[00:07:29] --- Simon Wilkinson has left [00:20:58] --- dev-zero@jabber.org has become available [00:21:15] --- dev-zero@jabber.org has left [00:21:23] --- dev-zero@jabber.org has become available [00:26:43] --- shadow@gmail.com/owlB97AB4FD has left [00:29:40] --- haba has left [00:43:05] --- abo has left [00:53:05] --- kaj has become available [00:58:17] --- reuteras has left [01:03:23] --- abo has become available [01:12:18] --- dev-zero@jabber.org has left [01:48:53] --- kula has left [02:00:34] --- haba has become available [02:45:16] --- Jeffrey Altman has left: Replaced by new connection [02:45:16] --- Jeffrey Altman has become available [02:56:58] * haba had "uncorrectable ECC error" on AFS server. But after pulling 1/2 of the RAM the OS came up and salvage did its thing. Now vos-moving out. Well well, it could have been worse. [03:29:53] --- kula has become available [04:51:16] --- kaj has left [04:54:29] --- Simon Wilkinson has become available [04:54:46] --- Simon Wilkinson has left [05:06:58] --- shadow@gmail.com/owlB07515D6 has become available [05:51:51] --- dev-zero@jabber.org has become available [06:52:56] --- abo has left [06:53:18] --- abo has become available [06:59:30] --- dev-zero@jabber.org has left [07:03:45] --- deason has become available [07:09:02] > We have one, and it would be trivial to implement, but I'm not sure > that's actually what we want. it would be really nice to have the possibility for directories that act like the unix "x but not r" directories, for which something like RXAFSLookup would be required [07:09:16] not as the primary dir lookup mechanism, but a fallback or something [07:13:56] --- reuteras has become available [07:45:02] Yeah, I could see doing that. Of course, you wouldn't be able to access such a directory at all on old clients. [07:46:29] I take it gerrit 1367/1368 will not be in 1.4.12? [07:50:06] --- jaltman has left: Replaced by new connection [07:50:07] --- jaltman has become available [08:06:48] --- abo has left [08:06:50] we could and should implement RXAFSLookup now on the server. especially since there's already code from hartmut istr [08:08:36] > I take it gerrit 1367/1368 will not be in 1.4.12? i think they're just added risk this late in the game. [08:09:28] the comment Lookup made it sound like it's just for DFS xlators, and we'd need to standardize on it or something before actually using it for normal service [08:09:39] "the comment by Lookup" [08:11:03] it is just for xlators but the clients do all implement it. i think we standardize it as "working how it already works" [08:11:15] actually, unsure if windows implements it [08:13:37] if we could use that for traversing directories without listing contents, what would the necessasry permissions be? 'r' but not 'l' on the parent; or do we need to establish some new bit? [08:16:24] i assume at the moment you can only use it for where you have l, and *new* behavior does need standardization [08:24:28] --- sxw mobile has become available [08:30:56] The OpenAFS UNIX client implements it, but only uses it on volumes which it thinks are translator volumes. hostafs depends on this, and it works fine. [08:33:44] --- reuteras has left [08:35:52] --- sxw mobile has left [08:41:13] --- sxw mobile has become available [08:42:44] --- sxw mobile has left [08:48:31] --- haba has left [08:52:09] --- sxw mobile has become available [08:54:00] --- sxw mobile has left [08:56:50] the Windows client doesn't use it and can't use it because it does not implement case insensitive lookup semantics or Unicode normalization [09:01:28] --- sxw mobile has become available [09:02:03] --- sxw mobile has left [09:06:07] I just sent an e-mail to afs3-standardization regarding the licensing issues. It appears that when IBM OpenAFS 1.0 was imported into the openafs repository, the IBM Public License 1.0 should not have been applied to all files in the tree that contained an IBM or Transarc copyright. The rx, rxkad, rxgen and util directories for example had a much more liberal license left over from the DARPA grant work. Many other files were copyrighted from the ITC CMU project days and more than likely should have not been relicensed. What I think we need to do now is go back and reconstruct what we think the licensing on all of these files should be. Document the differences and present them to IBM for review. If we can get