[05:03:36] --- Stephan Wiesand has become available
[06:24:21] --- Marc Dionne has become available
[06:34:07] <Stephan Wiesand> Hello Marc. Any bad news regarding Linux 3.12?
[06:34:20] <Marc Dionne> Hi Stephan
[06:35:08] <Marc Dionne> Just tested this morning with the 1_6_5_x branch, adding the proposed commits, on fedora 19 with current kernel git (3.12-rc3+).  Looks good
[06:35:18] <Stephan Wiesand> :)
[06:35:38] <Stephan Wiesand> Thanks a lot!
[06:36:19] <Marc Dionne> the gcc warnings are not a problem if you build without --enable-warnings and --enable-checking (which I normally use)
[06:43:37] <Stephan Wiesand> I'd really like to keep the list of changes for 1.6.5.1 short. But I think that one would be ok if it's considered important. The only actual change is the "*aCL = NULL" hunk, right?
[06:44:48] <Marc Dionne> gcc 3.8 complains about that, also unused typedefs, and also a memcpy that is an actual bug (have to locate it again)
[06:45:39] --- Derrick Brashear has become available
[06:47:44] <Stephan Wiesand> Hmm, I can't find a memcpy in 10242, and it's not mentioned in the commit message.
[06:47:49] <Stephan Wiesand> Master only, maybe?
[06:48:26] <Marc Dionne> it's fiexd in master with 7076, it's a memset actually
[06:48:52] <Marc Dionne> commit 34afb3e6b903fc
[06:49:56] <Stephan Wiesand> That's 10243 on 1.6.x
[06:50:20] <Marc Dionne> ok, wasn't sure if it had been pulled up
[06:51:04] <Derrick Brashear> I'll be slightly slow on the draw today fwiw. Operating ~1handed
[06:51:37] <Stephan Wiesand> Marc: apparently, you did ;-)
[06:51:42] <Stephan Wiesand> Hello Derrick
[06:51:56] <Marc Dionne> Stephan: ok, short memory.. :)
[06:53:06] <Marc Dionne> Derrick, yeah that cast/splint can't be great for typing...
[06:53:36] <Stephan Wiesand> Ouch. What happened?
[06:53:46] <Derrick Brashear> Bike wreck
[06:54:11] <Stephan Wiesand> Very sorry to hear that.
[06:54:17] <Derrick Brashear> Broke one bone in my hand. That's it
[07:00:51] <Stephan Wiesand> My EL5 build of the proposed 1.6.5.1 just finished successfully.
[07:01:11] --- meffie has become available
[07:02:04] <Stephan Wiesand> Hello Mike
[07:02:48] <Derrick Brashear> If you want a tag pushed I can get that within the hour but I need to reshuffle sandboxes first
[07:03:00] --- Derrick Brashear has left
[07:03:25] <meffie> hello. 
[07:03:33] --- Derrick Brashear has become available
[07:04:48] <deason> ping?
[07:05:08] <Derrick Brashear> Echo reply
[07:05:19] <Stephan Wiesand> Derrick: Thanks. Nothing has been merged yet. Buildbot isn't active on that branch, so we should have at least a few success reports before.
[07:05:20] <deason> hmm, I can't see anyone else, but apparently messages are fine
[07:05:30] <Stephan Wiesand> Hello Andrew.
[07:05:56] <Derrick Brashear> I'll build macos shortly then
[07:06:37] <Stephan Wiesand> I see Derrick, Marc, Mike and Jeffrey (who hasn't spoken up yet though).
[07:07:19] <Stephan Wiesand> Thanks Derrick (Macos is not a 1.6.5.1 target platform though?)
[07:07:52] <Derrick Brashear> No, but might as well confirm no breakage
[07:08:16] <Stephan Wiesand> Right, thanks.
[07:09:17] <Stephan Wiesand> Ok, I guess we won't become more today. Let's start.
[07:09:58] <Stephan Wiesand> http://gerrit.openafs.org/#q,status:open+project:openafs+branch:openafs-stable-1_6_5_x,n,z
[07:10:13] <Stephan Wiesand> Anything else that should go into 1.6.5.1?
