[06:03:05] --- Marc Dionne has become available
[06:40:51] --- meffie has become available
[06:59:24] --- deason has become available
[06:59:42] --- wiesand has become available
[07:00:16] <wiesand> Hello. Just about made it...
[07:00:27] <meffie> hello
[07:01:05] <kaduk@jabber.openafs.org/barnowl> I won't look at the sub-minute timestamp precision ;)
[07:01:36] <wiesand> budget meetings... lengthy...
[07:02:16] <wiesand> Ok, how about going for a pre2 release?
[07:02:57] <wiesand> Do you think we don't need one (yet)?
[07:03:48] <kaduk@jabber.openafs.org/barnowl> "What has changed since pre1?"
[07:04:12] <wiesand> Mavericks fix, touching upon files used on other platforms.
[07:04:20] <wiesand> RHEL5 build fix.
[07:04:40] <wiesand> SLES 11 SP3 build fix.
[07:04:41] <Marc Dionne> has there been much feedback on pre1
[07:05:03] <wiesand> None of this is merged yet, but I think we want to before 1.6.6.
[07:05:20] <wiesand> Of course not.
[07:05:46] <kaduk@jabber.openafs.org/barnowl> A pre2 after merging those listed changes seems fairly reasonable.
[07:06:21] <meffie> agreed
[07:06:58] <wiesand> Thanks.
[07:07:16] <wiesand> Have I missed anything coming up on master we'd like to include in 1.6.6?
[07:08:06] <wiesand> Marc, any dark clouds on the Linux sky?
[07:08:17] <Marc Dionne> 10530 is for 3.13, but it's a warning fix only
[07:08:35] <Marc Dionne> we're still good with current linux git, as of this morning
[07:08:42] <kaduk@jabber.openafs.org/barnowl> -Werror is set for my (freebsd) kernel builds ;)
[07:10:23] <Marc Dionne> with current being 3.13-rc2+, probably another 6+ weeks before release
[07:11:00] <wiesand> [looking at 10530]
[07:12:05] <kaduk@jabber.openafs.org/barnowl> I guess I could phrase my comments in the form of a patch targetted
for master, if Ken isn't going to do anything about it.
[07:12:53] <wiesand> Ben, which one?
[07:14:17] <wiesand> The lack of feedback on pre1 is a real problem.
[07:14:39] <wiesand> Pre2 is another chance to beg ;-)
[07:14:45] <deason> he means his comments on 10519
[07:14:56] <deason> (er, I think :)
[07:15:06] <kaduk@jabber.openafs.org/barnowl> (Er, on 10519, sorry)
[07:15:11] <deason> but those are not worth blocking on, so no sense on spending time on it here
[07:15:37] <kaduk@jabber.openafs.org/barnowl> ((was off in another window grumbling at libauth's apparent need for a
librxgk.a))
[07:16:37] <wiesand> Thanks for clarifying. Yes, corrections should go through master at this point, unless 10519 isn't acceptable.
[07:17:39] <wiesand> And the rest is rather straightforward IMO.
[07:19:59] <wiesand> So, merge 10519, 10472, 10518, update NEWS, then issue pre2?
[07:21:19] <wiesand> I'll take the silence as a yes.
[07:21:33] <kaduk@jabber.openafs.org/barnowl> okay.
[07:22:09] <wiesand> The 1.9 branch was planned for "after thanksgiving". Any news on this plan?
[07:22:38] <wiesand> (though that's not my business of course ;-)
[07:23:45] <wiesand> I'll take the silence as a no.
[07:24:07] <wiesand> Anything else to discuss today?
[07:24:16] <kaduk@jabber.openafs.org/barnowl> <more>
[07:24:48] <wiesand> "Anything more to discuss today?" ?
[07:24:59] <kaduk@jabber.openafs.org/barnowl> I'll just mention that we appear to be seeing RT 131780 more
frequently, of late.
[07:25:08] <deason> is that ben raising his hand to discuss something else?
[07:25:10] <kaduk@jabber.openafs.org/barnowl> I don't think we have anything actionable on that front, though.
[07:25:47] <wiesand> Is this the one with a patch from Anders on master which apparently didn't help?
[07:25:55] <kaduk@jabber.openafs.org/barnowl> Yup.
[07:26:12] <kaduk@jabber.openafs.org/barnowl> We used to only be seeing it on the scripts.mit.edu webservers, but
have started seeing it on dialups, too.
[07:26:26] <deason> I could spend some more time thinking about it at some point, but there is another way to fix this
[07:26:43] <deason> which is implementing mountpoints on linux as bind-mounts to /afs/.:mount, or whatever it's called
[07:26:44] <kaduk@jabber.openafs.org/barnowl> I think Anders had an idea or two as well.
[07:27:06] <deason> which I have a half-finished implementation for back when we were discussing the 5-ish different ways to implement this
[07:27:19] <kaduk@jabber.openafs.org/barnowl> Maybe there should be mail to -devel to hash things out; I'm not very
familiar with the linux internals.
