sounds good
17:59:39 <hellais> so if there are no more burning questions I would say this meeting is adjourned
17:59:45 <hellais> thanks for attending!
18:00:02 <hellais> #endmeeting