Home
release-team@conference.openafs.org
Friday, June 8, 2018< ^ >
Room Configuration
Room Occupants

GMT+0
[03:26:28] mvita leaves the room
[11:52:48] kaduk@jabber.openafs.org/barnowl leaves the room
[11:53:03] kaduk@jabber.openafs.org/barnowl joins the room
[11:53:14] <kaduk@jabber.openafs.org/barnowl> [jabber client roulette ... or, can I send today?]
[11:53:28] <kaduk@jabber.openafs.org/barnowl> answer: yes, but only if I use the right outgoing account.
[11:53:46] Marcio Barbosa joins the room
[12:53:36] meffie joins the room
[12:54:07] <meffie> good morning
[12:54:14] <kaduk@jabber.openafs.org/barnowl> greetings
[12:54:22] wiesand joins the room
[12:54:25] wiesand leaves the room
[12:54:38] wiesand joins the room
[12:54:43] <kaduk@jabber.openafs.org/barnowl> I think you should look at 13121 again
[12:54:53] <wiesand> Hello
[12:54:59] <wiesand> https://drive.google.com/file/d/1MjD2TUdMV02_RTBwLHRSCPpcsgeRWTHZ/view?usp=sharing
[12:55:25] <kaduk@jabber.openafs.org/barnowl> !!!
[12:55:39] <wiesand> This shall be my new home
[12:55:54] <kaduk@jabber.openafs.org/barnowl> Congratulations!  It's a beautiful site
[12:56:08] <meffie> 13121?
[12:56:42] <wiesand> and it's less than half a mile from my workplace
[12:56:58] <meffie> congrats!
[12:57:07] <kaduk@jabber.openafs.org/barnowl> Bonus.  I'm quite spoiled being less than 30 meters from my workplace
[12:57:57] <wiesand> So that's what I've been occupied with lately, to the disadvantage of openafs.
[12:58:36] <wiesand> 30 meters? isn't that a bit too close?
[12:59:13] <kaduk@jabber.openafs.org/barnowl> My "office" is the second bedroom of the two-bedroom apartment.  And
20 meters probably would work, too, but I didn't want to go measure.
[13:01:31] <wiesand> Ok. We're a bit oldfashioned here here in the old world, so I still am supposed to be present in the office during "core hours".
[13:02:30] <kaduk@jabber.openafs.org/barnowl> Yeah, it's somewhat unusual, and was specifically agreed upon during
hiring.  But since the nearest office is, IIUC, a 5-hour train ride
away, they could take me or leave me.
[13:03:35] Marcio Barbosa leaves the room
[13:03:36] mvita joins the room
[13:04:31] <mvita> Oh, wow, it's gorgeous!
[13:05:17] <wiesand> Thank you all.
[13:05:20] <mvita> about 8 metters for me
[13:05:24] <mvita> meters
[13:05:42] <wiesand> Ok, you win ;-)
[13:05:47] Marcio Barbosa joins the room
[13:06:17] <meffie> wiesand can host the next openafs hackathon now :)
[13:06:38] <mvita> YASSSS!
[13:07:13] <wiesand> No, it's definitely too tiny for that, unless you want to hack away on the lawn ;-)
[13:07:46] <meffie> you just know we will drink too much and wreck the place :)
[13:08:05] <wiesand> that's alright
[13:08:28] <mvita> >  Ok, you win
[13:08:34] <mvita> no, I think I lose….
[13:09:41] <mvita> business?
[13:10:02] <wiesand> yes, back on topic. sorry for distracting everyone
[13:10:17] <kaduk@jabber.openafs.org/barnowl> No need to apologize; it's a very exciting major life change!
[13:10:33] <mvita> no, we wanted to see - but there seemed to be a lull
[13:11:02] <wiesand> I don't have much to talk about today for obvious reasons.
[13:11:21] <wiesand> 13121 may need a 2nd look, I'm not sure
[13:11:46] <meffie> i'll take a look and retest. perhaps i didnt test enough.
[13:12:00] <wiesand> Ben merged tons of the 1.8 pullups - thanks!
