15:58:06 <onyinyang[m]> #startmeeting tor anti-censorship meeting
15:58:06 <MeetBot> Meeting started Thu Aug  3 15:58:06 2023 UTC.  The chair is onyinyang[m]. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:58:06 <MeetBot> Useful Commands: #action #agreed #help #info #idea #link #topic.
15:58:49 <onyinyang[m]> hello everyone!
15:58:51 <onyinyang[m]> here is our meeting pad: https://pad.riseup.net/p/tor-anti-censorship-keep
15:58:55 <meskio> hello
15:59:16 <ggus> hello o/
16:03:16 <onyinyang[m]> ok let's start
16:03:36 <onyinyang[m]> Does anyone want to address any points from the discussion from last week?
16:04:08 <meskio> shell is not around today, I guess he didn't have the time to reproduce it before going AFK
16:04:08 <onyinyang[m]> if not, let's move on to the discussion items for this week
16:04:29 <meskio> neither I heard back from guardian project, I'll bring this topic in our next in voice meeting
16:04:35 <meskio> EOF
16:04:55 <onyinyang[m]> Ok, so let's continue these discussions at a later time and move on to this week.
16:05:08 <onyinyang[m]> The first item is on the ptspec status version support
16:05:44 <onyinyang[m]> ptspec status version support... (full message at <https://matrix.org/_matrix/media/v3/download/matrix.org/dseuSDUiUIfensRbmXFOjZYe>)
16:05:55 <meskio> since a year (or more??) we've being working on a change on the ptspec to report the PT version
16:06:07 <meskio> so we know what version and implementation is using each bridge
16:06:22 <meskio> c-tor is planning to include support for that in the next release 0.4.8
16:06:36 <meskio> https://gitlab.torproject.org/tpo/core/tor/-/merge_requests/731
16:06:50 <meskio> I have implemented it in goptlib:
16:07:23 <meskio> https://gitlab.torproject.org/tpo/anti-censorship/pluggable-transports/goptlib/-/merge_requests/1
16:07:56 <meskio> AFAIK the change is backward compatible both ways, so it should be fine to implement it in our PTs even if they are run with a version of tor that doesn't support
16:08:13 <dcf1> I iwll review that MR. Core team just wanted a quick implementation to test their side against though, goptlib is not blocking them?
16:08:43 <meskio> no, is not blocking them, I did some basic testing and saw the version and implemenation appearing in the bridge descriptor
16:09:13 <meskio> I don't think we need to hurry on this, but will be nice to move it so we get it working at some point :)
16:09:14 <dcf1> I'm looking at https://gitlab.torproject.org/tpo/core/tor/-/merge_requests/731, one thing I am wondering is what happens if there are two different server pts configured
16:09:39 <dcf1> if there is something to disambiguoate their transport-version and transport-implementation
16:09:54 <meskio> yes, I wrote about that, AFAIK the likes will come together one after the other, so is possible to guess wich goes with which, but not so nice
16:10:01 <dcf1> I'll leave some comments on tpo/core/tor!731
16:10:07 <meskio> thanks
16:10:19 <dcf1> meskio: aha, I see your comment
16:10:22 <meskio> we could change the proposal to don't support spaces on the implemenation or version to fix that
16:10:31 <meskio> I don't see a clear need for them
16:10:34 <dcf1> imo prohibiting spaces is worse
16:11:06 <meskio> I don't have a strong opinion here
16:11:12 <dcf1> I wish tor would stop invienting data formats that are so cumbersome to parse and process
16:11:53 <meskio> +1
16:12:03 <meskio> but we live in a legacy world...
16:13:16 <onyinyang[m]> Anymore on this topic or should we move on to the next one?
16:13:32 <meskio> not from me
16:14:03 <onyinyang[m]> ok, let's move on to the next discussion item.
16:14:14 <onyinyang[m]> Webtunnel soft release and next phase: https://gitlab.torproject.org/tpo/community/team/-/issues/94
16:14:47 <meskio> ggus ??
16:14:48 <ggus> so, we are at the end of july and almost all the items for phase #1 are done
16:14:57 <meskio> nice :)
16:15:31 <ggus> the feedback that we collected so far from operators is about: improving the docs for compiling from the source
16:15:48 <ggus> and some ppl asked apache instructions and not just nginx
16:16:01 <ggus> we can work on this during phase #2
16:16:29 <ggus> from users: 1. some ppl in china reported that didn't work. in russia it worked without prblem
16:16:32 <ggus> problem
16:16:37 <ggus> we still need to test in iran
16:16:54 <ggus> (in tm didn't work, but probably because of the ip blockage rule)
16:17:21 <ggus> any question about this first phase? if not, we can move to discuss the next steps
16:17:45 <meskio> I wonder why it didn't work in china, but we can move along and investigate that in the future
16:18:15 <ggus> meskio: here are the logs - https://gitlab.torproject.org/tpo/community/support/-/issues/40119#note_2923287
16:19:15 <meskio> thanks, I'll try to dump some of that into webtunnel issues
16:19:17 <ggus> and this comment in net4ppl: https://github.com/net4people/bbs/issues/263#issuecomment-1621447664
16:19:36 <ggus> ops, this one: https://github.com/net4people/bbs/issues/263#issuecomment-1613856517
16:20:18 <meskio> +1
16:20:34 <ggus> for phase 2, i'm planning to work on moving the readme instructions to the community portal next week
16:20:49 <meskio> sounds good
16:20:57 <meskio> let me know if you need any help to improve or review them
16:21:12 <meskio> shell will be AFK most of next week, but I can help there if needed
16:21:37 <ggus> ok!
