- Wednesday, September 9, 2015

release-team@conference.openafs.org

Wednesday, September 9, 2015< ^ >

Room Configuration

Room Occupants

GMT+0
[00:05:13] mvita joins the room
[00:13:09] shadow@gmail.com/barnowlABEE2063 leaves the room
[00:13:17] shadow@gmail.com/barnowlABEE2063 joins the room
[00:16:01] Jeffrey Altman joins the room
[01:30:37] shadow@gmail.com/barnowlABEE2063 leaves the room
[01:30:46] shadow@gmail.com/barnowlABEE2063 joins the room
[01:33:19] mvita leaves the room
[13:19:59] stephan.wiesand joins the room
[13:54:18] meffie joins the room
[13:55:18] Daria Brashear joins the room
[13:58:27] mvita joins the room
[14:00:26] <stephan.wiesand> Hi
[14:00:26] jhutz@jis.mit.edu/owl leaves the room
[14:00:35] <mvita> hello
[14:00:50] <stephan.wiesand> Today's image: https://www-zeuthen.desy.de/~wiesand/snf.jpg
[14:01:28] <stephan.wiesand> Have you ever seen a complete rainbow, with both ends touching the ground?
[14:01:46] <mvita> I think so, in western USA
[14:01:47] <Daria Brashear> i have.
[14:02:06] <Daria Brashear> well. i guess the most recent was "both ends touching water"
[14:02:09] <stephan.wiesand> I hadn't, before today.
[14:02:26] <mvita> that's a very nice photo
[14:02:59] <mvita> where was it taken?
[14:03:24] <stephan.wiesand> It's the view from the porch of the cottage I'm staying in. Snaefellsnes peninsula, Iceland.
[14:03:32] <meffie> nice!
[14:03:54] <Daria Brashear> i'll not rehash my icelandic name joke
[14:04:28] <stephan.wiesand> What you can't see, because the weather is quite bad, is the Snaefell volcano. Though it's quite close.
[14:05:03] <stephan.wiesand> Jules Verne's story about descending to the center of the earth starts there.
[14:05:56] <stephan.wiesand> Anyway, enough about my current setting. Though it may explain why I haven't been that active the past few days,
[14:06:08] <stephan.wiesand> Any Linux news?
[14:08:06] <stephan.wiesand> Ok, that probably means we should release 1.6.14.1 "as is".
[14:08:38] <stephan.wiesand> Ben made a proposal for rewording the NEWS change in gerrit (thanks).
[14:09:02] <stephan.wiesand> I'll follow that, and then push it out. Should happen tomorrow.
[14:09:19] <stephan.wiesand> What else?
[14:09:33] <stephan.wiesand> 1.6.15?
[14:09:35] <Daria Brashear> i can tag. lemme know what.
[14:10:21] <stephan.wiesand> Daria: thanks. I'll beg for the tag as soon as I've done my homework.
[14:10:55] <stephan.wiesand> I've been really slow the last week, but I guess I have a good excuse...
[14:11:16] <mvita> heh
[14:12:45] <stephan.wiesand> NB I'm left-handed today. Just didi the dishes. And my general purpose traveling knife's blade is really sharp...
[14:13:11] <mvita> eeg
[14:13:25] <stephan.wiesand> I'll survive.
[14:13:45] <stephan.wiesand> So, anything else to discuss? 1.6.15? 1.8?
[14:14:08] <stephan.wiesand> How about that C99 issue?
[14:14:30] <meffie> i suppose we can just fix the code to avoid it.
[14:14:47] <mvita> well re: c99
[14:14:49] <meffie> it's just a static array initialization for debugging
[14:15:14] <stephan.wiesand> Did I get this right that C99 features will remain unavailable for a long time due to Windows?
[14:15:15] <mvita> it would be nice to be able to verify (w/ buildbot, I suppose) whether a given c99 feature is supported
[14:15:29] <mvita> by all openafs platforms we intend to support
[14:16:10] <mvita> I think jaltman said not all c99 was in windows, but that some features were - iirc
[14:17:38] <mvita> ah, here's his direct quote:  "Most C99 features cannot be used on Windows."
[14:17:49] <Daria Brashear> push a change that uses it, see if it explodes, done
[14:17:56] <meffie> well, i'm sticking with ANSI C :)
[14:18:04] <Daria Brashear> C89 or GTFO
[14:18:15] <stephan.wiesand> K&R?
[14:18:18] <mvita> here we go
[14:18:26] <meffie> if it was good enough for k&r, it's good enough for me.
[14:18:58] <mvita> what is current OpenAFS standard?  is there one?
