Here's the summary of the IRC meeting.



Place: #openvpn-meeting on irc.freenode.net
Date: Thursday 9th May 2019
Time: 20:00 CEST (18:00 UTC)

Planned meeting topics for this meeting were here:


Your local meeting time is easy to check from services such as



chipitsine, cron2, dazon, mattock and syzzer participated in this meeting.


Discussed the option of enabling Cirrus-CI for FreeBSD testing. Agreed
that it might benefit some users as it would allow FreeBSD testing of
PRs on GitHub. Nevertheless it was decided to not enable it as we have
much wider *BSD tests in Buildbot and we provide Vagrant *BSD VMs for
people to use for development and testing.


Discussed tap-windows6 HLK testing. Mattock made good headway on the
infrastructure today and believes that HLK test results should be ready
in <2 weeks. Jamallx on #openvpn-devel has ran tap-windows6 through all
the HLK tests in his environment, so tap-windows6 is definitely
HLK-compliant now and no big HLK issues are to be expected.


Decided to arrange the second OpenVPN 2.5 mini-hackathon this Friday
(=today) "after lunch" (CEST). The idea is to have a mini-hackathon
every other Friday after Thursday's 20:00 (CEST) meeting if that suits
everyone. This way it is less likely that people forget about them.


Full chatlog attached.

