[00:24:30] Ubuntu packages of 1.6.0pre2 are building in the PPA: https://launchpad.net/~openafs/+archive/master (Russ, my Git repository has an orig tarball from get-orig-source that we can share.) [01:02:13] --- Russ has left: Disconnected [05:18:32] --- cudave has left: Replaced by new connection [05:18:33] --- cudave has become available [07:27:50] --- deason has become available [08:12:30] --- Simon Wilkinson has left [08:42:54] > 14-Feb-2011 - OpenAFS 1.6.0 prerelease 2 available > OpenAFS 1.6.0 prerelease 2 is the second candidate for the OpenAFS > 1.6.0 release. > The 1.6 series is intended to provide new features including the > Demand Attach File Service and ^^^^^^^ Uh. [08:47:36] what? [08:48:25] I can't tell what emotion goes with your surprise, Andrew. [08:49:46] I'm not sure what you're trying to point out [08:49:52] so... "confusion", I suppose :) [08:50:20] "I didn't know 'DA File Service' (as opposed to Server) was a thing" [08:52:28] speling [08:52:35] eh, potayto, potahto [09:10:09] the service provided by the demand attach fileserver would be... what, then? [09:29:09] > the service provided by the demand attach fileserver would be... the file service. The service isn't demand-attach; the volume package implementation is. 1.6.0 does not provide a new service in this regard, only a new server that provides the same service [09:29:57] sort of. i didn't use parens in the text. the (demand attach) (file service) as opposed to the (non demand attach) (file service) [10:17:12] It doesn't matter. demand-attach is not a property of the file service. It doesn't modify "service" and it doesn't modify "file service". It modifies "fileserver", which word you didn't include. [10:22:34] --- rra has become available [10:28:20] --- sxw has become available [10:29:42] --- sxw has left [10:30:41] --- jaltman/FrogsLeap has left: Disconnected [10:32:34] --- mfelliott has become available [12:01:56] --- mfelliott has left [12:30:57] --- mfelliott has become available [13:12:31] --- Simon Wilkinson has become available [14:33:40] --- Simon Wilkinson has left [14:46:43] --- sxw has become available [14:57:17] --- sxw has left [15:07:32] --- jaltman/FrogsLeap has become available [15:27:39] --- deason has left [18:02:55] --- rra has left: Disconnected [18:26:28] --- Russ has become available [20:21:04] --- phalenor has become available [20:36:24] --- deason has become available [21:04:08] The debian packaging has issues when the running kernel and the kernel being built for are different, right? Can these be resolved just by passing --with-afs-sysname to configure? (I am assembling a chroot farm and `uname` comes from the running kernel, which is not what I want.) [21:07:06] I'm pretty sure I successfully built a module for not the running kernel, but I don't remember how. I don't know that I've ever tried building for not the running architecture. [21:07:32] But for the chroot farm case, I'd expect that configure switch to work, provided your chroot actually has the correct tools for the platform you're building for. [21:13:28] I build Debian kernel module packages for modules other than the running kernel all the time. [21:13:34] module-assistant makes that pretty trivial to do. [21:13:51] But you have to tell it what kernel to build for; it defaults to figuring it out from the running kernel. [21:20:15] Okay, sounds good. (I think but am not sure that FreeBSD's jail(8) will let me tweak the uname output, but I'm too sleepy to get through the man page.) [21:26:51] Are there parallel-make fixes not in 1.6.0pre2? (I.e. should I be running with -jN?) [21:29:06] there my be one. i use -jN anyway [21:31:48] --- deason has left [21:45:50] In addition to setting the sysname, for Debian you also have to point it at the correct kernel headers for whatever you're building against. [21:46:03] In fact, if you get the kernel headers right, you shouldn't have to set the sysname. [23:03:52] --- Russ has left: Disconnected [23:39:51] --- Russ has become available