Home
release-team@conference.openafs.org
Friday, December 6, 2019< ^ >
Room Configuration
Room Occupants

GMT+0
[12:01:51] mbarbosa joins the room
[14:19:58] meffie joins the room
[17:00:01] <kaduk@jabber.openafs.org/barnowl> Greetings all
[17:04:10] yadayada joins the room
[17:04:21] <yadayada> Hello all
[17:04:22] <meffie> good day.
[17:04:41] <meffie> sorry, was getting coffee...
[17:05:54] <kaduk@jabber.openafs.org/barnowl> I just stepped out to get my tea :)
[17:07:14] <meffie> thank for the recent merged kaduk!
[17:07:34] <kaduk@jabber.openafs.org/barnowl> I was excited to finally have the time to do something useful, yeah
[17:10:49] <kaduk@jabber.openafs.org/barnowl> Is Marcio with is or just idling?
[17:11:35] <meffie> idle i think. i can ping him tho.
[17:11:36] yadayada leaves the room
[17:11:49] <kaduk@jabber.openafs.org/barnowl> It's not urgent; I can just leave comments in gerrit
[17:11:55] <meffie> ok, thanks.
[17:12:50] <kaduk@jabber.openafs.org/barnowl> I guess we can talk a bit about master while wiesand is not here
[17:13:25] <meffie> ok.
[17:13:27] <kaduk@jabber.openafs.org/barnowl> I made a pass through the afs-dir-pkg topic, which is generally pretty
good and just needs a few tweaks.  So hopefully Mark will have some
time to update those
[17:14:13] <kaduk@jabber.openafs.org/barnowl> Now I'm looking at the macos-catalina stuff, since we have a buildbot
but it's not doing anything useful with the build broken
[17:15:03] yadayada joins the room
[17:15:12] <kaduk@jabber.openafs.org/barnowl> It's amazing how much compression the Xcode rewrite of .xib files
produced!  Though, I think maybe some of the removed bits were just
dead code; it's a little hard to read just from the XML
[17:16:17] <kaduk@jabber.openafs.org/barnowl> I made a bit of progress on the FBSD stack that Andrew has been
working on with Tim Creech (mostly in that I made it through the giant
commit of 13812 that I wanted to look at closely), so hopefully that
will stay unblocked as well
[17:16:39] <kaduk@jabber.openafs.org/barnowl> I also took a look at 10831, which we've been talking about on and off
for a long time. (logdeswarning)
[17:17:13] <meffie> ah, should we make a 1.6.x specific commit for that issue?
[17:17:23] <kaduk@jabber.openafs.org/barnowl> I don't think we can take it directly due to API stability
considerations, but that does inspire an alternative that will give us
the functionality we want without breaking the API, so I don't think
we'll need a 1.6.x-specific commit.
[17:18:41] <kaduk@jabber.openafs.org/barnowl> I also see we had some traffic on the ipv6-prep-format-helpers changes
after my review, but I lost track of whether they're fully updated and
ready to re-review
[17:19:55] <meffie> i feel we should drop rx_inet2str() in that patchset. that was scope creep imho
[17:20:30] <kaduk@jabber.openafs.org/barnowl> I didn't find it particularly bothersome, I guess
[17:21:28] <kaduk@jabber.openafs.org/barnowl> I think that's about all the updates I had
[17:21:45] <meffie> ok. also, cheyenne and i have started staging a more compete set of patches for ipv6 migration on github. we want to have a something more complete before pushing tons of drafts to gerrit.
[17:21:54] <kaduk@jabber.openafs.org/barnowl> Makes sense
[17:22:33] <meffie> i'll push something to gerrit to show the location on github an the topic branches.
[17:22:43] <kaduk@jabber.openafs.org/barnowl> okay, thanks
[17:23:15] <meffie> did you see andrew's rxgk-phase2 stack on github?