Samuli Seppänen
Community Manager
OpenVPN Technologies, Inc

irc freenode net: mattock

(20:56:25) L'argomento di #openvpn-meeting è: Next meeting on 13/Feb/2019 at 
11:30CET. Agenda at https://community.openvpn.net/openvpn/wiki/Topics-2019-04-11
(20:56:25) L'argomento per #openvpn-meeting è stato impostato da 
dazo!~freenode@openvpn/corp/developer/dazo a 21:18:02 su 11/04/2019
(20:56:31) mattock: almost meeting time
(20:57:20) mattock: who do we have here?
(20:58:09) ***cron2 hides under the table
(21:01:59) ***syzzer present
(21:02:45) mattock: hi syzzer!
(21:02:55) syzzer: I made it to the meeting :')
(21:03:01) mattock: most excellent!
(21:03:04) syzzer: hi mattock1 :)
(21:03:38) chipitsine ha abbandonato la stanza (quit: Read error: Connection 
reset by peer).
(21:03:47) mattock: ok let's start
(21:03:50) chipitsine [~chipitsin@] è entrato nella stanza.
(21:03:56) mattock: chipitsine: here?
(21:04:04) chipitsine: yep
(21:05:14) mattock: https://community.openvpn.net/openvpn/wiki/Topics-2019-05-09
(21:05:16) vpnHelper: Title: Topics-2019-05-09 – OpenVPN Community (at 
(21:05:24) mattock: so we actually have a few pre-created topics
(21:05:30) mattock: start from the top?
(21:05:30) cron2: hi syzzer!
(21:05:38) mattock: hi chipitsine!
(21:06:28) mattock: anyone opposed to "Enabling Cirrus-CI for FreeBSD testing 
in GitHub"?
(21:06:36) mattock: I am not
(21:06:58) chipitsine: I'd suggest to enable it and to see how it goes
(21:07:02) cron2: I am opposed to spending more and more CPU cycles on things 
that have little extra benefit
(21:07:17) mattock: that's somebody else's CPU
(21:07:23) mattock: :)
(21:07:24) cron2: we already have more FreeBSD testing than Cirrus-CI would 
give us (full buildslaves from 7.x to 11.x, 12.x coming up)
(21:07:35) cron2: mattock1: still waste for little good
(21:08:12) syzzer: the only benefit I see for cirrus-ci is that it's easier to 
use for non-core team members
(21:08:18) mattock: yep
(21:08:30) cron2: how so?
(21:08:43) mattock: they can create a PR and have that checked?
(21:08:52) cron2: we do not want PRs
(21:08:54) syzzer: they can just fork the openvpn repo, push to their forked 
repo, and have the tests run
(21:09:15) cron2: syzzer: sure, but for that, *we* do not need to carry 
(21:09:20) syzzer: instead of posting to the mailinglist, waiting for review, 
then waiting for the patch to be applied
(21:09:31) mattock: we would probably want the config file in our repo
(21:09:34) mattock: right?
(21:09:39) mattock: as with travis
(21:10:23) syzzer: cron2: (I'm mostly with you here, hence "the *only* 
benefit") having a config in our repo does make it a lot easier for occasional 
committers to run the tests
(21:10:27) cron2: we could carry it but I do not think we should enable it and 
I'm not overly keen to have to handle patches to it every now and then
(21:11:12) chipitsine: why not to enable it ?
(21:11:16) syzzer: nah, I say either carry it and run it (so we know when it 
breaks) or not at all
(21:11:22) mattock: agree with syzzer
(21:11:39) cron2: chipitsine: it adds cost.  Someone has to pay for electricity 
for this.  It has no benefit.
(21:11:59) cron2: it adds maintenance costs (if little).  It has no benefit.
(21:12:12) syzzer: and cron2 or dazo have to spend time processing patches for 
it, which they could also spend on real functionality
(21:12:17) chipitsine: I do not think it actually adds cost. servers run all 
the time, they do not have tasks to run
(21:12:34) cron2: chipitsine: this is not how modern clouds work
(21:12:35) syzzer: that's not how it works...
(21:13:18) mattock: I have to admit - I don't know how it works, so I will 
quickly check
(21:13:38) syzzer: basically "travis for freebsd" 
(21:13:40) cron2: now... if we did PRs, it has an obvious benefits.  But the 
way we do process patches today, PRs are actually contraproductive, because 
people will create PRs and then we get to chase them to submit patches
(21:14:28) cron2: I can see the benefit of "how do I do freebsd tests?" for 
submitters not having access to the buildmaster repo
(21:14:28) mattock: is it as automatic as travis? do _we_ need to run something?
(21:14:47) chipitsine: it is automated
(21:14:49) mattock: or just enable some plugin in GitHub and have config file 
in the repo
(21:14:52) cron2: but then, we do have the vagrant boxes... and *most* patches 
are unlikely to break *BSDs
(21:14:55) chipitsine: I'm not sure how stable it is
(21:15:06) chipitsine: if it appears not to be stable, we'll drop it
(21:15:42) mattock: could we do it on a separate openvpn GitHub fork that gets 
synced with the main one every now and then?
(21:15:51) mattock: I mean, like chipitsine's
(21:15:58) mattock: if it is not too much work
(21:16:08) cron2: anyone can do anything they want in their github fork :-)
(21:16:13) mattock: yeah
(21:16:53) mattock: and if it seems to work fine then consider activating it 
for OpenVPN/openvpn
(21:17:09) cron2: there we go again.  "No benefit, adds costs"
(21:17:12) chipitsine: it works fine on my own repo
(21:17:16) mattock: ok
(21:17:17) cron2: even "consider" is too much effort
(21:17:21) chipitsine: it does not add costs to us
(21:17:34) mattock: except patch review time cost
(21:17:35) chipitsine: we do not have to worry for someone's electricity
(21:17:39) syzzer: chipitsine: that's a rather self-centric way to look at it