[07:10:58] <Stephan Wiesand> Marc mentioned 10242 and 10243
[07:11:05] <Derrick Brashear> Looks fine to me
[07:11:58] <Derrick Brashear> Lib admin. I'm indifferent
[07:12:10] <Marc Dionne> but not strictly necessary unless you enable the extra warnings and checking
[07:12:10] <deason> 10242 and 10243 are fine by me
[07:13:01] <Stephan Wiesand> I'd rather leave those to 1.6.6, unless anyone feels strongly about having them in 1.6.5.1.
[07:13:08] <Derrick Brashear> I bet Xcode 5 (clang not gcc) gives a slew of new warnings
[07:13:11] <Marc Dionne> users who have a recent kernel (ex: fedora 19) are also likely to have a newer gcc (ex: 4.8)
[07:13:42] <Stephan Wiesand> Are they likely turn on the extra checks?
[07:14:20] <deason> I don't think non-developers are likely to know they exist
[07:14:20] <Marc Dionne> don't know - i would think most users who compile don't
[07:14:58] <Stephan Wiesand> And those who do should have filed a bug against 1.6.5.
[07:16:57] <Marc Dionne> i think it's fine either way, take them or not
[07:17:06] <Stephan Wiesand> It builds on EL6 too. I have to find machines to actually run it on, but should be able to do that tomorrow.
[07:17:41] <Stephan Wiesand> Marc tested F19 with 3.12-rc3, Derrick will build on Macos.
[07:17:45] <Derrick Brashear> Is Derek able to try on fedora?
[07:18:04] <Derrick Brashear> Ok. Marc did
[07:18:17] <Marc Dionne> yeah, tested this morning
[07:18:24] <Stephan Wiesand> Anyone else who could contribute testing?
[07:18:32] <Stephan Wiesand> (what else do we need?)
[07:22:31] <Stephan Wiesand> Hmm... ok, F19 + EL5/6 should suffice for merging the changes. Then provide preliminary tarballs and ask for testing on the release-team list?
[07:23:32] <Stephan Wiesand> (and Macos, of course)
[07:23:52] <Derrick Brashear> Sure
[07:24:04] <meffie> ok
[07:24:41] <Stephan Wiesand> This brings us to the merge policy for the 1_6_5_x branch. Do we insist on +1s for the pullups?
[07:25:18] <Derrick Brashear> You control the horizontal 
[07:25:29] <Derrick Brashear> you control the vertical
[07:26:06] <meffie> (thanks for not calling us monkeys again)
[07:26:24] <Stephan Wiesand> please explain...
[07:26:33] <Stephan Wiesand> (I'm dumb again)
[07:26:46] <Marc Dionne> missing cultural reference? :)
[07:27:03] <Stephan Wiesand> I hope it's just that :-)
[07:27:04] <deason> it's just another way of saying "it's your circus"
[07:27:05] <Derrick Brashear> Colloquialism. Means your call
[07:27:23] <deason> I think you've already got agreement from in here, though
[07:27:53] <Stephan Wiesand> Thanks. I'll take that as a "not strictly required".
[07:28:19] <Stephan Wiesand> Reviews of 10316/7 would be nice though.
[07:29:36] <meffie> ok
[07:29:48] <deason> sure, those make more sense, as they aren't cherry picks :)
[07:30:14] <Stephan Wiesand> I take it noone here insists on buildbot verification either?
[07:30:52] <Derrick Brashear> For what
[07:31:10] <Derrick Brashear> We never merge without it or knowing why it failed
[07:31:21] <Jeffrey Altman> I think Stephan is asking about buildbot for 1.6.5.1
[07:31:32] <Stephan Wiesand> Hello Jeffrey. Right.
[07:32:01] <Derrick Brashear> Oh. For this, don't care. It's same as 1.6.x.
[07:32:08] <Jeffrey Altman> It isn't necessary.   The changes are supposed to be single platform and have already passed master and 1.6.x
[07:32:19] <Stephan Wiesand> Ok, thanks.
[07:33:55] <Stephan Wiesand> I think this concludes 1.6.5.1. The rest should be straightforward and rather swift.
[07:35:23] <Stephan Wiesand> On to 1.6.6. Anything on RT #131737 ?
[07:36:42] <deason> if that actually is a problem, I think it's been there since 1.6.0
[07:37:01] <deason> I'm not sure what he needs that header for, but we can talk to him
[07:37:27] <deason> my opinion is, don't block onit
[07:37:32] <meffie> +1
[07:37:55] <deason> (immediate cause is switching from 'struct AFS_UCRED' -> afs_ucred_t; the former works without the struct definition, but the latter doesn't work without a typedef)
[07:38:48] <Stephan Wiesand> Ok, won't block on it. I guess few have this problem, and todd has a workaround.
