- Thursday, June 10, 2021

release-team@conference.openafs.org

Thursday, June 10, 2021< ^ >

kaduk@jabber.openafs.org/barnowl has set the subject to: openafs release team

Room Configuration

Room Occupants

GMT+0
[05:52:03] kaduk@jabber.openafs.org/barnowl leaves the room
[05:55:46] kaduk@jabber.openafs.org/barnowl joins the room
[11:41:08] mbarbosa joins the room
[13:13:28] meffie joins the room
[13:42:09] Cheyenne joins the room
[15:58:17] <meffie> greetings.
[15:58:35] <kaduk@jabber.openafs.org/barnowl> good (just barely still) morning!
[16:00:08] wiesand joins the room
[16:00:26] <wiesand> Good morning
[16:00:36] <Cheyenne> Hello
[16:02:02] mvita2 leaves the room
[16:02:05] <wiesand> Hi Cheyenne, hope all is going well?
[16:02:57] <Cheyenne> yep.. will need some minor tweaks.. but things are looking clear :)
[16:03:28] <wiesand> great!
[16:04:18] <wiesand> I'm uploading 1.8.8pre1 sources right now
[16:04:52] <kaduk@jabber.openafs.org/barnowl> yay!
[16:05:23] <meffie> hooray!
[16:05:29] <Cheyenne> yay!
[16:06:15] <wiesand> verifying…
[16:09:01] <wiesand> Loks good. Maybe it's time to tag commit c978f92cbf9306c0b12eec68cd64ebac045c65e6 as openafs-stable-1_8_8pre1 . Ben, would you?
[16:09:19] <kaduk@jabber.openafs.org/barnowl> My pleasure!
[16:13:04] <wiesand> Got it. Thanks!
[16:13:19] <wiesand> I also release the volumes. So it's out there now.
[16:14:08] <kaduk@jabber.openafs.org/barnowl> Thank you!
[16:14:21] <Cheyenne> master and 1.8.x are still building cleanly with latest linux-5.13-rc5+
[16:14:38] <wiesand> What's missing is the web change and the announcement. I may not get around to those today, but should tomorrow the latest.
[16:15:05] <wiesand> Thanks Cheyenne, good to know our shiny new pre1 is still good with recent kernels.
[16:15:38] <Cheyenne> I believe latest compilers are all okay as well
[16:16:56] <wiesand> Even batter
[16:17:13] <wiesand> Anything else on 1.8.x today?
[16:18:29] <kaduk@jabber.openafs.org/barnowl> Just to ask whether there was anything we knew we planned to put in
after pre1 but before 1.8.8 final -- my memory is apparently not so
great on this front.
[16:18:58] <wiesand> some more FBSD catchup, maybe
[16:19:05] <kaduk@jabber.openafs.org/barnowl> Ah, right.
[16:20:25] <wiesand> I think the other items on the wish list are rather for 1.8.9
[16:20:35] <kaduk@jabber.openafs.org/barnowl> That makes sense
[16:21:50] <kaduk@jabber.openafs.org/barnowl> In terms of master/1.9, I seem to have been moving slowly this week,
possibly due to this cold or something that I picked up.  But I did
merge and review some things this morning already.
[16:22:05] <kaduk@jabber.openafs.org/barnowl> Thanks Mike for quickly looking at the NoAuth un-documentation :)
[16:22:34] <meffie> no problem. good catch on the admin guide.
[16:22:37] <wiesand> 14594 would be another pre2 candidate, I just have to disentangle the connections with 10831 & others (among those, FBSD)
[16:24:46] <Cheyenne> Oh.. BigSur changed a compiler default (-Werror-implicit-function-declaration).  14631 fixes a problem during configure
[16:25:36] <kaduk@jabber.openafs.org/barnowl> Oddly enough, I was just about to push the approve button on that one
for master :)
[16:26:05] <meffie> neat :)
[16:26:45] <meffie> speaking of Big Sur, marcio is testing Big Sur on M1 with good results.
[16:27:08] <meffie> some minor changes are in progress.
[16:28:04] <wiesand> Given Apple's pricing policy, M1 will probably take the lead qickly
[16:28:14] <Cheyenne> just a side note on the 14631, there are some additional implicit-function-declaration errors during config for bswap16/32 -- but after investigating them they are "okay" -- I was tempted to push a comment only commit about it.
[16:28:35] <wiesand> (pu 14631 on the pre2 candidate list)
[16:29:31] <kaduk@jabber.openafs.org/barnowl> nothing wrong with comment-only commits
[16:31:17] yadayada joins the room
[16:34:34] <kaduk@jabber.openafs.org/barnowl> I don't think I had other particular master/1.9 topics to discuss.
[16:35:02] <kaduk@jabber.openafs.org/barnowl> I will note that the workshop is next week ... I haven't started on my
slides yet for the openafs/release-team update
[16:35:29] <meffie> and without a plane ride to work on them...
[16:36:32] <wiesand> arr missed the chance to offer a talk, and still have to register…
[16:37:03] <wiesand> but at least the time of the 1.8.8pre1 release is kind of right ;-)
[16:37:20] <meffie> yes
[16:37:30] <kaduk@jabber.openafs.org/barnowl> yes :)
[16:38:07] <meffie> nothing like a conference to help get a tag out the door.
[16:38:40] <wiesand> well this time it's pure coincidence
[16:39:00] <kaduk@jabber.openafs.org/barnowl> I'll also note that I pushed a new rev of 14465 and welcome further
reviews
[16:40:13] <meffie> ok
