[00:00:21] --- reuteras has become available [00:14:09] --- Russ has become available [01:55:24] --- lars.malinowsky has become available [02:26:44] --- Simon Wilkinson has become available [02:32:35] --- lars.malinowsky has left [02:32:40] --- lama has become available [03:05:10] --- Simon Wilkinson has left [03:54:06] --- Russ has left: Disconnected [04:09:51] --- steven.jenkins has become available [05:21:25] --- meffie has become available [07:40:31] --- deason has become available [08:29:26] oh, readdir how fun [09:23:20] --- reuteras has left [09:24:11] --- lama has left [10:28:36] is anyone actually able to read the readdir and readdir_move functions? [10:46:58] anything in particular you're trying to look at? [10:53:28] well, find/ls/whatever get stuck in a dead readdir() loop on larger directories [10:53:41] i've aready mucked about with it to get it to work at all [10:53:54] you're on... netbsd, iirc? [10:53:56] yeah [10:54:56] as in, infinite loop, or hanging on something? [11:05:10] infinite ... let me get a ktrace [11:06:39] getdents seems to fall over at about 4KiB of entries [11:07:05] seems to terminate the list incorrectly so userland loops [11:11:47] heh, it helps when i actually try to see what's going on, somewhat. [11:14:53] is setting ap->a_eofflag not sufficient? [11:15:14] maybe not [11:15:40] compare the eofflag and offsets returned compared to a working filesystem, I suppose [11:17:44] (and apparently we have an 'afs_debug' to make seeing these easier :) [11:19:50] hmm, this is suspicous, DIRPAD = 4 [11:20:15] should be _DIRENT_ALIGN() [11:20:41] but then i have another mess to muck through, because depending on if userland is legacy, the dirent changes alignment [11:21:06] (i.e. have to use the macro which takes an argument everywhere) [11:21:42] i really want to ignore the existing code :P [11:21:52] even though it almost works [11:27:59] it might help to indent the ifdef maze in there; if you want to submit a change that just does that, feel free :) [12:27:34] --- rra has become available [12:39:02] --- CUDaveWork has left [13:01:41] --- jakllsch has left [13:06:54] --- Simon Wilkinson has become available [13:09:03] --- Simon Wilkinson has left [13:09:17] --- Simon Wilkinson has become available [13:11:52] i think the phasers should be set for kill. no mercy. [13:12:38] That's 1.6.0 itself, surely? [13:15:21] and dont call me shirley! [13:15:27] :) [13:35:21] --- jakllsch has become available [13:37:21] i can't even figure out how this works as "well" as it does. sigh. [13:37:42] i'm about ready to reimplement readdir based on the linux implementation [13:40:02] Anything particular up with it, or is it just a tangled mess? [14:26:29] --- Simon Wilkinson has left [14:55:37] --- Simon Wilkinson has become available [15:10:38] --- meffie has left [15:36:42] --- jaltman/FrogsLeap has left: Disconnected [16:23:45] --- Simon Wilkinson has left [16:24:57] --- deason has left [16:25:43] --- Simon Wilkinson has become available [16:51:29] tangled mess mostly, if i can't understand it, i can't make it work :-) [19:05:30] --- deason has become available [19:06:24] --- rra has left: Disconnected [19:25:56] --- Russ has become available [19:50:32] --- jaltman/FrogsLeap has become available [21:03:04] --- jaltman/FrogsLeap has left: Disconnected [21:07:04] --- jaltman/FrogsLeap has become available [21:37:31] --- Russ has left: Disconnected [21:43:44] --- deason has left [22:13:25] --- Russ has become available [23:11:42] --- Russ has left: Disconnected [23:55:18] --- reuteras has become available [23:59:06] --- Simon Wilkinson has left