[13:12:09] <meffie> yes, thank you.
[13:12:37] <kaduk@jabber.openafs.org/barnowl> wiesand: you made it super-easy with your table of "merge in this
order" and having the pullups all ready!
[13:13:28] <wiesand> Thanks, I'll try to keep that up.
[13:13:53] <wiesand> And I really should get that scripted in some way…
[13:14:14] <meffie> yes, i meant to say last week, i have started some tools that could help
[13:14:48] <meffie> i "ported" some shell scripts to python and put them on github and pypi
[13:15:18] <meffie> see https://github.com/meffie/git-gerrit
[13:15:48] <wiesand> Well I really prefer shell (and most everything else) over python, but I'll take a look ;-) Thanks.
[13:16:04] <kaduk@jabber.openafs.org/barnowl> perl5?  <tiny>perl4?</tiny>
[13:16:18] <meffie> (i remember perl4 :)
[13:16:19] <wiesand> maybe not perl4
[13:16:29] <kaduk@jabber.openafs.org/barnowl> phew
[13:17:10] <mvita> linux builder sadness (intermittent?):
[13:17:12] <wiesand> I'll also take care of pulling up the rest of public-api
[13:17:13] <mvita> checking if linux kbuild requires EXTRA_CFLAGS... yes
checking if linux kernel module build works... no
configure: error: in `/home/buildbot/openafs':
configure: error: Fix problem or use --disable-kernel-module...
See `config.log' for more details
Connection to 192.168.122.15 closed.
WARNING: Command failed with code 1: './configure'.
[13:17:48] <kaduk@jabber.openafs.org/barnowl> sadness indeed
[13:20:21] <meffie> also, Joe Gorse has some time starting next week for openafs work.
[13:20:30] <wiesand> good news
[13:20:54] <meffie> so, he could help with linux builds
[13:21:32] <wiesand> what's missing for 1.8.1pre1?
[13:21:40] <meffie> we expected our lucky streak would not last forever.
[13:21:49] <meffie> for linux builds that is.
[13:22:44] <kaduk@jabber.openafs.org/barnowl> 1.8.1pre1 I think just needs the public-api stuff on master to get
finished and pulled up, and NEWS.
[13:22:58] <kaduk@jabber.openafs.org/barnowl> Unfortunately I think I found a bug in one of the yet-to-be-merged
public-api changes on master
[13:23:12] <meffie> oh?
[13:23:33] <meffie> (did i miss something in a review?)
[13:23:34] <kaduk@jabber.openafs.org/barnowl> I don't think we need to export xdr_prname, since the prname is just a
fixed-length character array with no allocated storage.
[13:23:44] <kaduk@jabber.openafs.org/barnowl> Only if you expect to read your mail from one minute ago :)
[13:24:02] <meffie> ok
[13:27:04] <meffie> ben, did you say last week there is more public-api work to do for librx.a ?
[13:27:57] <kaduk@jabber.openafs.org/barnowl> Maybe I said something about libafsrpc (which is roughly the public
verison of librx)?  But I think I did go through that all; I just
forget if it was after the meeting or before.
[13:30:02] <meffie> oh, yes, libafsrpc.a. i started to dust off the collectd plugin i had written some time ago to see what was missing.
[13:30:06] <kaduk@jabber.openafs.org/barnowl> Thanks for looking at the early afsauthent ones in the stack, at least
-- it's reassuring to get a second pair of eyes before we commit to
exporting them in the public API
[13:30:19] <meffie> i remember doing a work around...
[13:31:31] <kaduk@jabber.openafs.org/barnowl> We also don't have to get the public API pristine and fully complete
for 1.8.1; I just want to get it in a vaguely reasonable and useful
for at least some things state.
[13:32:18] <meffie> ok, i just didnt want to miss the bus, in case it was the last one leaving, if you know what i mean.