[16:40:50] <meffie> kaduk, may i ask about our use of regex libraries?
[16:41:33] <meffie> it's a bit of a mess now. there is HAVE_POSIX_REGEX and then a fall back to something in src/util
[16:42:13] <kaduk@jabber.openafs.org/barnowl> Oh gosh, there was a security advisory many years back where I had to
worry about the regex libraries.
[16:42:22] <kaduk@jabber.openafs.org/barnowl> It's definitely a bit of a mess!
[16:42:23] <meffie> yes
[16:43:39] <kaduk@jabber.openafs.org/barnowl> But I don't think I've looked at it since then.
[16:43:43] <meffie> i was thinking we could put a copy of an appropriately licensed and supported lib in src/external and get rid of the HAVE_POSIX_REGEX/libutil dance
[16:43:44] <kaduk@jabber.openafs.org/barnowl> Did you have a proposal?
[16:44:11] <kaduk@jabber.openafs.org/barnowl> Hmm.  That's worth considering, yes.
[16:44:27] <meffie> offhand, it looks like PCRE would be a good candidate.
[16:45:29] <meffie> it builds on windows and everywhere we care about
[16:45:58] <kaduk@jabber.openafs.org/barnowl> Debian will, of course, complain about bundled libraries and want the
option to use an "external" (system-provided) library.
But the main concern I would have is just how much of a backwards
compatibility break it would be.
[16:46:37] <kaduk@jabber.openafs.org/barnowl> I'm pretty amenable to just doing it and documenting it in the release
notes, though -- if there is breakage we can help people update their
scripts.
[16:47:07] <meffie> maybe we just have a opr_regex, that happens to use some subset of PCRE (or something else)
[16:48:42] <kaduk@jabber.openafs.org/barnowl> I don't even remember what we use regex for ... I don't think it was
very much, so we probably have a lot of flexibility about what we use
and what we claim to adhere to
[16:49:28] <meffie> ok, thanks.
[16:49:36] <kaduk@jabber.openafs.org/barnowl> On a different topic, it seems that the ubuntu buildbot workers are
showing test failures on master at the moment.
IIRC it's the superuser and rxperf tests, perennially sensitive to
environmental settings
[16:49:53] <meffie> hmm
[16:51:07] <Cheyenne> looks like it's failing in the tests not the actual build
[16:51:39] <Cheyenne> rx/perf test 3
[16:51:41] <kaduk@jabber.openafs.org/barnowl> right
[16:51:48] <meffie> not ok 3 - multi threaded client ran succesfully
[16:52:33] <meffie> i'll have to take a look.
[16:53:58] <kaduk@jabber.openafs.org/barnowl> I think a couple of the tests (we've seen similar issues on debian
buildds) are sensitive to the specific /etc/hosts contents, but now I
am forgetting exactly how.
[16:55:29] <meffie> i'll try to run them manually on the same box to see what is wrong
[16:56:48] <kaduk@jabber.openafs.org/barnowl> Thanks, that's going to be the most efficient way forward, I think
[16:57:49] <yadayada> One update from my side. On latest kernel non-gpl module build will fail because CONFIG_PPC_QUEUED_SPINLOCKS is enabled, and due to this spinlock operation use GPL shared_processor variable
[16:57:54] <yadayada> https://bugzilla.redhat.com/show_bug.cgi?id=1966245
[16:58:15] <yadayada> this is specific for ppc64le arch though
[16:58:46] <yadayada> the kernel patch which makes shared_processor GPL is submitted by an IBMer
[16:59:40] <wiesand> let's hope they'll accept it
[17:00:04] <yadayada> I am in discussion with him and I am hopeful that shared_processor will be made non-GPL.
[17:00:05] <wiesand> if they don't, any chance for a wokaround on our side?
[17:01:56] <yadayada> since all non-GPL modules will be imapcted, I think change shared_processor has to be reverted to nonGPL. Whatever discussions I had till now, I do not see any need for making shared_processor as GPL, so I am hopeful license will be reverted
[17:02:50] <yadayada> will keep you posted on the outcome and when patch for reverting of license will be ready
[17:02:56] <wiesand> [knocks on wood]
[17:03:06] <wiesand> Anything else to discuss today?
[17:03:35] <wiesand> If there isn't… adjourn?
[17:03:46] <kaduk@jabber.openafs.org/barnowl> nothing from here
[17:04:01] <Cheyenne> nothing here either
[17:04:41] <meffie> see everyone next week at the virtual workshop
[17:04:54] <kaduk@jabber.openafs.org/barnowl> Thanks everyone, and see you next week!
[17:05:27] <wiesand> Thanks a lot everybody. See you next week.
[17:05:38] <wiesand> Now of to watering the garden…
[17:05:45] <yadayada> thanks all
[17:08:06] <Cheyenne> have a good week everyone
[17:26:31] yadayada leaves the room
[17:46:31] yadayada joins the room
[18:46:04] mbarbosa leaves the room
[18:46:06] mbarbosa joins the room
[19:00:02] yadayada leaves the room
[21:42:37] meffie leaves the room
[21:53:05] wiesand leaves the room
[22:00:15] mbarbosa leaves the room
[22:41:31] Cheyenne leaves the room