- Wednesday, April 22, 2015

release-team@conference.openafs.org

Wednesday, April 22, 2015< ^ >

Room Configuration

Room Occupants

GMT+0
[08:34:05] Jeffrey Altman leaves the room
[08:34:05] Jeffrey Altman joins the room
[13:40:36] meffie joins the room
[13:44:23] kaduk joins the room
[13:45:00] <Jeffrey Altman> http://workshop.openafs.org/afsbpw15/cfp.html
[13:55:53] wiesand joins the room
[13:57:37] <wiesand> Is https://desypad.de/ysM1G2UgB8 accessible from outside?
[13:58:04] <meffie> i see it
[13:58:29] <wiesand> Good. If you spot a problem, please simply correct it ;-)
[13:58:39] <Jeffrey Altman> oh that is neat
[14:01:52] <wiesand> Seems to work :)
[14:05:23] <wiesand> So we’re at the “1.6.12pre1” topic. I released the volumes, the announcement is being worked on under the URL above.
[14:05:59] <wiesand> Daria, are you here? If so, could you tag c8d6d91dfca8c19f1c101ee6987141fd601ea2c3 as openafs-stable-1_6_12pre1 please?
[14:10:12] <Jeffrey Altman> how does that look?
[14:10:32] <wiesand> I like that version better than mine. Thanks a lot.
[14:11:40] <wiesand> Fixed one more leftover 1.6.11. I think it’s ok now. I’ll sen that to -announce once we have the tag.
[14:14:15] <wiesand> And that will be pre1 then.
[14:14:34] <wiesand> So what about pre2? Seems like we’ll have one.
[14:15:08] <wiesand> Thanks for the reminder of that GPLONLY problem. Does it seem likely that this can be sorted out?
[14:16:04] <shadow@gmail.com/barnowlE5B64A04> marc is trying
[14:16:22] <wiesand> Valiant.
[14:17:31] <wiesand> Does anyone know why this yosemite-cherrypicking stack apparently became needed?
[14:17:47] <shadow@gmail.com/barnowlE5B64A04> new compilers
[14:17:58] <shadow@gmail.com/barnowlE5B64A04> apple updated xcode
[14:18:12] <Jeffrey Altman> I believe the current status is.  The author of the commit that broke us is not going to revert or alter the change.  Nor is s/he going to submit a change on our behalf to remove the GPL_ONLY label from the symbol in question.   Instead we have been asked to research all of the affected non-GPL_ONLY interfaces and submit patches to each of them
[14:18:14] <wiesand> guessed that, thanks.
[14:18:41] <shadow@gmail.com/barnowlE5B64A04> linux kernel core is useless, film at 11
[14:19:27] <wiesand> sounds bad
[14:19:42] <Jeffrey Altman> its politics so yes its bad
[14:19:51] <Jeffrey Altman> or perhaps, its religion so ....
[14:20:21] <wiesand> probably both, and that’s the worst combination imaginable...
[14:20:51] Marc Dionne joins the room
[14:21:00] <wiesand> So, we should have those yosemite/clang fixes in pre2.
[14:21:05] <Jeffrey Altman> ah there is Marc.  Express your sympathies
[14:21:21] <Marc Dionne> reading back..
[14:21:34] <Jeffrey Altman> The Yosemite fixes can be pulled in.
[14:22:05] <wiesand> Well, not without a pull-up of 11846 or something similar, right?
[14:22:12] <kaduk> Right.
[14:22:13] <shadow@gmail.com/barnowlE5B64A04> they are all correct. so, yes
[14:23:35] <Jeffrey Altman> well, the other Yosemite changes can go in.  They aren't blocked on 11846
[14:23:54] <Marc Dionne> actually what was suggested was just to submit a patch changing the one problem symbol to non-GPL (so a one liner), and the developer said he would ack it.  I wanted to research the extent of the functions affected a bit in case there were questions about "why are you doing that"..
[14:24:36] <Jeffrey Altman> k
[14:24:49] <Jeffrey Altman> that sounds much more reasonable
[14:25:14] <Marc Dionne> i've been able to confirm that at least the Broadcom wifi driver and Nvidia will have the same issue
[14:25:21] <kaduk> Yay, I appear to have successfully un-locked myself out of gerrit.  Apparently Postgres SQL really wants you to use single quotes, which took a while to figure out.
[14:25:47] <Jeffrey Altman> oh? you are messing with the database?
[14:25:53] <Marc Dionne> my account(s) will need help
[14:26:19] <wiesand> April the 20th has passed?
[14:26:20] <Jeffrey Altman> and my duplicate needs destruction :)
[14:26:28] <meffie> heh, i now have two accounts as well.
[14:26:44] <wiesand> me too...
