[03:53:09] --- pod has left [03:58:11] --- pod has become available [05:12:22] --- meffie has become available [06:05:46] --- deason has become available [09:02:12] anyone know if it's possible to do something like 96e95c4000ce649591275eb4aefceb5f28f92c6b on solaris? I don't see a linker option for it, but I wouldn't know.... [09:02:48] the rokenafs .so will build, but anything trying to link with it, I get undefined symbol errors for rk_strcasecmp et al [09:04:27] (or alternatively, how do others/heimdal deal with it? just not specifying a mapfile seems to work but...) [09:06:09] -z nodefs i think [09:09:25] ld -h suggests i am correct [09:09:55] that's already the default, and it allows the .so to be created, but the undefined refs are still there [09:10:34] the opposite, -z defs, specifies the undef symbols to be an error when creating it [09:10:43] (at least, AIUI) [09:28:28] actually, -B reduce might work, depending how it deals, i suppose. [09:29:08] --- jhutz@jis.mit.edu/owl has become available [09:29:10] --- jhutz@jis.mit.edu/owl has left: Disconnected [09:29:17] --- jhutz@jis.mit.edu/owl has become available [09:29:20] I thought I tried that... or maybe it was -B eliminate; I can try to fiddle some more [09:29:52] -B eliminate not working suggests -B reduce won't work either [09:29:57] just figured I'd ask once I realized I was just trying options without realizing what they do [09:30:17] well, the explanation in the ld documentation kinda sucks [09:50:05] --- rra has become available [10:26:41] [test message from MIT server] [10:28:14] ack [10:37:20] --- ksumner has become available [12:17:44] --- abo has left: Lost connection [12:17:44] --- mho has left: Lost connection [12:19:01] --- abo has become available [14:20:24] --- shadow@gmail.com/owlDFDA9887 has become available [14:20:25] --- shadow@gmail.com/owl2AA04E7C has left [15:01:20] --- deason has left [15:40:40] --- mfelliott has become available [16:27:47] --- meffie has left [17:33:12] --- deason has become available [17:52:29] --- jakllsch has become available [17:57:05] --- rra has left: Disconnected [18:37:08] --- Russ has become available [20:21:29] okay, on that solaris linking thing, I still don't see a way to specify an ld option to ignore undefined mapfile symbols, but it looks like I can add some stuff to the mapfile to make it do that [20:25:49] add what? [20:30:25] a trailing " = EXTERN" to every symbol [20:30:58] that's not it's intention, obviously, which is probably why it was hard to find, but it appears to work [20:31:08] "This keyword is only meaningful when generating a symbol reference. Should a definition for this symbol occur within the objects combined at link-edit, then the keyword is silently ignored." [20:31:19] "its intention", that is [20:31:49] evil. well, if it works, good enough... [20:32:10] I'll get a patch tomorrow with some other fixes... for now, it's rather late [20:32:30] yup [20:32:55] --- deason has left [21:03:42] Have we already covered the lack of the doc/ directory in the release tarball? I want to say it was mentioned somewhere but I'm not sure. [21:05:46] there is a separate tarball for docs [21:19:50] Ah. [21:19:59] That I can deal with :) [21:32:05] Anyway, this guy on IRC (#bsdports / efnet) who is trying out my packaging gets a build error that causes me to notice that my FBSD syscall-declaration fixup didn't make it into 1.5.78. I don't suppose Derrick has any interest in cutting a 1.5.78-1 tarball? [21:32:44] Hm, but I suppose I should fix this crypt() thing first. [21:42:40] Did gerrit/2932 only make it into openafs-stable-1_6_x and not master? I am not entirely sure what I should be looking for. [21:43:55] (And don't have a tree on this machine and should be asleep. Goodnight!) [22:53:36] --- geekosaur has left [22:54:01] --- geekosaur has become available [23:30:15] Hmm, it helps if I look at a clean tree.