[07:40:35] <Stephan Wiesand> Open changes in gerrit... which ones of the top 6 from http://gerrit.openafs.org/#q,status:open+project:openafs+branch:openafs-stable-1_6_x,n,z are for 1.6.6?
[07:41:44] <Stephan Wiesand> Or, which clearly aren't?
[07:42:49] <Jeffrey Altman> the parallel build changes and doc changes might as well go in.   
[07:42:59] <deason> top 5 are for 1.6.6
[07:43:07] <Jeffrey Altman> I don't think aklog should be a blocker
[07:43:49] <deason> 10309 10310 and 10275 are unlikely to break anything besides the build failing
[07:43:53] <Derrick Brashear> Aklog is "would be nice" IMO
[07:44:19] <deason> 10314 is just a doc change
[07:45:45] <deason> 10274 is the only one that could conceivably cause a runtime bug, but it's pretty simple
[07:46:23] <Stephan Wiesand> Well, pre1 is meant to be tested. So why not.
[07:46:50] <deason> 10310 also contains a slight divergence from master, but it probably doesn't matter
[07:46:59] <deason> (that is, "as-is")
[07:51:35] <Stephan Wiesand> So 10310 should be cleaned up -  wait for that before merging it?
[07:52:31] <meffie> one sec.
[07:52:55] <deason> mike says he'll get it
[07:53:01] <deason> should just take a moment
[07:53:09] <Stephan Wiesand> Great.
[07:53:33] <Stephan Wiesand> Anything else yet to pull up?
[07:54:01] <Jeffrey Altman> there will always be something else tomorrow :)
[07:54:21] <Stephan Wiesand> Of course. We have a few days - builders will be busy with 1.6.5.1 for that time anyway.
[07:55:41] --- Derrick Brashear has left
[07:56:33] <meffie> (10310 pushed)
[07:57:51] <Stephan Wiesand> Thanks Mike. Once that's verified and reviewed, 10274/5, 10309/10, 10314 seem ok to merge then.
[07:57:59] --- kaduk has become available
[07:58:26] <kaduk> I hope to have patches to build on freebsd 10.0 in the next few days, ideally those could get into 1.6.6, too.
[07:58:55] <Stephan Wiesand> Hello Ben.
[07:59:43] <kaduk> Sorry I'm late.
[08:00:05] <Stephan Wiesand> If they touch FBSD only, there has been no dispute in the past.
[08:00:55] <kaduk> I think I have a couple of cleanups outside afs/FBSD, but those may only be needed on master.
[08:01:55] <kaduk> (We are using a few headers as <foo.h> that are supposed to be used as <sys/foo.h> or <machine/foo.h> on FreeBSD, and the kernel module build glue doesn't seem to like just adding extra -I arguments to the compiler.)
[08:02:44] <Stephan Wiesand> Let's discuss them once they're ready. If it's just cleanups, I wouldn't want to block 1.6.6 on them though.
[08:03:06] <kaduk> Sure.  I'll work on getting actual concrete things to talk about.
[08:05:00] <Stephan Wiesand> Ok. In the next days, I'd like to prioritize getting 1.6.5.1 out. Then let's see what else has come up for 1.6.6.
[08:05:41] <Stephan Wiesand> Tomorrow's a public holiday here. I should be able to get 1.6.51 to the stage of preliminary tarballs (w/o) tag.
[08:05:54] <Stephan Wiesand> (and catch up on the minutes - shame on me...)
[08:06:01] <Jeffrey Altman> freebsd support can always be 1.6.6.1 if the changes are small and platform specific
[08:07:30] <Stephan Wiesand> Jeff: Ok. Let's see what they actually look like.
[08:07:40] <Stephan Wiesand> Anything else to discuss today?
[08:08:36] <Stephan Wiesand> Thanks a lot for your participation. And the +1's etc in gerrit.
[08:09:02] <Stephan Wiesand> Bye.
[08:09:08] --- Stephan Wiesand has left
[08:10:23] --- kaduk has left
[08:12:25] --- meffie has left
[08:16:22] --- Marc Dionne has left
[12:47:46] --- deason has left
[12:47:46] --- deason has become available
[15:28:57] --- deason has left