- Wednesday, December 7, 2016

release-team@conference.openafs.org

Wednesday, December 7, 2016< ^ >

Room Configuration

Room Occupants

GMT+0
[06:14:16] Jeffrey Altman leaves the room
[06:14:16] Jeffrey Altman joins the room
[12:39:13] Jeffrey Altman leaves the room
[12:39:13] Jeffrey Altman joins the room
[14:25:09] meffie joins the room
[14:53:53] <meffie> good day.
[14:57:00] <mvita> hello
[14:59:13] <kadukoafs@gmail.com/barnowlE45450E9> gretings
[15:01:14] <kadukoafs@gmail.com/barnowlE45450E9> meffie: are you equipped to set up people with secrets for new
buildbot builders?
[15:02:15] <meffie> yes, i believe so.
[15:02:34] Stephan Wiesand joins the room
[15:02:53] <Stephan Wiesand> Hi
[15:03:25] <mvita> hi
[15:03:28] <kadukoafs@gmail.com/barnowlE45450E9> Hi Stephan.
[15:03:39] <kadukoafs@gmail.com/barnowlE45450E9> meffie: hmm, let me add you to this mail thread, then.
[15:03:46] <meffie> ok
[15:04:53] <Stephan Wiesand> So, any Linux news regarding 4.9rc8?
[15:06:09] <mvita> I built it okay (header build)
[15:06:34] <Stephan Wiesand> Ok, thanks.
[15:06:36] <mvita> Joe built and tested fine with rc7
[15:06:51] <mvita> I don't know if he's tested -rc8 yet
[15:08:14] <Stephan Wiesand> 4.9 will probably be released sunday
[15:08:15] <mvita> -rc8 had some linux regression issues so it's at least possible there will be an -rc9
[15:08:31] <Stephan Wiesand> shall we aim for a 1.6.20.1 next wednesday?
[15:08:36] <mvita> but I agree, more likely he will release it
[15:09:43] <kadukoafs@gmail.com/barnowlE45450E9> We should probably discuss whether updates to NEWS for 1.6.x should go
via master, once 12392 lands.
[15:09:48] <mvita> sounds good, Stephan
[15:11:30] <Stephan Wiesand> ben: via master, or via stable-1_8_x?
[15:11:59] <kadukoafs@gmail.com/barnowlE45450E9> Well, both, eventually.  But I was not sure that I was going to branch
just for an alpha.
[15:12:05] <kadukoafs@gmail.com/barnowlE45450E9> We could talk about that, too... ;)
[15:13:15] <Stephan Wiesand> We updated NEWS directly on the 1.6 branch because on master it hadn't been updated for ages.
[15:14:18] <kadukoafs@gmail.com/barnowlE45450E9> My logic for not branching being something like "we don't make any
ABI/API stabilty guarantees at alpha stage, so it's okay to (e.g.)
leave src/rxgk around even though it will be removed for the release,
and it's just silly extra work to merge things to master and a branch
when there's no real benefit".
[15:14:21] <meffie> in the future, it would be good to have a running NEWS on master, with the top section set to [UNRELEASED]
[15:14:23] <Stephan Wiesand> Thus going through master would have been silly IMO.
[15:15:02] <kadukoafs@gmail.com/barnowlE45450E9> I guess the argument for the other side would be that doing the work
on master directly would leave stale version information on master
post-branch, though we could probably revert or otherwise do something
about that for everything other than 'git describe'.
[15:15:43] <kadukoafs@gmail.com/barnowlE45450E9> Whoops, I jumped topics, sorry.
[15:16:49] <kadukoafs@gmail.com/barnowlE45450E9> Agreed that going through master given that state would have been
silly.
[15:18:06] <kadukoafs@gmail.com/barnowlE45450E9> I guess having the NEWS change be in the same commit that implements
the feature is probably not a good idea, since it would cause huge
congestion on NEWS for merge conflicts and such.
[15:18:32] <mvita> agreed
[15:18:38] <Stephan Wiesand> right
[15:19:30] <kadukoafs@gmail.com/barnowlE45450E9> And one can also argue that having NEWS entries only in (e.g.) 1.6.25
for things that are also new in (e.g.) 1.8.1 doesn't make sense, as
users shouldn't be expected to look at new versions on the older
branch in order to see what's new on their current branch.
[15:20:15] <kadukoafs@gmail.com/barnowlE45450E9> So I guess I'm leaning towards "NEWS updates should be done
independently on the stable branches".  I'm still undecided about
whether master should get them once openafs-stable-1_8_x exists.
[15:21:03] <meffie> "NEWS updates should be done independently on the stable branches". yes i think so.
[15:21:17] <Stephan Wiesand> Let's try. I may well not work out, but we'll find out ;-)
[15:21:28] <kadukoafs@gmail.com/barnowlE45450E9> Fair enough :)
[15:21:58] <meffie> and it would be good to have a running NEWS on the master branch, and just dont mix NEWS updates with other files.
