- Friday, February 22, 2019

release-team@conference.openafs.org

Friday, February 22, 2019< ^ >

Room Configuration

Room Occupants

GMT+0
[13:39:17] meffie joins the room
[13:57:15] wiesand joins the room
[14:00:43] <meffie> good afternoon
[14:02:02] <wiesand> good morning
[14:03:36] <mvita> present
[14:06:00] <wiesand> sorry, came straight from a previous meeting and had to get an overview of recent gerrit activity first
[14:06:08] <mvita> okay
[14:06:18] <mvita> I just took a nap in the interim ;-)
[14:07:03] <wiesand> I see one new change on 1.8.x (13503) - is it on the 1.8.3 wish list?
[14:08:02] <meffie> that is a pullup for 1.6.x
[14:08:14] <meffie> oh, no i'm looking at 13505
[14:08:20] <wiesand> Nothing pulled up after Feb 5 has received review yet, those may not make it
[14:08:39] <mvita> not urgent - this problem is relatively rare
[14:08:55] <wiesand> ok, thanks
[14:09:21] <mvita> and the affected site already has a 1.8 patch
[14:09:21] <wiesand> and yes, I also noticed the recent 1.6.x pullups
[14:10:01] <wiesand> looks like many of those failed to get verified due to build errors on Ubuntu 18.04
[14:10:08] <mvita> yes
[14:10:33] <wiesand> is it due to the missing build fix we discussed a few weeks ago?
[14:11:44] <wiesand> I also noticed a couple of path conflicts already. So getting those merged will take some work/time…
[14:11:47] <meffie> the whole stack is needed to avoid the --enable-checking build errors on gcc 7.3 (on the ubuntu 18.04 worker)
[14:12:10] <mvita> I don't know, I didn't look into it much except to see that it's one of those string-truncation warnings
[14:12:35] <meffie> they should be merged in the given order, and there will be no merge conflicts.
[14:13:25] <kaduk@jabber.openafs.org/barnowl> (here now; sorry I'm late)
[14:14:02] <wiesand> mike, what's the bottom of the stack?
[14:16:06] <meffie> https://pastebin.com/zzmAvpRX
[14:17:00] <meffie> 13495 is the first one.
[14:17:16] <wiesand> got it, thanks
[14:17:48] <meffie> thanks! yes, it seems we have not been doing --enable-checking on the 1.6.x branch for some time.
[14:18:38] <wiesand> I really ought to work on the 1.6.x backlog
[14:18:48] <meffie> fortunately these pullups are all that is needed to avoid warnings on gcc 7.3
[14:19:15] <wiesand> I set a aside some time during the weekend, that should help get into a saner state again…
[14:19:35] <wiesand> thanks for those pullups
[14:19:50] <meffie> you are welcome. cheyenne helped too.
[14:21:09] <wiesand> right, thanks again
[14:22:43] <wiesand> Any new candidates for the stable series which haven't been pulled up or discussed lately?
[14:23:23] <meffie> 13493
[14:24:14] <meffie> it is very old, on master and 1.8.x, but has never been on 1.6.x, so md5 ino is broken on 1.6.x without this fix from Andrew.
[14:25:38] <wiesand> especially because it's so old, I'm a bit curious why that feature needs to be added to the old stable series now…
[14:27:23] <meffie> because it has been just being applied to make md5 work on solaris clients
[14:27:50] <kaduk@jabber.openafs.org/barnowl> applied where, by whom?
[14:28:54] <wiesand> and why now, not years ago?
[14:28:55] <meffie> applied to 1.6.x for a very large installation that requires md5 inos
[14:30:03] <meffie> well, because i suppose it was done by andew, and i didnt remember to pull it up to 1.6.x
[14:30:12] <meffie> so, it's on me.
[14:31:20] <meffie> if we dont include it in 1.6.x then, we should just document the afsd -inumcalc feature is broken?
[14:31:57] <wiesand> ok - review, review review, and let's see - but it will have lower priority than most other recent things in the 1.6.x queue for the time being
[14:32:45] <meffie> yes, agreed. it is not urgent, just an oversight on my part. apologies.
