16:59:20 #startmeeting OONI dev gathering 2016-04-11 16:59:20 Meeting started Mon Apr 11 16:59:20 2016 UTC. The chair is hellais. Information about MeetBot at http://wiki.debian.org/MeetBot. 16:59:20 Useful Commands: #action #agreed #help #info #idea #link #topic. 16:59:31 greetings people! 16:59:35 who is here? 16:59:40 hello 16:59:59 here 17:00:00 * sbs here! 17:02:47 apparently with the daylight savings the weekly meeting has the time changed right? 17:03:19 used to be 18:00 CEST 17:04:42 we've redefined 1800 cest to be the time the ooni meeting starts 17:05:04 for a while 17:06:32 and is now changed to 19:00 CEST? 17:06:55 or it was just a DST confusion? 17:07:14 yeah, it's been like this since this time around daylight savings was introduced in europe 17:07:36 since it's pegged to UTC and we didn't discuss changing it, this is what happenned 17:07:56 I am ok with changing it, but we should probably talk about this on the mailing list 17:08:05 anyways, what have you been up to last week? 17:08:25 nothing new here EOF 17:09:27 lots of small MK patches to prepare for release - EOF 17:10:03 What I did last week: * Worked on adding JSON support to MK * Implemented a quick and dirty workaround to resume processing reports in the pipeline * Started rebasing web_connectivity branches for ooni-probe and ooni-backend * Looked into ways of refactoring ooni-probe for instrumenting it via the GUI 17:11:26 mostly travelling last week, started updating the installation of ooni-probe and ooni-backend - EOF 17:15:20 Is there something that you would like to talk about during this gathering? 17:16:50 yes, it would be helpfull if there was a way of invoking ooniprobe without any prior knowledge. that you can say: just do your business without the need to read tons of documentation 17:17:23 flok: this is what i'm doing right now 17:18:06 flok: if you have some spare cycles you can report bugs or issues with the installation process 17:18:07 flok: the thing that comes closest to that currently is the oonideckgen. You can just run that command and it will tell you what is the one line you need to run to make it "do it's business" 17:19:51 ok thanks 17:21:49 exceptions.IOError: [Errno 2] No such file or directory: '/usr/share/ooni/resources/citizenlab-test-lists/test-lists-master/csv/global.csv' 17:22:12 flok how did you try to run ooniprobe? 17:22:42 no I just ran oonideckgen 17:23:02 flok: how did you install it? 17:23:09 apt-get install ooniprobe 17:23:11 debian 17:23:23 version 1.2.2-1 17:24:14 I believe that the 1.2.2 version of the debian package had a bug where it didn't download the resource files on installation. 17:24:35 you can fix that by running sudo ooniresources --update-inputs 17:25:56 1.2.2 though is a very old version. In testing and unstable there is 1.3.1 and 1.3.2 17:26:42 nope, same error 17:27:30 it is doing something though 17:27:40 let mee see what happens 17:28:47 have you seen this: https://iclab.org/explore ? 17:29:50 no, it just sits idle. 17:30:03 waiting for epoll_wait(4, ...) 17:31:24 flok: which command is this? 17:32:50 still the deckgen 17:34:53 flok: can you try updating to 1.3.1? 17:35:13 (or 1.3.2 even better) 17:35:43 anadahz: no I hadn't seen that, seems interesting, but it looks like there are no detected blockpages 17:36:45 anadahz: do you know what they mean by "TTL anomaly" and "Sequence number anomaly"? 17:38:03 hellais: No not really trying to figure out if they have published the source somewhere 17:41:05 hellais: those are probably tests to spot anomalous, possibly spoofed TCP segments via unexpected TTL values or sequence numbers 17:43:33 see for example the related works of this paper: http://www.icir.org/vern/papers/reset-injection.ndss09.pdf 17:44:30 it says that the great firewall could send RST packets with sequence number incremented of 1460 and "TTL that differ from other packets from the purported source address" 17:44:36 sbs: that seems plausible, though I wonder how they get that information given the fact that centinel doesn't have raw packet capabilities :| 17:45:14 hellais: do they capture traces? you can probably get this information by processing the pcap 17:47:06 sbs: yeah it does support packet captures, so they probably do it as part of a postprocessing task 17:47:45 (something recent and somewhat relevant re: tcp injection: http://arxiv.org/pdf/1602.07128v1.pdf) 17:47:59 anyways I guess we should move into next steps 17:48:25 landers: thanks! 17:49:33 landers: looks interesting, thanks! 17:50:23 I will continue work on the web_connectivity branch 17:52:02 I will look into open pull requests, guess will probably reach the point where I could zap the proverbial 1k lines of code in favour of less, easier to manage code for HTTP, continue to develop to the NDT branch, start a branch to completely replace YAML-CPP with JSON - EOF 17:53:43 anadahz: this looks like the code for the iclab explorer: https://github.com/lianke123321/PiResultsWebServer 17:54:30 hellais: ah nice thx! 17:58:26 Next steps: Contine working on ooni-probe and ooni-backend installation/easy updating, Coordinate with Greenhost to set a bigger storage space for hosting ooni measuremtsn reports -EOF 17:59:34 ok great, thanks for attending and until next time 17:59:39 #endmeeting