[13:34:48] <wiesand> er, I don't - being dumb again…
[13:35:27] <wiesand> er, forget it
[13:37:03] <wiesand> 13132 could be replaced by a pullup from 1.8 now, and it still needs review
[13:37:42] <wiesand> again, I can pull it up to a new change
[13:37:55] <mvita> thank you Stephan
[13:39:42] <mvita> so the public-api items are high priority for 1.8?
[13:39:49] <meffie> yes
[13:40:07] <mvita> okay, I can take a look at those
[13:40:13] <meffie> (at least that's what i understand from ben)
[13:41:01] <kaduk@jabber.openafs.org/barnowl> Or at least some subset of them.  But it's hard to know exactly what
functions are being consumed without searching for consumers -- we ran
into python-afs at MIT, but there are probably other things that
haven't tried using 1.8 yet
[13:41:27] <mvita> hmm
[13:42:11] <kaduk@jabber.openafs.org/barnowl> the xdr_foo functions are maybe not critical, especially since IIRC
they are just for variable-length array typedefs, and not more
complicated (nested) structures.
[13:42:14] <meffie> regarding 13132, sorry for the mis understanding on the backport policy. it use to be patches for 1.6.x would be packported from master, but now we want them from the 1.8.x branch, right?
[13:42:22] <kaduk@jabber.openafs.org/barnowl> right
[13:42:41] <wiesand> right
[13:42:48] <meffie> seems we need to make that more clear somehow, in a backport policy statement
[13:43:06] <meffie> so we can point people to that.
[13:43:36] <wiesand> pull up from the nearest upstream branch only
[13:43:39] <meffie> i put that in the summary for today at least.
[13:43:42] <kaduk@jabber.openafs.org/barnowl> I suppose so.  Does it need to be on not-the-wiki?
[13:44:09] <meffie> i think the wiki is fine.
[13:44:27] <meffie> (the wiki always needs love)
[13:44:35] <wiesand> pullups to 1_6_x should be from1_8_x
[13:44:52] <wiesand> pullups to 1_6_22_branch should be from 1_6_x
[13:45:47] <wiesand> It doesn't speed things up, but it makes sure we won't miss anything.
[13:46:09] <mvita> +1
[13:46:12] <wiesand> it avoids any potential need for merges too
[13:46:19] <kaduk@jabber.openafs.org/barnowl> yup
[13:46:48] <wiesand> consent :)
[13:48:50] <meffie> i'll update das wiki
[13:48:55] <meffie> thanks!
[13:49:25] <meffie> (pat was bugging me to update the wiki too. maybe he can help with that)
[13:50:04] <mvita> maybe jgorse can help with that ;-)
[13:50:08] <meffie> no!!
[13:50:10] <meffie> :)
[13:50:24] <meffie> (last time he nuked all the pages :)
[13:50:48] <mvita> <meffie trolled>
[13:50:54] <meffie> (which would have been ok, if he would have written new ones)
[13:51:04] <wiesand> occasionally, changes pulled up from higher branches or maybe even not pulled up at all may make sense (as a reminder, poc,,… - Andrew used to do that in a few cases) - but they should immediately marked -1 or -2 explaining why
[13:51:32] <kaduk@jabber.openafs.org/barnowl> +1
[13:52:17] <meffie> ok. (btw, gerrit "drafts" could be used for that usecase)
[13:52:17] Marcio Barbosa leaves the room
[13:52:48] <mvita> speaking of the wiki - is there any documentation about the various OpenAFS libraries (static and dynamic) and how they are (intended to be) used?
[13:53:05] <meffie> i do not believe so.
[13:53:09] <mvita> I have a hard time keeping them straight, even with my notes
[13:53:10] <wiesand> drafts? another gerrit feature I don't know about
[13:53:54] <kaduk@jabber.openafs.org/barnowl> There's not really much documentation about the libraries, no.
[13:54:27] Marcio Barbosa joins the room
[13:54:34] <meffie> i've only read about gerrit drafts, but it seems to be for the case you push something to gerrit but just for feedback, it is not ready to merge yet.