[07:27:29] <wiesand> I believe we've seen 131780 a few times.
[07:27:40] <kaduk@jabber.openafs.org/barnowl> I mostly just created the ticket so that we had a place to track
discussion/progress.
[07:29:58] <wiesand> If I understand the comment in 131780, not fixable?
[07:30:39] <kaduk@jabber.openafs.org/barnowl> I am hearing (not from Jeffrey) that there are things we can do, they
may just be painful.
[07:30:47] <deason> ben: if you/he wants, there is some prior discussion in a thread here: https://lists.openafs.org/pipermail/openafs-devel/2012-June/018886.html
[07:31:24] <kaduk@jabber.openafs.org/barnowl> I can certainly link to that from RT; Anders should see that.
[07:32:46] <deason> wiesand: jeff's not saying it's not unfixable, but that it's not worth the effort of maintaining the module
[07:32:57] <deason> er, "not saying it is unfixable"
[07:34:03] <wiesand> Yes, he's been doing that since I first met him in person ;-)
[07:34:09] <deason> but that is much too large of a discussion to just have in here
[07:35:03] <wiesand> My personal belief is that getting the FUSE client fully functional is much more realistic. And sufficient.
[07:35:36] <kaduk@jabber.openafs.org/barnowl> "what's your story for pioctls?"
(no need to answer that here)
[07:36:11] <wiesand> Someone I believe once said it's doable.
[07:36:31] <deason> it's doable, but there's a bunch of different ways to implement it so someone just needs to do one
[07:36:41] <deason> and this isn't a discussion for here
[07:36:59] <wiesand> Agreed.
[07:37:52] <wiesand> I wouldn't want to delay pre2 for 131780, unless a solution is near.
[07:38:25] <Jeffrey Altman> sorry I'm late.   Its been a busy morning.
[07:39:00] <deason> anything else, then?
[07:39:01] <wiesand> Hello.
[07:39:13] <wiesand> Let's give Jeff a minute?
[07:39:28] <Jeffrey Altman> as a process point.  it would be useful if when the discussions unrelated to the 1.6.x branch are brought up that everyone switch to the "openafs" room.
[07:39:44] <deason> sorry, my messages are a bit delayed
[07:40:01] <kaduk@jabber.openafs.org/barnowl> I don't really expect a 131780 workaround in 1.6.6 at all, really.
[07:40:51] <Jeffrey Altman> I do not believe that any 1.6.x release should not be delayed for a fix to 131780.   Any such change is a significant change that should not be introduced in a stable release series.
[07:41:11] <Jeffrey Altman> s/should not be delayed/should be delayed
[07:41:27] <deason> that's not necessarily true; we don't know what a fix would be
[07:42:08] <Jeffrey Altman> there is no trivial fix.   this is a problem that has been known for a decade.
[07:42:28] <deason> those two statements have nothing to do with one another
[07:42:43] <Jeffrey Altman> if there was a trivial fix jhutz, shadow, deason, mdionne and others would have implemented it long ago
[07:43:01] <deason> there have been several minor fixes dealing with mountpoint processing and linux vfs integration
[07:43:26] <deason> some of them are very difficult to find out what they are, but the actual implementation can be small and limited and scope
[07:43:41] <deason> "limited in scope"
[07:44:35] <Jeffrey Altman> when you have a such a solution please propose it.   At that point we can evaluate what the potential side effects might be.   However, I am going to stand by my statement.
[07:44:47] <deason> nobody _expects_ it to be fixed in 1.6.x, which is what ben said
[07:44:58] <deason> but a fix for it may be appropriate for 1.6.x if it appears
[07:45:03] <meffie> well, it seems more work is to be done, but in the meantime, we need to get more people to test 1.6.6pre*
[07:45:30] <kaduk@jabber.openafs.org/barnowl> (Well, I said 1.6.6, but maybe I should have said 1.6.x)
[07:45:58] <deason> oh, okay, well, I'm saying 1.6.x, then :)
[07:46:12] <wiesand> Ok, no need to wait for such a fix. If a miracle happens - fine.
[07:46:12] <Jeffrey Altman> then there is no point arguing with me
[07:46:25] <meffie> it would be great to find a solution that would work for a stable release. if not, then it would be good to find a solution on a later release :)
[07:46:28] <deason> "we're not blocking on it (at least for now)" is probably the most useful take-away
[07:47:08] <Jeffrey Altman> I would love to see such a fix applied to a 1.9 series release for field testing.
[07:48:02] <wiesand> Bringing us back to my question regarding plans for 1.9...
[07:48:32] <Jeffrey Altman> I ended up working through the holiday so didn't get the prep work I wanted to accomplish for 1.9.
[07:49:51] <wiesand> prep work on Windows?
[07:50:01] <Jeffrey Altman> In particular, I want to try to summarize the differences between master and 1.6.x at this point so we know what is being targeted for 1.9/1.10 and what work needs to be accomplished to be able to start issuing bi-weekly releases
[07:50:19] <Jeffrey Altman> not for windows.   
