Hi,
Here's the summary of the IRC meeting.
---
COMMUNITY MEETING
Place: #openvpn-meeting on libera.chat
Date: Wed 24th August 2022
Time: 10:30 CEST (9:30 UTC)
Planned meeting topics for this meeting were here:
<https://community.openvpn.net/openvpn/wiki/Topics-2022-08-24>
Your local meeting time is easy to check from services such as
<http://www.timeanddate.com/worldclock>
SUMMARY
d12fk, dazo, djpig, kp, lev, mattock, MaxF, ordex, Pippin and plaisthos
participated in this meeting.
---
Agreed that ommunity services migration tomorrow (Thursday) during EET
working time (9-17) is ok. Mattock will take care of it.
---
Talked about the hackathon. It seems like November 26th is the best
option right now.
---
Talked about OpenVPN 2.5. The next release, which is not urgent, should
include an updated tap-windows6 driver.
---
Talked about OpenVPN 2.6. Lev had dco-win pretty comprehensively tested
by OpenVPN Inc. internal QA. A report of the results has been sent to
the mailing list. On the Linux dco side some changes were made which
required simultaneous userspace (openvpn) and kernel-space (ovpn-dco)
changes to maintain compatibility.
Set the tentative release schedule for first 2.6 RC to mid-September,
which would make 2.6.0 November 1st a possibility.
--
Full chatlog attached
(11:30:44) ordex: ay
(11:30:52) MaxF: hi!
(11:30:57) mattock: hi!
(11:33:18) mattock: who do we have here today?
(11:33:41) dazo: o/
(11:33:46) mattock: https://community.openvpn.net/openvpn/wiki/Topics-2022-08-24
(11:33:50) lev__: guten tag
(11:34:00) mattock: I have one topic I want to discuss before I head to my
customary lunch
(11:34:23) mattock: "Community services migration tomorrow (Thursday) during
EET working time (9-17) is ok?"
(11:34:32) mattock: more details on the topic page
(11:35:02) mattock: trac will be affected briefly tomorrow, otherwise you
(=developers) should not notice anything
(11:35:06) mattock: will/would
(11:35:22) ordex: fine with me, as long as we have some time to double check
that things are working before you disappear :D
(11:35:36) mattock: I've allocated a full day (plus Friday) for fixing any
issues
(11:36:03) mattock: I can typically focus on Thursdays from morning to late
evening
(11:36:10) dazo: LGTM; and I agree with ordex
(11:36:14) ordex: :-D
(11:36:16) mattock: +1
(11:36:52) mattock: hackathon?
(11:36:54) d12fk: ooh late, hi
(11:37:05) mattock: hi!
(11:37:28) MaxF: hackathon! Has everyone responded to the poll?
(11:37:36) mattock: I have not!
(11:37:38) mattock: forgot
(11:37:40) mattock: link?
(11:37:43) mattock: I'll do it now
(11:38:06) Pippin_: https://doodle.com/meeting/participate/id/dRgEwERe
(11:38:07) vpnHelper: Title: Doodle (at doodle.com)
(11:38:15) mattock: thanks!
(11:40:12) mattock: done
(11:40:41) MaxF: looks like nov 26 is the least worst
(11:40:58) mattock: that late we could all stay for Christmas :)
(11:41:38) dazo: yeah ... that date might mean I need to leave late on Saturday
.... I'll see what I can manage
(11:42:40) mattock: Sync up on OpenVPN 2.5 and 2.6?
(11:42:42) MaxF: is this everyone? what about corp people?
(11:45:17) plaisthos: Hmpf I still don't see the overview
(11:45:59) dazo: MaxF: Charlie, James, Johan and Mark are the corp folks
missing ... I'll follow up internally, but I'd say the community folks are more
important in this context
(11:47:14) plaisthos: argh and now I removed my choices
(11:47:15) MaxF: the main point is, I need to have a list with everyone's name
on it, and if you're not on it, I can't let you in
(11:47:43) ordex: by when is this required?
(11:48:17) MaxF: not sure, but if we settle on a date in November, I don't need
it right now
(11:49:20) plaisthos: yeah I think we can get that very soon after we finalise
the date
(11:49:26) d12fk: can we fix the date today?