[15:22:36] <kadukoafs@gmail.com/barnowlE45450E9> I'm happy to try that out in terms of merging submissions.  I don't
expect to be able to keep on top of things all by myself in terms of
creating the entries.
[15:23:31] <meffie> yes, of course.
[15:23:42] <kadukoafs@gmail.com/barnowlE45450E9> I guess we can move to the next topic ... any comments about strategy
for when to branch 1.8.x?
[15:24:44] <kadukoafs@gmail.com/barnowlE45450E9> (Also I will be asking for code review on 12473 later, in addition to
the 12392, 12393, and 13470 that are already obvious, if people are
bored with the slow pace of conversation.)
[15:24:55] <meffie> i suppose that makes sense when you need to merge something that will not ship in 1.8.x.
[15:25:01] <Stephan Wiesand> It depends on when there will be a strong desire to merge anything not headed for 1.8
[15:25:21] <meffie> ha, we are in accord
[15:25:44] <Stephan Wiesand> surprised?
[15:25:57] <Stephan Wiesand> (and yes, I *am* slow today - sorry)
[15:26:01] <kadukoafs@gmail.com/barnowlE45450E9> > strong desire to merge anything not for 1.8
Which we probably have some level of control over.
I mean, a bunch of stuff has piled up and been waiting for a year
already, and the world hasn't ended, so clearly it can continue
waiting...
[15:26:41] <mvita> heh
[15:26:41] <kadukoafs@gmail.com/barnowlE45450E9> So, that leads me to my current inclination of branching once we've
had enough testing to want to call things "beta" or "release
candidate" instead of "alpha".
[15:26:57] <meffie> yes, sounds good.
[15:26:58] <Stephan Wiesand> np
[15:27:32] <kadukoafs@gmail.com/barnowlE45450E9> which leads to the question of how we are going to (get testing to)
convince ourselves that it should be called "beta"...
[15:28:11] <mvita> yes, that's a tough one
[15:28:35] <kadukoafs@gmail.com/barnowlE45450E9> Anyway, to stop beating around the bush, I did get to reintegrate
current master into debian packaging over the past couple days, and
did some smoke testing yesterday, so I do not know of anything
blocking doing an alpha release.
[15:28:54] <meffie> excellent!
[15:28:56] <Stephan Wiesand> good
[15:30:49] <kadukoafs@gmail.com/barnowlE45450E9> The only thing I had trouble with was the libafsdep files needed for
make_libafs_tree.pl (which dates back to the original IBM AFS?), which
I didn't update for the fortuna import into the kernel.
So that's 12473
[15:31:46] <meffie> ah. i wondered what triggered that commit.
[15:32:52] <kadukoafs@gmail.com/barnowlE45450E9> I guess the debian packaging uses it to effect not building the kernel
module at debian-package-build-time, leaving it for the user via DKMS
or otherwise.
[15:38:31] <kadukoafs@gmail.com/barnowlE45450E9> (I fixed the commit-message typo in 12470, thanks for spotting that.)
[15:43:21] Jeffrey Altman leaves the room
[15:43:22] Jeffrey Altman joins the room
[15:49:32] <kadukoafs@gmail.com/barnowlE45450E9> Are there further topics for discussion, or are we all doing code
review of NEWS?
[15:49:51] <kadukoafs@gmail.com/barnowlE45450E9> (To be clear, I expect to have a 1.8.0pre1 before next week.)
[15:51:49] <meffie> nice.
[15:51:49] <mvita> no other topics atm
[15:52:18] <mvita> YAY
[15:52:37] <meffie> sounds like we better start finding testers.
[15:52:39] <mvita> "It's beginning to look a lot like CHRISTMAS…"
[15:55:13] <Stephan Wiesand> oh yes :)
[15:55:43] <Stephan Wiesand> no further topics today from my p.o.v.
[15:56:13] <Stephan Wiesand> adjourn?
[15:56:57] <meffie> have a good evening.
[15:57:22] <mvita> thank you everyone
[15:57:28] <Stephan Wiesand> Looking forward to 1.8.0pre1. Bye!
[15:57:36] <kadukoafs@gmail.com/barnowlE45450E9> Thanks all
[15:57:40] Stephan Wiesand leaves the room
[16:23:20] <kadukoafs@gmail.com/barnowlE45450E9> (Was anyone planning to look at 12473 or should I just merge it based
on experimental validation?)
[16:32:05] <mvita> looking….
[16:34:21] <kadukoafs@gmail.com/barnowlE45450E9> The change kind of wants more context, but I'm not sure that I can
provide it, having only encountered the libafsdep files as a result of
my debian build being broken.
[16:42:39] <mvita> so when does make_libafs_tree.pl run?  only in RH packaging?
[16:47:47] <kadukoafs@gmail.com/barnowlE45450E9> Also in the debian packaging, but that's not in-tree anymore.
[17:45:46] Jeffrey Altman leaves the room
[17:45:47] Jeffrey Altman joins the room
[18:16:20] meffie leaves the room
[22:16:59] mvita leaves the room