16:21:58 <ggus> i'll include a new section here: https://community.torproject.org/relay/setup/
16:22:28 <meskio> should we rename 'bridge' to 'obfs4 bridge' or something?
16:22:31 <ggus> and probably rename "bridge" to "obfs4 bridge"
16:22:32 <ggus> yes
16:22:34 <ggus> hehe
16:22:39 <meskio> :D
16:23:43 <onyinyang[m]> Anything more on this topic?
16:23:43 <ggus> and TB 13 will be release in septmber, are we going to include webtunnel in stable?
16:23:53 <meskio> there was an issue from jacobo to pay attention when fixing the doc:
16:24:00 <meskio> https://gitlab.torproject.org/tpo/anti-censorship/pluggable-transports/webtunnel/-/issues/20
16:24:29 <ggus> yes, i saw that
16:24:41 <ggus> nvjacobo: *
16:24:51 <meskio> I think is a good idea to include webtunnel and conjure in the next TB
16:25:02 <ggus> +1
16:25:05 <meskio> we should poke the applications team about that
16:25:15 <meskio> I mean include support, no UX to enable them
16:25:28 <meskio> (besides pasting a bridgeline)
16:26:07 <gaba> for conjure the idea is still not to have it as a default bridge, right?
16:26:20 <ggus> right
16:26:25 <meskio> +1
16:27:56 <meskio> the plan looks good
16:28:05 <meskio> this is all for me on this topic
16:28:37 <onyinyang[m]> ok cool, let's move on to the final point?
16:28:49 <onyinyang[m]> HTTP connect pts
16:29:02 <meskio> I added that point
16:29:22 <ggus> meskio: we should discuss next week in s96 or here when we will do the public call as we'll both be busy/traveling to ccc
16:29:40 <meskio> ggus: let's do it in the s96 meeting
16:29:58 <ggus> ok!
16:30:13 <meskio> on HTTP connect PTs, nickm was proposing in the past that socks5 is not a great protocol to do PTs and it will be better to use HTTP connect
16:30:40 <meskio> HTTP connect has things like headers, were we can pass arguments fixing our bridge line length and maybe be able to do many other things
16:31:17 <meskio> in the in person meeting we did discuss this a bit, and saw it as a nice change
16:31:38 <meskio> we might want to move that forward to happen sometime in arti
16:31:41 <dcf1> Maybe not just HTTP CONNECT but also MASQUE more generally? Hopefully could open the door to datagram-based protocols (MASQUE has projects for UDP proxying, for example)
16:32:17 <meskio> I don't know anything about MASQUE, but that sounds good
16:32:21 * ggus needs to leave. o/
16:32:36 <onyinyang[m]> thanks for coming ggus !
16:32:44 <meskio> bye ggus
16:33:05 <meskio> dcf1: I wanted to hear your opinion on this, but it sounds like you like the idea
16:33:21 <onyinyang[m]> I am a little bit familiar with MASQUE but not enough to have anything useful to say for now
16:33:29 <dcf1> SOCKS is annoying too, there are like no good standard *server* implementations of SOCKS in commonly used packages, which is necessary for implementing client PTs, and is why goptlib includes a SOCKS server.
16:33:49 <gaba> meskio: do you havea  ticket for this?
16:34:07 <dcf1> Much better HTTP server support in common programming language libraries.
16:34:13 <meskio> gaba: no, that will be the next step, is being mentioned in a ticket about the bridge line size
16:35:32 <dcf1> Even Proteus (new PT presented at FOCI 2023) had to implement its own SOCKS server: https://github.com/unblockable/proteus/tree/99751539b78782d4477411786e4df03b68213e5d/src/net/proto/socks
16:35:46 <meskio> :(
16:35:52 <dcf1> (Proteus implemented in Rust, not using goptlib.)
16:36:17 <meskio> we might see more rust PTs in the future, having something that is easier for more languages will be nice
16:37:47 <meskio> being realistic, this change will take time and I don't expect c-tor ever to support it, but we could work on getting a spec and include it in arti
16:38:23 <meskio> I'll create an issue to discuss that and poke the arti people to see if they are interested in helping (or leading) writting the spec
16:39:56 <onyinyang[m]> anything else on this topic?
16:40:40 <meskio> not from my side, I just wanted to start the conversation
16:41:53 <onyinyang[m]> ok, that brings us to the end of the discussion points. Is there anything else anyone else wanted to bring up?
16:42:59 <meskio> not from me
16:43:21 <onyinyang[m]> ok then, let's end the meeting!
16:43:24 <onyinyang[m]> #endmeeting