[00:03:05] --- dev-zero@jabber.org has left [00:26:22] --- abo has left [00:30:11] --- abo has become available [02:11:47] --- dev-zero@jabber.org has become available [03:41:36] --- SecureEndpoints has left: Replaced by new connection [05:49:57] --- dev-zero@jabber.org has left [05:51:20] --- Simon Wilkinson has become available [05:53:46] --- dev-zero@jabber.org has become available [06:00:20] --- Simon Wilkinson has left [06:25:43] --- SecureEndpoints has become available [06:26:08] --- Simon Wilkinson has become available [06:30:32] --- Simon Wilkinson has left [06:31:51] french spam said something about vos. i opened it. realized duh "vos has nothing to do with afs" [07:50:58] --- reuteras has left [07:51:28] --- matt has become available [08:45:34] it is now possible to build the rx library without RXDEBUG defined [08:47:10] so far there is no performance benefit from doing so. however, on memory contrained devices being able to reduce the memory footprint is beneficial [08:54:26] --- Russ has become available [09:00:44] --- Simon Wilkinson has become available [09:31:57] --- mcohan has become available [11:22:19] --- Moose has become available [11:42:09] --- dev-zero@jabber.org has left [12:15:14] --- Moose has left: Replaced by new connection [12:15:15] --- Moose has become available [13:12:28] --- Moose has left: Replaced by new connection [14:16:28] Thoughts appreciated on /afs/inf.ed.ac.uk/user/s/sxw/Public/openafs-disconnected-extend.patch [14:17:49] This fixes the problem edgester found with extending files in disconnected mode, at least on Linux. I have some concerns about the VM architectures of other systems, and whether we have to inform them of file extensions. My suspicion is that it needs an additional ubc_setsize() on DARWIN, but I don't know about Solaris, and I don't have access to anything else to test with ... [14:26:27] --- dmontuori has become available [14:27:22] --- dmontuori has left [15:28:45] --- dev-zero@jabber.org has become available [17:28:42] simon - I could possibly dig up a Solaris system to test on. [17:29:00] not for certain, though. [17:29:23] I could certainly do a build test -- I'm just not sure if I can get access to anything except a Solaris zone (which won't let me run kernel modules) [17:29:58] You might find it useful to use a vm. [17:31:13] I do have a solaris vmware image somewhere...just unsure of its status. [17:31:22] and I know the zone we rent is in decent condition. [17:44:02] --- matt has left [17:50:32] --- matt has become available [17:56:07] SecureEndpoints: reading back, the Andreas did get a bit accomplished, with per file acls, and that could be carried forward. But if you'll recall, in that too, I'd assumed we could follow a short-path approach which would yield per-file ACLs, but not address ACL limits, and you found that less desirable. I think I'd find myself arguing for a short-path solution there too, if that started back up. [17:57:08] DAFS bugs have lots of eyes focused. [18:17:53] question..for function declarations, is there a preference for fun(a, b) int a, b; vs fun(int a, int b) ? [18:19:19] There is a detailed byte-range locking design draft published. It depends on extended callback information, and that is implemented and being prepared for (hopefully) future integration into openafs. [18:19:24] sjenkins: ansi [18:20:14] matt - tx; thought so, but I also thought I've seen k&r style in the code (but couldn't find any) [18:20:40] You have. I think Simon and Derrick are busy about improving prototypes, though. [18:24:59] --- matt has left [19:27:22] rt.central.org is taking forever... [19:27:40] (where 'forever' means 'more than 5 seconds') -- I should sleep instead and do this tomorrow. [19:53:07] ansi>>k&r [20:18:19] Matt: if I disagreed with the per file ACL approach I would not have permitted the gsoc project to go forward [20:19:03] more importantly I wouldn't have voted in favor of funding it [23:02:49] --- Russ has left: Disconnected [23:16:05] --- reuteras has become available [23:53:26] --- manfred furuholmen has become available