[13:54:59] <kaduk@jabber.openafs.org/barnowl> https://packages.debian.org/sid/libafsauthent2 and
https://packages.debian.org/sid/libafsrpc2 have stanzas (presumably)
written by Russ which are thus somewhat authoritative -- afsrpc is the
core Rx and RPC library, and afsauthent is client authentication and
token management
[13:55:24] <kaduk@jabber.openafs.org/barnowl> jaltman showed me a change as a draft once and I was a little confused
that the rest of you couldn't see it (having not been added as
reviewers).
[13:55:48] <kaduk@jabber.openafs.org/barnowl> Oh, and I pushed some new commits to public-api for master, hopefully
addressing the issues I found.
[13:57:01] <meffie> i wish gerrit had a "public" star feature, or something to highlight stuff stephan and ben think are critical
[13:57:12] <mvita> kaduk: thank you for those links, they are very helpful
[13:57:30] <kaduk@jabber.openafs.org/barnowl> I think there's also some nuggets of wisdom in commit messages from
jaltman
[13:58:59] <mvita> kaduk: re: jaltman - do you mean about libraries?
[13:59:21] <kaduk@jabber.openafs.org/barnowl> Yeah, about the libraries' intended use.
[13:59:28] <mvita> ah, okay
[13:59:34] <kaduk@jabber.openafs.org/barnowl> e.g., e2ef4ce23d0cd0eddadfb39a39a29593093f0b6f
[14:00:55] <mvita> okay, thanks, that is good stuff
[14:01:18] <kaduk@jabber.openafs.org/barnowl> (I had discovered these while trying to figure out what to export for
the 'public-api' topic.)
[14:01:29] <mvita> <nod>
[14:01:31] <meffie> ah, thanks!
[14:01:51] <meffie> man, some in-tree docs would be nice :)
[14:02:15] <kaduk@jabber.openafs.org/barnowl> It would be nice, yes.
[14:07:07] <meffie> (and people in heck want ice water)
[14:07:08] <wiesand> Looks like we're finished for today?
[14:07:10] <meffie> ok
[14:07:24] <mvita> cool
[14:07:24] <kaduk@jabber.openafs.org/barnowl> probably
[14:07:25] <meffie> oh, wiesand
[14:07:27] <meffie> btw
[14:07:32] <wiesand> yes?
[14:07:58] <meffie> andrew's sysname patches are targeted for docker support, just in case that was not clear from the messages.
[14:08:10] <meffie> just to let you know.
[14:08:49] <wiesand> I have to admit I haven't looked at them yet, shame on me. But thanks for the heads up.
[14:09:24] <wiesand> Docker support sounds interesting. Any chance Andrew could look at singularity too?
[14:10:30] <meffie> singularity?
[14:12:06] <meffie> https://singularity.lbl.gov/ ?
[14:12:37] <wiesand> Singularity is an increasingly popular container solution at least in the HPC world. AFS "just works" with it, since your credentials inside a container are just the same as outside. The only drawback is that the sysname is always the one from the host.
[14:12:52] <wiesand> yes, that singularity :)
[14:13:03] <wiesand> amazing stuff, really
[14:14:32] <wiesand> the "next (or rather current) big thing"
[14:14:38] <meffie> heh, ok.
[14:15:07] <meffie> ok we can try to look.
[14:15:36] <meffie> ok, have a good weekend.
[14:15:51] <wiesand> If you have an idea how to change the "sysname" inside a singularity container, that would be great.
[14:16:13] <wiesand> And with that note, let's adjourn. Thanks a lot everyone!
[14:16:28] <kaduk@jabber.openafs.org/barnowl> thanks everyone!
[14:40:35] meffie leaves the room
[14:47:41] wiesand leaves the room
[15:19:20] Marcio Barbosa leaves the room
[15:21:28] Marcio Barbosa joins the room
[16:54:13] mvita leaves the room
[16:55:47] mvita joins the room
[17:28:01] Marcio Barbosa leaves the room
[17:30:09] Marcio Barbosa joins the room
[19:34:59] Marcio Barbosa leaves the room
[19:37:09] Marcio Barbosa joins the room
[21:00:21] mvita leaves the room
[22:03:27] Marcio Barbosa leaves the room