[14:33:32] <mvita> Although, if you do happen to encounter an inode number collision, that's not a good day.
[14:34:05] <meffie> well, i mean sited that need md5 inos should apply this patch.
[14:34:16] <mvita> aye
[14:34:41] <meffie> until it is included in a release. you have to enable such with the afsd option
[14:35:11] <meffie> so it is off by default anyway.
[14:36:48] <wiesand> ok, fixes an actual bug encountered on a real site, it's a candidate for a 1.6.x release
[14:37:08] <meffie> thank you!
[14:38:02] <wiesand> but my feeling is that we should become gradually more reluctant to fixing off-by-default features which have been b0rken ~forever in 1.6.x
[14:39:34] <wiesand> more on the stable series to discuss today?
[14:40:04] <meffie> agreed. (in this case it was something that should have been part of the new afsd option, we will not have more options for 1.6.x i think)
[14:41:10] <wiesand> fine.
[14:41:21] <wiesand> on to master?
[14:41:30] <meffie> yes
[14:41:52] <meffie> https://wiki.openafs.org/devel/GerritsForMaster/
[14:42:27] <kaduk@jabber.openafs.org/barnowl> I don't have much this week; it was pretty rough with IETF stuff and my
mom's thesis defense
[14:43:08] <mvita> 13501 and 13502 are not for master, they are based on (and require) rxgk-phase1
[14:43:55] <mvita> 13500 could/should apply to master
[14:44:27] <mvita> 13489 fixes a minor makefile bug that is already merged to master, so it also is targeted for master
[14:46:58] <wiesand> ah, make dest
[14:47:05] <mvita> yes
[14:47:20] <mvita> I'm the only person in the world that still uses it, so it's not that critical
[14:47:24] <mvita> ;-)
[14:47:47] <wiesand> no you aren't ;-)
[14:47:57] <mvita> BROTHER!
[14:48:08] <meffie> i use it too
[14:48:47] <kaduk@jabber.openafs.org/barnowl> I guess we could try to investigate a "make install to fake DESTDIR then
munge" replacement
[14:48:49] <mvita> <seems to have something in his eye just now, give me a moment>
[14:49:57] <meffie> i have an old topic branch called "death to make dest", so a top level "make dest" just uses 'make install' under the covers.
[14:50:35] <meffie> oh, it is called "make dest wrapper for ben"
[14:50:42] <kaduk@jabber.openafs.org/barnowl> :)
[14:50:47] <mvita> heh
[14:51:16] <meffie> i'll retry after the fileserver -auto patches
[14:51:30] <kaduk@jabber.openafs.org/barnowl> Sounds like a plan
[14:51:49] <meffie> (at least we killed make dest in the openafs.spec file)
[14:52:23] <kaduk@jabber.openafs.org/barnowl> I don't suppose anyone got a chance to look at my tweaks in the latest
10567 rev?
[14:53:08] <mvita> looking
[14:53:16] <kaduk@jabber.openafs.org/barnowl> Also I started looking at 13463 again and it seems that my last fix
there only addressed part of the set of related issues, sigh.
[14:54:21] <mvita> no, I have not
[14:56:53] <meffie> ok, i'll look at 13463 again.
[14:57:30] <kaduk@jabber.openafs.org/barnowl> (I will, too; I got interrupted the first time)
[14:57:52] <meffie> cheyenne pushed 13494 for the gcc8 topic
[14:58:37] <kaduk@jabber.openafs.org/barnowl> I saw the mail about that; pretty cool
[15:00:43] <meffie> ok, it's the top of the hour.
[15:01:50] <wiesand> right - are there other topics to dicuss, or shall we adjourn?
[15:01:50] <kaduk@jabber.openafs.org/barnowl> I don't have anything else
[15:02:02] <kaduk@jabber.openafs.org/barnowl> Sorry again for the slow week.
[15:02:16] <wiesand> (echoed)
[15:02:51] <wiesand> Let's adjourn then. Thanks a lot everybody!
[15:03:28] <meffie> thank you. have a good weekend
[15:03:35] meffie leaves the room
[15:03:37] wiesand leaves the room
[15:04:25] <kaduk@jabber.openafs.org/barnowl> thanks everyone!
[18:08:20] meffie joins the room
[21:28:46] meffie leaves the room