(21:17:55) ***dazo is here
(21:17:58) mattock: hi dazo!
(21:18:40) mattock: so basically cron2 is not sold, but we agree there is some 
small benefit for people who want to have their code tested on FreeBSD
(21:18:59) cron2: "and can neither run vagrant nor have access to the 
buildmaster repo"
(21:19:04) mattock: if they first realize that Cirrus-CI is available, then 
they could make a PR to trigger
(21:19:07) syzzer: so, trying to bring this to an end: I'm with cron2 on this 
one. We have travis and the windows-travis (what's it called again) already. 
Those catch most problems. For the real testing we have buildbot running much 
more extensive tests. That should do it.
(21:19:12) syzzer: We need less work, not more.
(21:19:13) mattock: many people don't want to run Vagrant because of <n>
(21:19:23) mattock: syzzer: appveyor
(21:19:37) mattock: I don't have strong opinions
(21:19:39) dazo: I'm also not sold on the cirrus-ci stuff
(21:19:57) mattock: let's let it be then
(21:20:09) mattock: ok next topic
(21:20:30) cron2: that is fairly easy - I've seen the ACK but I had no time yet 
to give it an extra quick glance and then merge it
(21:20:37) mattock: https://patchwork.openvpn.net/patch/579/
(21:20:38) vpnHelper: Title: [Openvpn-devel,v2] Fix various compiler warnings - 
Patchwork (at patchwork.openvpn.net)
(21:20:40) cron2: propably tomorrow
(21:21:04) cron2: s/pa/ba/
(21:21:28) mattock: next in line: https://patchwork.openvpn.net/patch/725/
(21:21:30) vpnHelper: Title: [Openvpn-devel] tests: remove dependency on base64 
- Patchwork (at patchwork.openvpn.net)
(21:21:41) cron2: that gets a solid "superceded" :-)
(21:21:57) mattock: that was quick
(21:21:59) mattock: tap-windows6 updates 
(21:22:05) cron2: https://patchwork.openvpn.net/patch/728/ is the new one
(21:22:06) vpnHelper: Title: [Openvpn-devel,v2] tests: remove dependency on 
base64 - Patchwork (at patchwork.openvpn.net)
(21:22:10) mattock: noted
(21:22:16) dazo: I think I was fine with an earlier patch, just some minor 
changes (kicking out some wrapper functions which looked superfluous)
(21:22:19) cron2: syzzer: I still think you should have gone with the roman 
numbering :-)
(21:22:36) cron2: dazo: which one?
(21:22:50) syzzer: cron2: weren't you hiding under some table? :p
(21:23:03) dazo: cron2: "Fix various compiler warnings"
(21:23:55) cron2: dazo: that one got an ACK, I was just busy with other stuff.  
Nothing wrong with the patch (AFAICS :) ), it's the v2 with the feedback 
incorporated anyway
(21:24:09) dazo: right
(21:24:46) dazo: (saw the ack now)
(21:26:20) dazo: patch 728 looks reasonable 
(21:26:38) cron2: yep
(21:26:39) dazo: but should probably be tested on our buildbots 
(21:27:45) syzzer: yeah, that.  I still need to re-figure out how to push to 
mattocks repo that triggers builds :p
(21:28:08) dazo: syzzer: you got access to the community vpn?
(21:28:10) mattock: any particular challenges there?
(21:28:22) cron2: build is not in the community vpn
(21:28:27) syzzer: so if anyone with that working could push the patch that 
way... O-)
(21:28:28) cron2: so it's just about account & keys
(21:28:29) syzzer: mattock1: time
(21:28:31) syzzer: I think
(21:28:44) dazo: cron2: oh, I'm not up-to-date then :)
(21:28:45) mattock: if I recall correctly you can push just fine without VPN
(21:28:54) mattock: I can check
(21:29:01) cron2: you can
(21:29:14) syzzer: ooh, that sounds easy enough
(21:29:15) cron2: my repo machine is not my desktop machine, and only the 
desktop machine is in the community VPN
(21:29:34) cron2: (I think mattock changed this because I'm so lazy)
(21:29:45) dazo: hehe :)  okay, that'll work
(21:30:21) cron2: mattock: could you do a quick check if syzzer has an account 
& pubkey?
(21:30:31) mattock: yes, should be possible to push directly via SSH
(21:30:32) chipitsine: being lazy to use vpn when developing vpn :)
(21:30:50) mattock: :)
(21:30:55) mattock: VPNs are a PITA
(21:30:57) mattock: :P
(21:30:57) cron2: being lazy about "how many things do I need to start before I 
can go about pushing code changes"
(21:31:26) mattock: syzzer does have an account and an authorized key
(21:31:33) cron2: \o/
(21:32:36) mattock: all looks good
(21:32:38) syzzer: I even had it half-configured, with a wrong path
(21:32:40) syzzer: but works now
(21:32:41) cron2: my openbsd buildslave needs to have cmake installed, 
otherwise the test won't excercise this code
(21:32:43) syzzer: thanks
(21:32:46) cron2: working on it...
(21:33:01) mattock: I need to split in 20 mins at latest
(21:33:09) mattock: done with this topic?
(21:33:11) cron2: so... tap-windows6?
(21:33:13) mattock: yes
(21:33:27) mattock: I've made good headway in HLK testing infrastructure - the 
openvpn setup is now ready but I need to clear all the firewalls in between the 
"client(s)" and "server(s)"
(21:33:43) mattock: I'm going to test on virtualized Windows Server 2019 HLK 
clients, assuming that rozmansi's patch to make tap-windows6 advertise itself 
as a virtual device does what it is supposed to
(21:33:51) cron2: nice
(21:33:54) mattock: in which case those EC2 Windows Server 2019 instances 
"should be good"
(21:34:18) mattock: jamallx successfully passed the HLK tests for tap-windows6 
so I don't see any major issues ahead when the infrastructure is in place
