[00:08:52] --- Simon Wilkinson has become available [00:34:17] --- jaltman has left: Replaced by new connection [00:34:18] --- jaltman has become available [01:07:12] --- dev-zero@jabber.org has left [01:25:43] --- manfred has left [02:21:30] --- dev-zero@jabber.org has become available [04:12:38] --- meffie has left [04:22:48] --- manfred has become available [04:36:26] --- haba has left [05:04:35] --- sxw has become available [05:09:33] --- haba has become available [05:09:37] --- sxw has left [05:24:09] --- sxw has become available [05:30:11] --- sxw has left [06:36:17] --- meffie has become available [06:54:56] --- reuteras has left [06:57:07] --- Jeffrey Altman has left: Replaced by new connection [06:58:59] --- jaltman has left: Replaced by new connection [06:59:00] --- jaltman has become available [07:03:39] --- jaltman has left: Replaced by new connection [07:03:39] --- jaltman has become available [07:16:06] --- deason has become available [07:19:09] --- haba has left [07:29:02] --- haba has become available [08:04:05] --- stevenjenkins has left [08:07:17] --- stevenjenkins has become available [08:52:47] --- stevenjenkins has left [08:53:22] --- stevenjenkins has become available [09:25:21] --- meffie has left [09:37:03] --- haba has left [10:23:40] --- sxw has become available [10:24:30] --- summatusmentis has left [10:28:26] --- summatusmentis has become available [10:53:56] --- sxw has left [11:08:07] I'd invite anyone that is familiar with softsig to take a look at gerrit 693 [11:11:39] i am trying to ferret out any notes from zephyr which didn't end up in rt 1453. so far none [11:16:52] sigwait manpages say you shouldn't sigwait() on a signal not being blocked... some others say you wouldn't block SEGV... it leads me to believe we shouldn't be blocking these at all, but I'm afraid of breaking things [11:17:34] What do you mean, "wouldn't block SEGV" [11:18:25] err, shouldn't [11:18:43] I've been looking at a lot of these... lemme try to find what I'm thinking of [11:18:51] Well, you probably don't want to _catch_ SEGV and then not die usefully. [11:19:19] yes, and everything !HPUX seems to ignore that we're trying to catch segv and just cores anyway [11:19:57] I believe the old issue to which you refer in your notes has to do with old-style linuxthreads, in which each thread is actually a separate process. That model had a lot of signal-delivery issues. [11:20:15] That's bizarre, because you generally _can_ catch SEGV [11:21:11] at least; they core anyway the way we're doing it [11:21:23] the way it's done seems wrong anyway to be since we don't block the signals before we sigwait() on them [11:24:09] one further note in RT; comments from chaskiel on zephyr now in gerrit [11:26:44] The pipe thing is clever because it allows something to use select(2) to block on a combination of signals (represented as fd's) and actual fd's. [11:33:07] --- dev-zero@jabber.org has left [12:06:29] --- manfred has left [12:26:36] --- Jeffrey Altman has become available [12:38:04] --- dev-zero@jabber.org has become available [13:02:17] --- Jeffrey Altman has left [13:05:48] does gerrit 690 mean we don't really need AFS_PTR_FMT at all? [13:06:31] --- haba has become available [13:06:33] when AFS_PTR_FMT was first discussed %p could not be used on all Unix hosts [13:06:45] I forget why, but I don't think it does. [13:07:33] I don't know why it was defined as %Ip on Windows since on Windows %p was in use before it was permitted in any of the non-Windows sources [13:12:50] --- Jeffrey Altman has become available [13:22:17] --- Jeffrey Altman has left [13:37:14] --- manfred has become available [14:09:07] --- meffie has become available [14:26:17] --- haba has left [15:26:09] --- deason has left [15:46:55] --- meffie has left [15:53:07] --- deason has become available [16:25:34] IIRC, it was %Ip on Windows because you said that's what the correct Windows usage was ... [16:33:19] deason: As far as I can see, if we're happy that Windows takes %p too, then we can get rid of AFS_PTR_FMT and use %p everywhere. [16:35:37] %p everywhere: please kill me [16:35:43] ? [16:36:00] In effect, its what we do now. [16:36:01] hey, we just finished putting it everywhere. LET'S UNDO IT! [16:36:04] Yeh. [16:36:16] i need another beer [16:36:18] Looks like we discussed this back in August. [16:36:34] http://jabber.openafs.org/openafs@conference.openafs.org/2009-07-14.txt [16:36:46] July, even. [16:37:00] august. 07. nice. [16:38:03] --- manfred has left [16:43:08] b41d124409589e2890dea60963db4dc158f4d19c introduced %Ip for Windows. As long as we're happy that ... [16:43:15] a) We're not going to need it back again and [16:43:47] b) That we're not supporting any platforms whose printf's _don't_ support %p [16:43:58] I'm happy to do the work in removing AFS_PTR_FMT [16:44:13] it's easy enough to unrevert if we need it back [16:44:45] I guess... [16:44:58] so, sure, if you wanna do it, go nuts [16:45:10] It's a distraction from a) the gerrit upgrade b) my ongoing fight with writepage deadlocks c) the work I should be doing ... [16:46:27] well, i'm happy to leave it alone [16:46:29] honest [17:03:59] --- gendalia@gmail.com/owlCF44859C has left [18:07:06] --- andersk@mit.edu/dr-wily has left [18:23:29] --- andersk@mit.edu/dr-wily has become available [20:35:12] --- jaltman has left: Replaced by new connection [20:35:13] --- jaltman has become available [20:39:57] --- jaltman has left: Disconnected [20:40:37] --- jaltman has become available [20:41:58] --- jaltman has left: Disconnected [20:45:27] --- jaltman has become available [20:48:40] --- jaltman has left: Disconnected [20:59:15] --- jaltman has become available [22:24:24] --- reuteras has become available [22:38:41] --- deason has left