[14:27:02] <kaduk> Yeah, April the 20th has passed.  Apparently I was not paying enough attention between the 17th and the 20th to do something about it.
[14:27:09] <Jeffrey Altman> anyone else going to buy Ben a beer?
[14:27:15] <Marc Dionne> it stopped working at some point yesterday afternoon
[14:27:39] <wiesand> Jeff: me. And Marc too.
[14:27:50] <kaduk> Adding a new ID to an existing account is not so bad; just insert one line into one table (account_id, email_address, external_id)
[14:28:31] <kaduk> Removing duplicate accounts is much more work, there are usually four tables that need to have things removed, and I like to sanity check that there isn't anything for that id in other tables as well.
[14:28:52] <meffie> > anyone else going to buy Ben a beer?
yes, i will, maybe even two.
[14:28:56] <wiesand> Yes, I learned that *after* I managed to create that duplicate :-(
[14:29:18] <kaduk> I haven't figured out merging accounts, yet.  I suspect it involves UPDATE, which I've not used before.
[14:29:42] <Jeffrey Altman> All that I think needs to happen for my duplicate is to remove the google identity and e-mail and replace it with something bogus
[14:29:49] <wiesand> It’s still working for me though (I created another u1 account and linked that)
[14:30:07] <kaduk> Anyway, people should probably send email with what they want to happen.
[14:30:16] <Jeffrey Altman> I removed the google identities from my primary account
[14:30:35] <meffie> i played a game of towers of hanoi with openid accounts to move my email back to the original gerrit id.
[14:30:59] <Jeffrey Altman> ouch
[14:31:39] <kaduk> ouch, indeed.
[14:31:55] <kaduk> Anyway, that done, I am trying to do a (hopefully last) review of 9985.
[14:31:56] <Jeffrey Altman> for 1.6.12 Marc is going to be pushing a replacement for http://gerrit.openafs.org/#change,11124
[14:32:31] <Marc Dionne> once I can sign in to my account... :)
[14:32:43] <Jeffrey Altman> sigh
[14:32:59] <kaduk> Shouldn't ssh keys still work even if web login doesn't?
[14:33:07] <Jeffrey Altman> good point
[14:33:08] <Marc Dionne> probably
[14:33:27] <Marc Dionne> so i should be able to push, just not comment through the web
[14:33:35] <kaduk> Yeah.
[14:33:44] <kaduk> But you can do more than just push over ssh, IIRC.
[14:33:52] <kaduk> I forget if comment is such a thing, though.
[14:34:15] <kaduk> Anyway, since I had a launchpad.net account sitting around, I just used that, and it worked okay.
[14:35:49] <Marc Dionne> prob is I logged in before doing the "link", so then it wouldn't let me link it because it already existed.  i suppose i should've tried using another openid to get around that (prob what Mike did)
[14:36:26] <wiesand> Daria: Thanks for the tag. I just sent the announcement.
[14:36:30] <kaduk> Maybe.  I did that, too, when I wanted to link another google account
[14:37:09] <kaduk> (the logging in first part, that is)
[14:37:13] <Marc Dionne> a web UI option for "nuke this account" would be useful
[14:37:38] <Jeffrey Altman> I'm happy with 9985.
[14:37:44] <kaduk> I think newer versions of gerrit have a command that admins can run, to delete an account, but I don't remember anything web-accessible.
[14:38:46] <kaduk> I was just rediscovering that FilepathNormalize(NULL) is not harmful.
[14:40:19] <kaduk> I have lost track of where we are in the agenda -- wiesand, are we at the 1.8 topic already?
[14:40:53] <Jeffrey Altman> or we could go back three weeks and discuss windows code signing.   I'm going to become a pumpkin in 20 minutes
[14:41:38] <kaduk> Don't get made into pie ;)
[14:41:45] <wiesand> Fine, let’s go on to 1.8 or whatever.
[14:41:55] <kaduk> Anyway, it sounds like we can probably merge 9985 today.
[14:42:23] <Jeffrey Altman> Ben, if you +1 it I will submit it
[14:42:46] <kaduk> working on it
[14:43:20] <kaduk> After that, I guess we should (continue to) look at Mike's logging changes for next week.
[14:43:56] <meffie> softsig?
[14:44:06] <kaduk> Oh, right, that was first.
[14:44:21] <kaduk> So, softsig for this week; logging for next week (maybe?).