(11:51:21) plaisthos: yeah I would say we should
(11:53:29) dazo: +1
(11:54:24) MaxF: nov 26 looks the best with one "cannot attend" and one "if
need be" (dazo might leave on Saturday evening)
(11:54:40) MaxF: nov 12 has one "cannot attend" and two "if need be"
(11:55:07) dazo: yeah ... Just checked flights and trains, if I need to leave
on Sat, I need to fetch a train 17:24 from Delft
(11:55:21) ordex: we don't have much choice then, unless we ask the "if need
be" on the 12 if they can still make it to the whole meting
(11:55:28) dazo: and arrive 12:00-12:30-ish
(11:55:29) ordex: then 12 would be better to we really have everybody for the
whole meeting
(11:55:29) MaxF: that sounds pretty close to cannot attend
(11:55:52) dazo: (arrive on Friday)
(11:56:22) ordex: still stretched I'd say. what about asking cron2__ and kp
their plan on Nov 12th? maybe they can still do it reasonably
(11:56:23) MaxF: I haven't heard from the "if need be"s on nov 12. If there's
*two* people who have to leave at 17.00, that's worse
(11:56:38) dazo: 12th has 2 "need be", 26th has 1
(11:56:44) ordex: MaxF: it may be they don't favour that date, but still can
join the whole meeting if decided to be on that date
(11:56:56) dazo: 26th is the best, how I see it
(11:57:25) ordex: dazo: the "need be" is grey. your situation is clear, but
kp's and cron2__'s is not. maybe they can easily make it
(11:57:36) ordex: if they also have big troubles, then we go for 26th
(11:57:40) ordex: but I'd suggest to ask them
(11:57:58) ordex: no?
(11:58:13) kp: Hmm, I'm not sure why I didn't mark the Nov. 12th. My calendar
is clear on that day.
(11:58:14) MaxF: agree! are they here?
(11:58:25) MaxF: one down :)
(11:58:25) ordex: kp just emerged :)
(11:58:37) ordex: cron2__ is on vacation so not here - easier to reach via
email I think
(11:58:52) ordex: but may not reply soon
(11:58:59) ordex: so we are at 1 vs 1 need be
(11:59:33) ordex: kp: you marked nov 12th (fri to sun) as "if need be", but
you're saying it is actually a green date for you?
(11:59:50) kp: Yeah.
(11:59:56) ordex: cool!
(12:00:10) ordex: MaxF: then maybe you could send an email to cron2__ and after
that we make a decision?
(12:00:12) plaisthos: kp: it is the whole weekend
(12:00:13) ordex: dazo: makes sense to you?
(12:00:18) mattock2 [~ya...@mobile-access-bcee45-220.dhcp.inet.fi] è entrato
nella stanza.
(12:00:20) plaisthos: kp: the doodle has only one day
(12:00:21) kp: I think I meant the weekend after, the 19th, because I'll be at
concert on the 20th, but I could make that work if I had to.
(12:00:23) MaxF: ordex sounds good
(12:00:27) dazo: ordex: makes sense!
(12:00:29) kp: The whole weekend works.
(12:00:30) ordex: cool!
(12:00:38) ordex: kp: cool, thanks
(12:01:13) ordex: anything else for the hackathon?
(12:01:20) MaxF: also, if we already know it's in November, then I can tell
that to people at work, so it's less urgent
(12:01:35) MaxF: so no big deal if we don't select the date today
(12:01:43) ordex: it seems those two dates are the real candidates
(12:02:06) ordex: MaxF: would you be able to offer a couple of options for the
accommodation? (similarly as done on the wiki for the past editions)
(12:02:15) ordex: so we know what's more convenient/closer
(12:02:22) MaxF: sure!
(12:03:24) plaisthos: the ikea hotel was nice
(12:03:31) d12fk: o_O
(12:03:58) ordex: :D
(12:04:01) ordex: next?
(12:04:25) mattock: yes
(12:04:42) mattock: sync-up?
(12:04:45) ordex: yap
(12:04:53) ordex: 2.5: new windows bundle needed?