[14:18:58] <stephan.wiesand> I'm even old enough to have learned BASIC ;-)
[14:19:31] <Daria Brashear> i had a computer that did basic
[14:19:35] <stephan.wiesand> Mark: good question
[14:19:45] <mvita> looking
[14:20:15] kaduk joins the room
[14:20:23] <meffie> c89 by defacto
[14:20:27] <stephan.wiesand> Ah, Ben. Hi.
[14:20:30] <Daria Brashear> i also used a punch card system and did hello world in … i think it was ada, in grade school
[14:20:42] <mvita> ANSI (from my notes)
[14:20:43] <kaduk> Sorry I'm late.  I have an a/c repairman here, so I may have to drop off.
[14:20:52] <mvita> looking for wiki reference now
[14:21:39] <stephan.wiesand> Daria: You win. They obsoleted punch cards slightly before I started with computing.
[14:22:30] <kaduk> There is the question of how much we want to support IRIX/AIX/HP-UX with no buildslaves available.
I think jaltman has threatened in public that they will not necessarily be supported if no one steps up to provide a buildslave or similar support efforts.
[14:22:49] <kaduk> (I don't remember which of those three had trouble with C99 features.)
[14:22:50] <Daria Brashear> westinghouse had a research center just up the street from where i went to grade school
[14:24:17] <stephan.wiesand> IRIX/AIX/HPUX are sure all candidates for having trouble with C99. Are these targets for 1.8+ at all?
[14:24:40] <stephan.wiesand> They probably shouldn't be.
[14:24:43] <Daria Brashear> ALGOL. it was ALGOL
[14:25:08] <mvita> http://wiki.openafs.org/CodingStyle/     "Functions should be written with the type information on the first line, followed by the function name on the next, followed by the parameter list, in ANSI C format."
[14:25:13] <mvita> that's the only mention of ANSI
[14:25:23] <kaduk> If I can jump ahead to 1.8 and drop a thought or two in while I'm still in the room, I'd like another opinion or two on 11955, in terms of what approach is best to put the signed/unsigned barrier.
[14:25:26] <stephan.wiesand> I heard of ALGOL. And PL/1.
[14:25:28] <meffie> of those 3, probably AIX is still in use.
[14:25:44] <kaduk> Ah, ANSI as opposed to K&R...
[14:26:23] <stephan.wiesand> I think there's even some K&R style function declarations left in the code base.
[14:26:34] <meffie> yes there are.
[14:26:47] <stephan.wiesand> Amazing.
[14:26:47] <meffie> (not many tho)
[14:27:50] <kaduk> Another 1.8-potential-discussion is 11782; IIRC the path in question is not /usr/afs on windows, and I can't quite convince myself that just removing the check is correct, since otherwise the permissions on the non-leaf directory node would not be enforced in startup.
[14:29:38] kaduk leaves the room
[14:30:34] <meffie> interesting.
[14:31:14] <Jeffrey Altman> I'm catching up
[14:31:20] <stephan.wiesand> Re 11782: in any sane installation, that directory ought to exist.
[14:31:24] <meffie> (i also think the bosserver should not make symlinks in /usr/vice/etc/)
[14:32:11] <stephan.wiesand> Mike: wasn't there a change for that too? (no symlinks in /usr/vice/etc)
[14:32:43] <meffie> there was a related change, just in the redhat init script.
[14:38:13] <meffie> i'd like to take a closer look at 11782.
[14:38:43] <Jeffrey Altman> kaduk: you are correct about 11782.  On Windows the paths referenced by AFSDIR_SERVER_AFS_DIRPATH and AFSDIR_SERVER_BIN_DIRPATH are not guaranteed to exist because they, like all of the other paths, are bos or afs specific
[14:39:45] <stephan.wiesand> [sheep invasion here]
[14:41:15] <Jeffrey Altman> The two most frequently used C99 features that fail on the versions of Visual Studio used by OpenAFS are declaring variables are arbitrary locations in a function and initializing structure members by name instead of by location.
[14:41:33] kaduk joins the room
[14:41:54] <kaduk> I think this is the first time I had seen the syntax for initializing array members by location (as opposed to struct members).
[14:42:27] <kaduk> As I noted in the change in question, afsd_kernel.c does not get compiled on windows, so buildbot didn't complain.
