Home
release-team@conference.openafs.org
Thursday, July 28, 2022< ^ >
kaduk@jabber.openafs.org/barnowl has set the subject to: openafs release team
Room Configuration
Room Occupants

GMT+0
[15:38:14] wiesand joins the room
[16:01:00] meffie joins the room
[16:01:09] <meffie> greetings
[16:01:37] Cheyenne joins the room
[16:01:41] <Cheyenne> Hello all
[16:02:18] <kaduk@jabber.openafs.org/barnowl> greetings
[16:03:06] <wiesand> Hello
[16:06:18] <wiesand> So I started looking closer at those stacks to merge. They are not completely linear, are they?
[16:06:30] <Cheyenne> THey are completely linear
[16:07:16] <meffie> the top one is 15065, correct?
[16:07:26] <Cheyenne> If you fetch 15065 you should have the entire stack ready to go
[16:08:33] <meffie> yes, that is what Alex Prengel did for fedora
[16:10:05] <meffie> Cheyenne, is 15068 needed for 1.8.x? it landed on master.
[16:11:22] <wiesand> Ah, right, thanks. I was a bit surprised about that 5.18-prereq stack sitting on top the 5.18 one.
[16:12:08] <kaduk@jabber.openafs.org/barnowl> Is 15068 needed?  Not really.  But it should be okay for 1.8.
It's cleanup for a case that we don't actually have any reason to
think gets executed with any regularity
[16:12:39] <meffie> ah, thanks for the explanation.
[16:12:51] <Cheyenne> correct (was just reviewing the commit message to see if it's really needed).  It's to handle an exceptional case
[16:13:15] <Cheyenne> but if wanted .. it's an easy (and small enough) commit to add.
[16:14:09] <meffie> just checking we didnt miss one. thanks.
[16:16:47] <Cheyenne> And if we want to be ahead of the curve, there are the 3 commits for 5.19 -- which was fairly simple
[16:19:06] <wiesand> I really really hope to get your stacks merged soon. I'll look into 15068 and the linux-5.19 ones then.
[16:21:23] <meffie> excellent! thanks wiesand
[16:21:48] <kaduk@jabber.openafs.org/barnowl> Looking forward to it :)
[16:22:24] <Cheyenne> Thanks Wiesand :)
[16:24:52] <wiesand> Do you know offhand about points in the stack where I should *not* interrupt merging?
[16:25:09] <meffie> ?
[16:26:09] <wiesand> (like, you really shouldn't merge the 5.18-topic, then go to bed and merge the 5.18-prereq ones the next day or so?)
[16:26:27] <meffie> it is one stack, the topics dont matter
[16:27:14] <kaduk@jabber.openafs.org/barnowl> There might be a little weirdness around the "Cleanup AFS_*LINUX_ENV
usage" changes, so maybe only do those four together, but I don't
actually know of specific problems with stopping in the middle there.
[16:27:15] <meffie> as long as you keep the order the same, you can stop at any time.
[16:27:49] <wiesand> *that's* what I wanted to hear, thanks :)
[16:28:12] <kaduk@jabber.openafs.org/barnowl> Well ... I think if you take "import of code from c-tap-harness" but
not the next commit ("tests: Accommodate c-tap-harness 4.7"), the
tests will not work.
[16:28:16] <Cheyenne> yes -- if you keep the order the same it should be "okay" -- If you look at the buildbot results, I think there were one or two "broken" commits that are fixed by the next commit in the stack
[16:28:39] <wiesand> I remember a few cases where changes should not be merged unless the next one (with a fix) is merged immediately after
[16:29:50] <wiesand> I'll try to watch out for those then.
[16:30:32] <kaduk@jabber.openafs.org/barnowl> I don't think I saw any "this change has a critical bug and you should
not have this code without the following bugfix" changes there.  Just
"the build will be sad" or "the tests will be sad" type things
[16:31:03] <wiesand> Ok, those are at least easy to identify.
[16:32:10] <wiesand> on to master/1.9?
[16:32:14] <kaduk@jabber.openafs.org/barnowl> ok
[16:32:39] <kaduk@jabber.openafs.org/barnowl> not too much from my end, I finally merged the bozo
no-fixed-length-strings changes
[16:33:18] <Cheyenne> still getting a clean build with linux-5.19-rc8+
[16:33:27] <kaduk@jabber.openafs.org/barnowl> I made it maybe a third of the way through 15067 (use make to build
the man pages), so Mike may want to look at some of those comments
(some of which will transfer to the other manpage categories' changes)
[16:33:28] <meffie> yay! Thank you. i'll be pushing the next set of patches for bozo (there are about 12 more commits)
[16:34:02] <meffie> (we still have fixed length strings in the BosConfig, for example)
[16:34:18] <kaduk@jabber.openafs.org/barnowl> Oof.  The joy of a legacy codebase ... thank you for taking this on.