(12:06:11) lev__: well not urgent
(12:06:32) lev__: but next time we do release we could include new tap driver
(12:06:34) ordex: ok
(12:06:45) mattock: +1
(12:13:31) dazo: on 2.6 ... I'm going to follow up with NM folks this week;
last week was too busy
(12:14:14) dazo: one question about the "Linux port builders aware" .... what's
that about, needing newer kernel package?
(12:17:08) dazo: nobody wants to do sync-ups it seems .... :-P
(12:17:38) lev__: on 2.6 - I asked corp QA to test dco-win ant it looks good -
he has covered many scenarios - I've sent report to mailing list
(12:20:11) lev__: also rob0 volunteered to test it
(12:21:07) ordex: dazo: that ports thing is something cron2__ added and it is
bsd specific, so...no clue :D
(12:21:26) ordex: also, I reviewed cron2__ MR_WITH_NETBITS patch. so that's
good for now
(12:22:00) MaxF: dco-win is relevant for me as well, but it will be some time
before I can get around to testing it
(12:22:18) MaxF: 2.5 weeks or so
(12:22:40) plaisthos: dazo: API changed basically
(12:23:37) dazo: plaisthos: ovpn-dco or core kernel APIs?
(12:24:26) ordex: well, not real API change
(12:24:33) ordex: but we implemented some fixes
(12:24:43) ordex: which were dual side (change a bit userspace and change
ovpn-dco too)
(12:24:57) ordex: so cron2__ wants to make sure that when userspace is
upgraded, we also upgrade ovpn-dco to be sure
(12:25:09) dazo: If it is tied to the ovpn-dco kernel module, then it's easy to
fix that. If the kernel package needs to be updated, that's a different beast
(12:25:25) ordex: no I think it was about ovpn-dco package only
(12:25:46) dazo: ahh, okay ... the bullet point said "kernel package", which
made me concerned
(12:25:50) ordex: basically we changed something in ovpn-dc and that something
allowed to make code in userspace easier. now if you update only userspace
without the change in ovpn-dco: kaboom
(12:25:53) ordex: yeah
(12:26:01) ordex: but he is probably referring to ovpn-dco
(12:26:05) dazo: But with ovpn-dco updates, I need to thread carefully though,
so I don't break openvpn3-linux users with DCO (some Cloud users has claimed
they wanted to test it)
(12:26:27) ordex: there was no API breaking change
(12:26:47) dazo: okay, good ... then I'll spin up some new ovpn-dco RPMs
(12:27:12) ordex: just openvpn2 required a newer ovpn-dco to make use of a new
behaviour (i.e. being allowed to not send some attributes in NEW_PEER)
(12:27:15) mattock: -> lunch, will follow the discussion on my mobile
(12:27:25) ordex: so yeah, nothing breaking
(12:27:36) ordex: should we try to agree on a release date for 2.6 ?
(12:27:45) ordex: cron2__ proposed Nov 1st
(12:28:06) dazo: brave! But I agree
(12:28:53) lev__: MaxF: ping me if you have any dco-win related questions
(12:29:00) ordex: that would be nice as it will allow perfect sync with some
linux distros
(12:29:10) lev__: perfect sync!
(12:29:16) ordex: in order to get there, by when should we freeze/branch master?
(12:31:00) dazo: with Nov 1st as target release date ... I'd say mid-september
to go through the RC cycles
(12:32:36) ordex: ay!
(12:32:50) ordex: that'd be nice to achieve
(12:33:07) ordex: anyway, for dco + ossl 3.0 build issue: I will probably send
a patch asap
(12:33:21) ordex: nothing major, just ubuntu has a broken .pc file and we want
to workaround things like this
(12:34:24) dazo: oh fun .... :/
(12:35:01) ordex: yeah, but it we revert the way LDFLAGS (or similar) are
provided, we should be able to easily overcome the issue
(12:35:54) dazo: revert or reverse?
(12:39:00) dazo: I believe I proposed a fix somewhere, where some flags was not
appended to the main CFLAGS/LDFLAGS ... but added in advance in some
Makefile.am files where needed
(12:39:30) dazo: IIRC; this was on some systems which had both distro openssl
package installed + self-built and installed openssl 3.0 in /usr/local
(12:52:53) ***dazo take it the meeting has concluded
_______________________________________________
Openvpn-devel mailing list
Openvpn-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/openvpn-devel
