- Wednesday, August 20, 2014

release-team@conference.openafs.org

Wednesday, August 20, 2014< ^ >

Room Configuration

Room Occupants

GMT+0
[13:21:16] stephan.wiesand joins the room
[13:47:33] meffie joins the room
[14:00:35] <stephan.wiesand> Hello ... o  ... o ... o ...
[14:01:48] deason joins the room
[14:02:03] <deason> hiiii.... i..... i....
[14:02:52] <stephan.wiesand> I see every message thrice today.
[14:03:27] <meffie> hello hello hello
[14:03:28] <deason> I see you in the room twice, which may be contributing
[14:03:47] <deason> (as 'stephen.wiesand', and 'wiesand')
[14:04:45] <stephan.wiesand> Well, only this one is here. The other one is the client on my office desktop - but that's not running.
[14:05:52] <stephan.wiesand> I guess this will be a brief meeting :)
[14:06:12] <stephan.wiesand> Any comments on any item on the agenda?
[14:07:33] <deason> on "testing"....
[14:07:44] <stephan.wiesand> Regarding 11392: Jeffrey changed his verdict from -1 to no score. I'm not quite sure what to make of it...
[14:07:46] <deason> a few sites have responded to me really quickly this time, but it's too soon for any results
[14:08:02] <stephan.wiesand> Good!
[14:08:49] <stephan.wiesand> NB I'm going to deploy 5 DAFS fileservers with 1.6.10 in the next days.
[14:10:19] <stephan.wiesand> And I just uploaded the Solaris binaries.
[14:10:32] <deason> debian packages should also be hitting debian experimental soon
[14:10:53] <deason> (prepared by ben)
[14:10:57] kaduk joins the room
[14:11:08] <stephan.wiesand> Ah, fine. I was going to ask that ;-)
[14:11:14] <kaduk> You worked on them, too ;)
[14:12:50] <stephan.wiesand> Ben, we're to few today to walk through the agenda. Anything you want to talk about?
[14:12:59] <stephan.wiesand> 1.8 I guess, but...
[14:13:19] <kaduk> Nothing especially comes to mind.  Let me pull up the agenda to see if that jogs anything...
[14:13:57] <kaduk> 11384 and 11385 got merged on master
[14:14:28] <kaduk> So, "someone (TM)" should pull those up for 1.6
[14:15:12] <deason> are they needed for 1.6.10? or can they be patches added in fbsd packaging for a release?
[14:16:24] <kaduk> I mean, they affect any system where the named types are 32 bits.
[14:16:56] <kaduk> I didn't look super-carefully, but probably there are not runtime issues.
[14:18:20] <kaduk> I prefer not having to keep local patches in the packaging, of course.
[14:19:03] <deason> oh, but they're all code for debugging or for local tools; no real risk
[14:19:54] <kaduk> (So, I would probably not want to keep these as fbsd-local patches since they are comparatively minor.)
[14:21:20] <stephan.wiesand> They look fairly innocuous to me. I have no problem imagining them in pre2 if we have one. If we don't, I'd rather defer them to 1.6.11 unless you can spell out to me why they're 0 risk.
[14:22:22] <kaduk> Spelling out that they are zero-risk would require understanding the ABI of all of our supported platforms.  I can do that for linux and freebsd on x86/x86_64, but probably not all of our platforms.
[14:23:31] <deason> to me, it's near 0 risk because the code paths are not executed except for 'debugging' situations, or when running fsprobe
[14:23:39] <deason> and we don't really care if the printed value is wrong
[14:23:45] <meffie> fsprobe and cbd
[14:23:59] <meffie> i'm not sure fsprobe even works :)
[14:25:58] <stephan.wiesand> We usually do have a pre2. Let's revisit those when we have to. But yes, someone (TM) should pull them up in any case. I can do it if they're straightforward.
[14:26:12] <stephan.wiesand> NB what's fsprobe?
[14:26:44] <meffie> it's a testing/debugging tools for the fileserver.
[14:27:16] <deason> you can execute specific rpcs with it
[14:27:19] <meffie> kind of like the vl client tool.
[14:27:33] <stephan.wiesand> Ah, thanks.
[14:28:39] <stephan.wiesand> So yes, no risk there. But the other hunks seem to be for code usually run.
[14:29:27] <meffie> the one in callback.c ? that's only used in the cdb tool (callback dump debugger)
[14:29:58] <meffie> (see the #ifdefs.)
[14:31:14] <stephan.wiesand> Ok. And afsd_debug in afsd.c ?
[14:32:25] <kaduk> afsd -debug
[14:33:35] <stephan.wiesand> And casting something we printf %lu to unsigned long doesn't look risky either ;-)
[14:33:54] <stephan.wiesand> So, fine even w/o pre2, barring ordinary review.
[14:34:11] <kaduk> Fine by me
[14:34:44] <deason> pullups pushed, 11402-11404
[14:35:05] <stephan.wiesand> 3?
[14:35:10] <kaduk> Thanks
[14:35:51] <deason> one extra dependency; similar debug-only-ness
[14:36:23] <stephan.wiesand> Ok, thanks.
[14:36:42] <stephan.wiesand> Ben & Mike: your opinion on 11392 would be appreciated.
[14:37:17] <stephan.wiesand> We don't have to merge it today, but I'd like this to be fixed in 1.6.10 final.
[14:38:47] <meffie> ok, i'll look again. (i only glanced at it)
[14:39:24] <kaduk> Same here.
[14:41:21] <stephan.wiesand> Thanks.
[14:43:52] <stephan.wiesand> So there's some material for minutes today :)
[14:44:02] <stephan.wiesand> Anything else to discuss today?
[14:45:15] <stephan.wiesand> Taking this as a no.
[14:45:35] <stephan.wiesand> Thanks a lot for having been here today!
[14:45:43] <meffie> thank you stephan.wiesand
[14:46:12] stephan.wiesand leaves the room
[14:49:31] meffie leaves the room
[14:49:40] kaduk leaves the room
[14:50:13] deason leaves the room
[15:03:39] kaduk joins the room
[16:43:02] kaduk leaves the room