[16:34:42] <meffie> yes, legacy.
[16:34:57] <kaduk@jabber.openafs.org/barnowl> I see that Andrew uploaded a 16-commit stack for viced "recently"
(maybe 19 July?); is there anything noteworthy about that?
[16:35:10] <meffie> thank you for looking at the 'make man pages'! :)
[16:35:32] <meffie> yes, more bugs in the host package in the fileserver.
[16:36:12] <kaduk@jabber.openafs.org/barnowl> Ok.  The two that I've opened so far are nice narrow/targeted fixes,
which should make the review straightforward (if not necessarily
"easy" due to the need to check wider-ranging consequences)
[16:38:05] <meffie> understood. yes, this is a critical area of the code.
[16:39:47] <kaduk@jabber.openafs.org/barnowl> I don't think I have other topics for today (it's IETF week, so even
though I'm only attending remotely, they are long days)
[16:41:11] <meffie> just wanted to mention. i added the 'rpm builder' on the buildbot. it runs on alma9 and does a test rpmbuild and unpacks the dkms-openafs package and does a test kernel module build.
[16:41:17] <Cheyenne> I want to take a look at a closer look at the issues around gerrit 15076.  The autoconf and where the libtool dependencies are need some attention
[16:42:00] <kaduk@jabber.openafs.org/barnowl> Thanks for taking that on, Cheyenne.  I merged the change to get a
quick band-aid in, but agree that the longer-term fix needs a closer
look.
[16:42:02] <Cheyenne> 15076 was a band-aid
[16:42:05] <meffie> thanks ben for merging the "quick" fix 15076
[16:42:58] <kaduk@jabber.openafs.org/barnowl> Thinking of that as the quick band-aid for fedora, reminds me that I
had to put a band-aid in for debian and didn't push to gerrit yet.
[16:43:08] <kaduk@jabber.openafs.org/barnowl> (actually, I think I didn't even upload it to debian yet, whoops)
[16:43:32] <meffie> fyi, 15076 is not related to fedora.
[16:43:56] <meffie> just happens to be where we saw it first.
[16:43:56] <kaduk@jabber.openafs.org/barnowl> Isn't that what alex prengel needed in order to build DKMS on his
fedora systems?
[16:44:02] <kaduk@jabber.openafs.org/barnowl> oh
[16:44:26] <meffie> is is related to how we build dkms on rhel like systems.
[16:44:45] <meffie> so, fedora and friends. but not just fedora.
[16:44:50] <kaduk@jabber.openafs.org/barnowl> The debian thing is basically bringing back strcasecmp_pag.o back
(removed by 14743) since afs_util.o is not compatible with the
stripped-down minimal afspag.ko module
[16:44:52] <Cheyenne> correct
[16:45:20] <meffie> ug, ok.
[16:45:37] <kaduk@jabber.openafs.org/barnowl> It looks like I just added a "config.h" include and copied the
afs_strlen() definition from afs_util.c to strcasecmp.c, so I will
hopefully get that submitted to gerrit for review soon.
[16:46:24] <kaduk@jabber.openafs.org/barnowl> In my local tree I did the "partially revert ..." and adding the new
stuff as separate commits, though.  Not sure if it's best to just
combine the changes into a single commit.
[16:47:56] <wiesand> Any other topics for today?
[16:48:09] <meffie> i have nothing more.
[16:49:12] <Cheyenne> Nothing here..
[16:49:13] <kaduk@jabber.openafs.org/barnowl> nothing here
[16:49:18] <wiesand> Fine, let's adjourn. I'll walk home, water the garden, run a smoke test of the 15065 checkout and then start merging (unless the neighbours invite me over to a cold beverage, which is what happened last week…)
[16:49:25] <wiesand> Thanks a lot everyone!
[16:49:27] <meffie> have a good evening all.
[16:49:46] <kaduk@jabber.openafs.org/barnowl> :) Hard to turn down a cold beverage.
[16:49:49] <Cheyenne> wiesand: Holler if you have any questions.. you have my contact info :)
[16:49:53] <kaduk@jabber.openafs.org/barnowl> have a good week all
[16:49:58] meffie leaves the room
[16:50:03] <wiesand> I do. Thanks!
[16:50:19] wiesand leaves the room
[16:50:29] <Cheyenne> have a good week everone..