[17:23:23] <kaduk@jabber.openafs.org/barnowl> I see Yadav pushed 13980; those swig bindings just keep causing
trouble (since only rarely do developers have all the dependencies
installed, I think)
[17:23:38] <kaduk@jabber.openafs.org/barnowl> I don't think I got to look at the rxgk-phase2 stack on github yet, no
[17:23:52] <kaduk@jabber.openafs.org/barnowl> Though I do have a note to get back to the pthread-bos stuff, since
IIRC phase2 depends on it
[17:24:06] <meffie> ok, just to let you know he has it there to see the stack
[17:24:12] <kaduk@jabber.openafs.org/barnowl> Thanks!
[17:24:43] <meffie> andrew says he is going to be looking at the pthread-bozo gerrits next too.
[17:25:44] <kaduk@jabber.openafs.org/barnowl> cool
[17:26:01] <meffie> i think the plan is to wait for 13947 to be merged, then more from github will be submitted to gerrit.
[17:26:28] <kaduk@jabber.openafs.org/barnowl> At least some of those require quite a bit of neurons working at once,
for tracking when we first start up threading, so they're not suited
to working on last thing at night
[17:27:10] <kaduk@jabber.openafs.org/barnowl> okay, that sounds good
[17:28:28] <meffie> with luck we will have a nice 1.9.x for people :)
[17:29:01] <kaduk@jabber.openafs.org/barnowl> indeed :)
[17:29:07] <yadayada> I was trying 1.8.5 on ppc64le and hit a build issue created a gerrit for same 13980.
[17:29:30] <meffie> thank you yadayada.
[17:29:45] <yadayada> I am in process of reviewing native mounts.
[17:30:04] <kaduk@jabber.openafs.org/barnowl> Yes, thank you, and the issue seems properly understood.  I just
commented on it, formatting nits mostly
[17:30:14] <kaduk@jabber.openafs.org/barnowl> Thank you for reviewing native mounts, as well!
[17:30:24] <meffie> thank you yadayada. i was running it on my desktop for a bit.
[17:30:36] <yadayada> sure Thanks
[17:32:20] <kaduk@jabber.openafs.org/barnowl> Any other topics for today?
[17:33:09] <meffie> andrew may have proposal for vldb
[17:33:13] <kaduk@jabber.openafs.org/barnowl> ooh
[17:34:00] <meffie> changes to support large number of volumes, support for rxgk and ipv6
[17:34:15] <meffie> just a heads up.
[17:34:26] <kaduk@jabber.openafs.org/barnowl> Definitely; wouldn't want to have to do that upgrade more than once
[17:34:44] <yadayada> There is one heads up ... I was looking into performance of OpenAFS and we know permission checks hits very often. I was thinking if we can cache some information so that permission checks are not performed for every lookup.  I had discussion with Andrew on this during workshop. I will come up with some initial poc, since it will need lots of iteration should I create gerrit or discuss on mail ?
[17:35:39] <meffie> yes, that is a "known" bottleneck. any work to resolve that would be nice.
[17:36:42] <kaduk@jabber.openafs.org/barnowl> gerrit probably makes it easier to comment on specific parts of the
code.  We've had patchsets with 30+ iterations before, so I don't
think that lots of iterations is a problem in its own right
[17:37:27] <yadayada> sure will go with gerrit
[17:37:33] <meffie> thanks yadayada
[17:38:35] <meffie> ok, that's all i had.
[17:38:54] <yadayada> nothing from my side
[17:39:07] <kaduk@jabber.openafs.org/barnowl> Okay.  Thanks everyone, and have a good weekend!
[17:39:08] <meffie> i do hope we have an openafs release by christmas :)
[17:39:22] <meffie> have a good weekend!
[17:39:31] <yadayada> thanks have a nice weekend !!
[17:41:35] meffie leaves the room
[17:54:41] mbarbosa leaves the room
[17:56:03] yadayada leaves the room
[18:31:10] mbarbosa joins the room
[20:17:08] mbarbosa leaves the room
[20:44:08] mbarbosa joins the room
[22:08:51] mbarbosa leaves the room