(21:35:00) mattock: plus according to wintun guys the combo of server 2019 + 
server 2019 core + windows 10 32-bit _should_ get us a "universal" Windows 10 
(21:35:12) mattock: so only two signatures required with fairly modest HLK 
(21:35:19) dazo: cool!
(21:35:37) mattock: I'd expect "real" results in <2 weeks
(21:35:50) mattock: as in: "some signature based on HLK testing from Microsoft"
(21:36:23) dazo: so in <2 weeks we can spin new updates without too much hassle 
any more?  Just issue the new driver update through some blackbox and get the 
signature back and release?
(21:36:24) mattock: the infrastructure should be functional tomorrow I hope
(21:36:46) mattock: dazo: well there will still be lots of hassle
(21:36:54) mattock: but we can make _releases_ :)
(21:37:10) dazo: but not of the epic proportions it has required so far, right?
(21:37:10) mattock: right now one release takes about an hour
(21:37:28) mattock: depends on what "epic" means
(21:37:28) dazo: (not being able to have a release in x months or so)
(21:37:34) mattock: HLK will run for a _long_ while
(21:37:38) mattock: not that yes
(21:37:54) mattock: it will probably take at least 24 hours to get a release 
out with all the tests running
(21:38:04) mattock: _unless_ rozmansi's patch makes a huge difference there
(21:38:05) dazo: that's reasonable
(21:38:08) mattock: which could be the case
(21:38:10) cron2: so, cmake added to netbsd and openbsd buildslaves... let's 
see what happens
(21:38:28) mattock: anyways, that's all I have on tap-windows6
(21:38:37) mattock: next topic?
(21:38:55) cron2: time for me to come up from under the table and frown in 
everybody's direction!
(21:39:16) mattock: next mini-hackathon with more participant? :P
(21:39:18) mattock: participants
(21:39:34) mattock: did we schedule that, or is the schedule automatic?
(21:39:34) cron2: yes, that would be nice :-) 
(21:39:53) cron2: we didn't schedule anything yet, last friday was the first 
attempt at, well, get things going
(21:40:09) mattock: would "after Thursday meeting" work in general?
(21:40:13) cron2: (and it succeeded, though in surprising fashion - sgstair was 
there and we got all the tap driver improvements reviewed and merged)
(21:40:27) mattock: that way it would be less easy to forget about the hackathon
(21:40:28) dazo: that's good at least :)
(21:40:43) mattock: yeah we made huge progress thanks to cron2 and sgstair
(21:40:45) dazo: I do promise to have it much higher on the agenda next time
(21:40:50) mattock: I was just working on HLK docs
(21:41:18) cron2: I'm at home tomorrow, but I have some chores to attend (like, 
cleaning up, and packing gifts for $kid's 9th birthday on Saturday)
(21:41:46) mattock: maybe have "something" tomorrow and start for real after 
two weeks
(21:41:50) cron2: so I won't be able to spend "a full day", but if people show 
up and send me nice things, I can work on it
(21:41:51) mattock: a mini-mini thing
(21:42:32) cron2: in two weeks time, I'll be sitting in a plane for 4h on 
friday afternoon :-) - but the point of having a community is that it's not all 
my tasks :-)
(21:42:32) dazo: cron2: which time of the day would you expect to be available? 
 I will try to block half a day for community dev tomorrow
(21:42:41) cron2: dazo: afternoonish
(21:43:28) dazo: sounds good!  I'll allocate "after lunch" with 4-5 hours 
community time
(21:43:30) cron2: syzzer: did you push?
(21:43:40) cron2: dazo: yeah, this sounds like we might get something done
(21:43:42) syzzer: cron2: yes
(21:43:55) syzzer: I won't have much time tomorrow
(21:44:13) cron2: mmmh, buildmaster didn't pick it up yet, I think
(21:44:25) mattock: it takes a while, it is polling
(21:44:36) mattock: and there is the "will another commit come in" delay
(21:44:45) mattock: can't recall the exact time
(21:44:49) dazo: cron2: I have a meeting at 20:00 which approx 1 hour from the 
office .... and I'm not planning to travel home in between (as that would 
easily add another 30-40 minutes)
(21:45:00) dazo: otherwise afternoon is open :)
(21:45:27) mattock: \o/
(21:45:29) cron2: let's give it a go :-) - I'll have family chores between 4 
and 5 and then again at 6...
(21:45:50) cron2: ok, remotes/origin/syzzer/base64-fix is there
(21:46:15) mattock: I will have to write the summary tomorrow, so if one of you 
could advertise this hackathon thing on some medium I'd appreciate it
(21:46:26) mattock: are we done? I need to split before I miss the bus
(21:46:37) cron2: if you could advertise it on the internal chat...
(21:46:41) cron2: plaisthos, ordex...
(21:46:42) dazo: run, mattock, run!
(21:46:42) mattock: lol yes
(21:46:46) cron2: run, mattock, run :-)
(21:46:50) mattock: I have 11 mins
(21:46:55) dazo: cron2: I added it to our team calendar ;-)
(21:46:59) mattock: but I will mention this on #openvpn-devel and our internal 
(21:47:06) cron2: I'm basically waiting for jjk, plaisthos, ordex, syzzer to 
give reviews... :-)
(21:47:15) cron2: well, jjk wanted to send a patch for --genkey
(21:48:13) mattock: ok, splitting now
(21:48:16) syzzer: yeah, I think we had off-list consensus on the direction to 
go with that
(21:48:22) mattock: "see" you tomorrow
(21:48:26) syzzer: run, mattock, run!
(21:48:26) cron2: *wave*
(21:48:29) mattock: bye!

Attachment: signature.asc
Description: OpenPGP digital signature

Openvpn-devel mailing list

Reply via email to