[07:50:35] <Jeffrey Altman> for windows I know exactly what needs to be done.
[07:51:17] <Jeffrey Altman> the libtool changes have impacts on packaging and I believe there are still places where libtooling is not complete.
[07:51:52] <Jeffrey Altman> there is work that is needed to stop distributing lwp versions of binaries, etc.
[07:51:55] <kaduk@jabber.openafs.org/barnowl> There are, yes. :-/
[07:52:37] <Jeffrey Altman> there are many patches in gerrit that should probably hit master before 1.9 is cut
[07:53:01] <kaduk@jabber.openafs.org/barnowl> Feel free to call them out for review, &c..
[07:53:49] <Jeffrey Altman> that is part of the document I started on but haven't made much progress on.
[07:57:07] <wiesand> Sounds like "no ETA"?
[07:57:20] <Jeffrey Altman> I have no ETA.  
[07:58:23] <Jeffrey Altman> I believe that wiesand should move ahead with pre2.   We can try to get additional testing but December tends to be very difficult for that.   
[07:59:31] <wiesand> At this point, 1.6.6 can probably wait until around the Linux 3.13 release date.
[08:00:03] <wiesand> In any case, I expect meaningful testing to happen after the release, as usual.
[08:00:29] <Jeffrey Altman> of course.   
[08:01:32] <Jeffrey Altman> the real testing is performed by those in this conversation and customers of support organizations that are suffering from a bug that is being fixed by the release.   
[08:02:41] <wiesand> And those who deploy the release but aren't support customers.
[08:02:59] <Jeffrey Altman> but not until after the release
[08:03:14] <wiesand> Exactly, *afterwards*.
[08:04:21] <wiesand> Let's see what pre2 gets us first.
[08:04:48] <Jeffrey Altman> its hard to ask a customer to test something that hasn't been released.   The response is usually "we pay you for support so that you will test it".    If the customer has a bug it is much easier to ask them to test the fix for their issue since they have a personal interest.
[08:06:11] <wiesand> Well, we fix a few bugs with 1.6.6.
[08:06:22] <Jeffrey Altman> one of the reasons we want to get to more frequent minor releases with fewer changes is to reduce the risk associated with deploying each release.
[08:07:21] <wiesand> But you also said customers prefer deploying at most twice a year.
[08:07:52] <Jeffrey Altman> and delayed by six months after a release :)
[08:08:47] <Jeffrey Altman> everyone wants to let the rest of the community find the bugs and do the testing.   
[08:09:23] <Jeffrey Altman> anyway, lets get pre2 finished.  if there is nothing else we can release it for the holidays as a present.
[08:09:30] <Jeffrey Altman> place a nice bow on it.
[08:09:46] <wiesand> Ok.
[08:09:47] <meffie> yay, presents. 
[08:10:28] <wiesand> I hope to get it out this week, or early next week.
[08:10:40] <Jeffrey Altman> thanks.
[08:11:09] <Jeffrey Altman> who here is planning to attend the EAKC at CERN?
[08:11:25] <wiesand> I'll go.
[08:11:42] <wiesand> Already have the travel permit.
[08:11:55] <deason> mike and myself
[08:12:17] <wiesand> Good.
[08:12:30] <Jeffrey Altman> A tour of the LHC is planned but the spots are extremely limited.   The slots will be given to those that submit talks and register on a first come first served basis.
[08:13:08] <Jeffrey Altman> so get your abstracts submitted.  registration will be open in a couple of weeks.
[08:13:31] <wiesand> We could do a release-team meeting live on stage ;-)
[08:13:42] <Jeffrey Altman> it will be a wednesday :)
[08:13:54] <kaduk@jabber.openafs.org/barnowl> I still need to apply for travel.
[08:14:11] <Jeffrey Altman> perhaps live pre-release testing?
[08:14:34] <meffie> we can ask the audience for gerrit reviews.
[08:14:44] <Jeffrey Altman> there is a long history of releasing a development release a day during conferences 
[08:15:14] <wiesand> "Hackathons"?
[08:16:16] <wiesand> Looking forward to seeing many of you there.
[08:16:25] <wiesand> But I believe we're finished for today?
[08:16:45] <Jeffrey Altman> I move that we be done
[08:17:02] <Jeffrey Altman> switching to "openafs" room to response to kaduk
[08:17:40] <wiesand> Ok, thanks a lot everyone.
[08:17:57] <wiesand> Bye.
[08:18:02] --- wiesand has left
[08:19:27] --- Marc Dionne has left
[08:33:58] --- meffie has left
[08:33:59] --- meffie has become available
[08:34:27] --- meffie has left
[09:19:58] --- deason has left
[09:19:58] --- deason has become available
[10:27:04] --- deason has left
[10:27:37] --- deason has become available
[13:00:13] --- deason has left
[13:00:14] --- deason has become available
[13:03:42] --- deason has left
[13:03:43] --- deason has become available
[15:56:21] --- deason has left