[14:44:31] <meffie> ok
[14:44:45] <shadow@gmail.com/barnowlE5B64A04> softsig: sigh
[14:44:48] <kaduk> Mike also got to test out 11786 (akeyconvert); I don't know if we want to give that more testing before merging it.
[14:44:51] <wiesand> What number was softsig?
[14:44:59] <shadow@gmail.com/barnowlE5B64A04> kill linux24. move on. yay softsig
[14:45:22] <kaduk> 6947 is softsig
[14:45:29] <wiesand> Linux24 was killed last week.
[14:47:03] <Jeffrey Altman> Ben, after 9985 is merged please rebase 6947 so it sits on top of the Linux24 removal
[14:47:40] <kaduk> *nods*
[14:47:48] <kaduk> ugh, I found a bug in 9985
[14:47:51] <Jeffrey Altman> thanks
[14:48:08] <kaduk> "but not in any code that anyone is really going to care about"
[14:48:08] <Jeffrey Altman> where?
[14:48:28] <kaduk> update/client.c; there's a strdup(path1) that should be strdup(path2)
[14:48:50] <kaduk> in the NT40 block...
[14:49:15] <Jeffrey Altman> yes there is
[14:49:17] <meffie> i care... oh
[14:50:51] <kaduk> Okay, so it sounds like we've got a plan, and Jeff has 10 minutes of non-pumpkin time left.
[14:51:32] <kaduk> Any gatekeeper thoughts about 11786?
[14:52:48] <Jeffrey Altman> it would be nice to have tests to go with it :)
[14:54:13] <Jeffrey Altman> I will add some comments on the man page
[14:54:17] <kaduk> I guess that's fair.
[14:54:34] <Jeffrey Altman> do you have specific implementation design questions?
[14:54:55] <kaduk> It's using hand-rolled argument parsing, of course, so adding a config directory argument involves a bit of work, but is probably still worth it.
[14:55:26] <Jeffrey Altman> is there a reason you didn't use the libcmd ?
[14:55:41] <kaduk> I don't remember any particular questions offhand, but it's been a while since I was in the code.
[14:55:49] <Jeffrey Altman> a -config option is definitely useful
[14:55:57] <kaduk> Mostly because I was using aklog as a reference/starting point :-/
[14:56:43] <Jeffrey Altman> any reason you are using the IPL10 for this submission since it is new code?
[14:57:54] <kaduk> Oh, for the man page? No reason.
[14:58:02] <kaduk> Well, actually, hrrm.
[14:58:38] <kaduk> I guess there's probably not enough similarity to the KeyFile manpage to worry about derivative work, so I should relicense it, yes.
[15:01:12] <wiesand> Ok, Jeff ought to be a pumpkin now, and it’s getting late for me too. Anything else to discuss today?
[15:02:10] <Jeffrey Altman> The only other thing that comes to mind is the age of the rxkad.keytab and whether some mention of when it is advisable to import non-latest keys is useful.
[15:02:47] <Jeffrey Altman> if there are no outstanding service tickets with an old kvno, there is no reason to keep the key around
[15:03:06] <Jeffrey Altman> and yes, I must go
[15:03:20] <kaduk> Okay, thanks.
[15:03:49] <Jeffrey Altman> Last comment for the day is please come vacation with your families in Pittsburgh the week of August 17th
[15:04:45] <Jeffrey Altman> If you have questions for those that designed and wrote this stuff we work on, this might be a once in a lifetime opportunity to ask them
[15:05:16] <Jeffrey Altman> http://workshop.openafs.org/afsbpw15/family.html
[15:05:23] <Jeffrey Altman> and with that I am a pumpkin
[15:05:50] <wiesand> Bye Jeff. Thanks.
[15:06:03] <meffie> lovely pittsburgh: https://plus.google.com/photos/yourphotos?pid=6139440575045854498&oid=113901143137527090881
[15:07:30] <wiesand> Am I supposed to see more than a shadow at midnight?
[15:07:48] <meffie> oh, sorry.
[15:08:38] <meffie> https://plus.google.com/113901143137527090881/posts/9Y3qtEdepoN?pid=6140580437269577218&oid=113901143137527090881
[15:09:27] <wiesand> Better :) Are you sure the cherry trees will blossom in August though?
[15:10:37] <meffie> maybe not.
[15:11:39] <meffie> have a good evening.
[15:11:55] <wiesand> Fine, let’s call it a day. Thanks a lot everyone!
[15:12:08] meffie leaves the room
[15:12:28] <wiesand> I should manage to send minutes later. Goodbye.
[15:12:32] wiesand leaves the room
[15:12:45] Marc Dionne leaves the room
[15:15:11] kaduk leaves the room