[00:48:33] --- abo has become available [00:50:56] --- abo has left [00:51:02] --- abo has become available [02:15:55] --- Simon Wilkinson has become available [05:30:36] --- mvita has become available [07:23:23] --- deason has become available [07:50:26] --- gendalia has left [07:52:21] --- gendalia has become available [08:02:25] --- meffie has become available [08:06:38] kaduk, was working on the patches for parallel make over the 3-day weekend. just dumped them on gerrit. [09:22:17] --- meffie has left [09:23:08] --- meffie has become available [09:42:07] Thanks, Mike -- I see I have some mail about it, too :) [09:42:08] --- Marc Dionne has left [09:45:31] --- Marc Dionne has become available [10:51:27] --- deason has left [10:58:08] --- deason has become available [11:52:35] Hmm, does separate build tree work on 1.6 these days? I'm poking at osx 10.9 and seeing "rxkad.h not found" while building akimpersonate_v5gen, which kind of looks like it's looking in the src tree and not the build tree for the compile_et droppings. [11:53:43] Possibly, I don't think objdir builds get much love. [12:09:33] I appear to be having a "how did this ever work?" moment. And I think maybe this is even code I wrote... [12:19:56] http://threepanelsoul.com/comics/2013-12-16-305.png [12:53:37] I had forgotten that there is no src/opr at all on 1.6; this means I get to figure out some other clever way to backport the sbrk()-avoidance. [12:54:52] Probably the lowest-effort is to just make a viced-specific helper. [12:56:18] Or just return 0 [12:56:56] I guess it is kind of a useless feature at this point, but I'd rather not neuter it for everyone when only OS X is complaining. [12:57:14] It's neutered for everyone anyway. [12:57:40] Pretty much every malloc() implementation that I'm aware of using mmap() to obtain memory, rather than sbrk() [13:14:06] Hmm, I see: Sun Jan 19 06:00:46 2014 There are 767 connections, process size -407375840 [13:30:04] This is proving to be quite the game of "cherry-pick Simon's commits from a year ago" :) [13:39:25] --- jaltman/FrogsLeap has left: Disconnected [13:39:47] --- jaltman/FrogsLeap has become available [14:20:51] Which ones? [14:21:21] The ones tagged 'mavericks' targetted for openafs-stable-1_6_x on gerrit :) [14:22:17] Which is to say, http://gerrit.openafs.org/#q,status:open+project:openafs+branch:openafs-stable-1_6_x+topic:mavericks,n,z [14:23:01] Ah, the "keep clang happy" series [14:23:41] That's pretty much what's going on, yup. [14:24:43] clang-analyzer is getting to be pretty good. It just needs a coverity style front end to help dealing with the false positives [15:01:52] --- Marc Dionne has left [15:11:15] --- Simon Wilkinson has left [15:11:34] --- Simon Wilkinson has become available [16:38:09] --- deason has left [16:43:04] --- meffie has left [17:04:25] --- mvita has left [18:15:41] --- ktdreyer has left: Replaced by new connection [18:15:42] --- ktdreyer has become available [18:44:03] --- Simon Wilkinson has left [18:47:56] --- Simon Wilkinson has become available [20:52:34] Sigh, the 1.6 branch doesn't have roken, either, and strlcpy is not portable enough. [20:55:04] this is why I believe that 1.9 needs to be cut from master instead of trying to turn 1.6 into master. [20:56:03] I was just about to write "if only there was some sort of 1.9 branch"... [20:56:24] But, for this particular strlcpy case, a workaround should not be too difficult. [20:56:26] it can be cut any day