[14:42:40] <meffie> first time i've seen it in openafs. i've seen it in gcc-only projects before.
[14:43:05] <meffie> it == initializing array members by location
[14:43:08] <Jeffrey Altman> Microsoft will be supporting full C99 by adding clang as an alternative compiler for visual studio 2015.  Reports are it will be available by November but converting Heimdal and OpenAFS to use clang and the new Universal Runtime Library (the runtime library finally becomes part of the operating system and not the compilation tool chain) is going to be a major project.
[14:43:47] <Jeffrey Altman> initializing array members by location is unlikely to work on Windows
[14:44:30] <meffie> i'm not crazy about the duplication of the names anyway, (e.g  [FOO] = "FOO")
[14:46:01] <kaduk> Jeff: do I remember correctly that you said it took some effort to get AuriStor building with VS2015 (as needed for win10 compliance)?
[14:51:42] <Jeffrey Altman> what I said is I have not finished getting AuriStor to build with VS2015
[14:52:08] <kaduk> Ah, thanks for helping my memory along.
[14:52:18] <meffie> kaduk: the -1 on 11599 makes me sad. :)
[14:53:34] <Jeffrey Altman> not to mention making all of the packaging changes that are required to use the new Universal Runtime.   The new runtime is part of the Windows 10 OS image but is not present in prior OSes and it must be installed by calling into Windows Update to fetch it on Windows 8.x but on Windows 7 there is a requirement that the installer call out to run an EXE to install it.   All of this is a royal pain in the rear.
[14:53:43] <meffie> do i just need to move the "create the local" config" later in the initialization?
[14:54:09] <kaduk> That does sound like quite the pain in the rear :(
[14:54:42] <kaduk> meffie: I think so, yes.  There were maybe two statements that needed moving; I'd have to check.
[14:54:42] <Jeffrey Altman> of course the universal runtime is not available at all pre-Win7
[14:55:17] <kaduk> I am lucky enough to no longer need to care about pre-win7; my condolences that you are not in the same boat
[14:56:21] <kaduk> meffie: looks like just assigning to tdir and to bozo_confdir, if memory serves correctly
[14:56:28] <Jeffrey Altman> I don't want to care but you might have seen http://arstechnica.com/information-technology/2015/06/navy-re-ups-with-microsoft-for-more-windows-xp-support/
[14:56:31] <meffie> ok, thanks.
[14:56:44] <kaduk> Yeah, I think I saw that.
[14:57:00] <Jeffrey Altman> and they use AFS
[15:00:25] <stephan.wiesand> I have to shop for food. Shop's 25 miles from here.
[15:00:35] <Jeffrey Altman> start walking
[15:00:51] <Jeffrey Altman> you should get there by the time they open in the morning :)
[15:01:05] <stephan.wiesand> I'll rather drive.
[15:01:08] kaduk leaves the room
[15:01:56] <stephan.wiesand> Still, I have to pass through the mountains. Which means driving right through the clouds.
[15:02:08] <Jeffrey Altman> be safe
[15:02:21] <Jeffrey Altman> watch out for moose (?)
[15:02:29] <meffie> have fun stephan.wiesand!
[15:03:00] meffie leaves the room
[15:04:19] <stephan.wiesand> Thanks. Thanks a lot for participating today! The current lack of activity is entirely my fault. I'm not feeling too guilty though...
[15:04:33] mvita leaves the room
[15:04:39] <stephan.wiesand> Bye.
[15:04:45] stephan.wiesand leaves the room
[15:17:35] <Jeffrey Altman> No need to feel guilty
[15:20:04] Daria Brashear leaves the room
[17:36:58] Daria Brashear joins the room
[19:27:48] Daria Brashear leaves the room
[20:14:46] Jeffrey Altman leaves the room
[20:15:22] Jeffrey Altman joins the room
[20:16:15] Jeffrey Altman leaves the room
[20:23:36] Jeffrey Altman joins the room
[21:28:53] Daria Brashear joins the room
[21:33:57] Daria Brashear leaves the room
[21:54:48] Daria Brashear joins the room
[21:54:59] Daria Brashear leaves the room
[23:50:33] Jeffrey Altman leaves the room