[03:57:40] --- Russ has left: Disconnected [05:27:00] --- Jeffrey Altman has left: Replaced by new connection [06:12:27] --- Jeffrey Altman has become available [08:22:59] --- matt has become available [09:03:18] --- cclausen has become available [09:52:55] --- Russ has become available [12:39:25] --- dev-zero@jabber.org has become available [12:39:30] --- dev-zero@jabber.org has left: offline [12:43:53] --- matt has left: Logged out [12:43:56] --- matt has become available [13:55:33] --- tkeiser@sinenomine.net/owl has left [14:05:09] I'd like to speak with someone who has recent experience with the ihandle package. [14:30:25] --- Russ has left: Disconnected [14:31:21] --- Russ has become available [15:14:28] Yuck. Our CVS tree is rather messier than I though. We've got the same delta names on HEAD and in branches in some places. [18:46:54] The fileno member of struct FILE is a short on the bsds, and an int on fedora 6 (not 8 bits). I wrote a patch that makes the fd cache sizes configurable and replaces the host hold bitmap with a refcount edited using atomic inc/dec, allowing to set MAX_FILESERVER_THREAD to be 16K. Someone tell me all the reasons why that doesn't work...maybe I'll get lucky and part of it eventually works. [20:15:11] --- cclausen has left [20:48:53] --- cclausen has become available [21:39:12] > We've got the same delta names on HEAD and in branches in some places. [21:39:19] that is, alas, accidentally true [22:43:10] --- Russ has left: Disconnected [23:13:19] Well, every delta on the stable_1_0_x branch that didn't have 'STABLE10' at the start of its name, does now. And that particular problem is solved ... [23:38:31] > fileno member of struct FILE is a short on the bsds I seem to recall that there was talk of changing that on fbsd ...