agreement from IBM that our interpretation of what should and should not be IPL 1.0 is accepted, then we will need to go back to all of the contributors to the affected source files and obtain permission to relicense their contributions under the appropriate license. I for one would be willing to authorize a relicensing of all contributions from myself, Secure Endpoints Inc. and Your File System Inc. to rx, rxkad, rxgen, and util under the original "this is free, no rights reserved" license that IBM published it under. [09:08:22] --- sxw mobile has become available [09:20:22] --- sxw mobile has left [09:23:33] And I replied to your message. The changes we made were correct, under our understanding of the terms under which IBM was releasing AFS. Some things had been released by IBM under more liberal terms before, and some were derivatives of things that were released under more liberal terms before IBM owned them, but it was fairly clear at the time that the terms that applied to what IBM released as OpenAFS 1.0 were the IPL 1.0. [09:24:07] you, jhutz, have already commented that not all places we applied the IPL should have had it applied [09:24:37] The items you're talking about, particularly including Rx and the xg files, were IBM's sole property. Any previous licenses on them before they released them as part of OpenAFS 1.0 were pretty much irrelevant. [09:25:06] Yes, I do think there were a couple of places where we were overzealous. But the .xg files with no message at all were not one of them. If we had not assumed the IPL applied to those, we could not distribute them at all. [09:26:19] In any case, I think it is a good idea to document what we think the licenses ought to be, and ask IBM and any other contributors to those files to relicense under those terms. My guess is that we will not get IBM to agree to a less restrictive license on Rx and related code, even though it was released before under such a license. _Maybe_ in 2001 we could have done that, but not today. [09:26:42] no harm in asking [09:27:14] Thus, we should push hardest on the .xg files, because relicensing those will significantly reduce pain for our protocol maintenance and standardizaton process. [09:28:54] sometimes I wonder if our lives wouldn't be easier if we let Peter decide what to do about IPR issues instead of having people like Jeff, Russ, Simon, and I argue about them. :-) [09:31:01] probably [09:31:33] At least, until we all got sued [09:49:22] --- sxw mobile has become available [09:59:17] which Peter? [09:59:33] --- sxw mobile has left [10:03:07] --- Simon Wilkinson has become available [10:08:39] Peter? [10:09:05] --- sxw mobile has become available [10:09:20] see jhutz's comment from earlier [10:10:46] Can't see a comment beyond "if we let Peter decide ...", which is what I was referring to. [10:10:55] --- sxw mobile has left [10:13:24] --- Russ has become available [10:13:27] --- Simon Wilkinson has left [10:13:37] --- Simon Wilkinson has become available [10:16:17] --- sxw mobile has become available [10:16:25] --- sxw mobile has left [10:16:27] oh. we were talking about intellectual property issues w.r.t openafs source licensing [10:16:54] sure. but you haven't answered the question of "which Peter"? [10:17:00] Yes, but who's the Peter who should be deciding/ [10:17:45] --- kaj has become available [10:17:59] well, i assumed he meant honeyman, at the time this started [10:18:03] not now [10:25:16] --- Simon Wilkinson has left [10:25:54] --- haba has become available [10:41:03] Russ: some weeks ago, you mentioned that you had some autogoo for translating a heimdal-like API to actual heimdal or MIT krb5, as appropriate. Do you have a pointer to where I could grab it? [10:41:35] http://www.eyrie.org/~eagle/software/rra-c-util/ See, in particular, portable/krb5.h, krb5-extra.c, and krb5-renew.c and the bits in configure.ac and m4/krb5.m4. [10:41:57] It's just a start, but it handles the stuff I've needed so far. Further additions are very welcome, of course. [10:42:02] Thanks! [10:43:00] --- jaltman has left: Disconnected [10:43:27] --- jaltman has become available [10:45:03] Now that I think of it, I bet neither Jeff nor Simon was at the AFS session at LISA in Boston several years ago, when Peter expressed his opinion regarding licensing issues related to OpenAFS and Linux. [10:45:50] Was this useful and informative Peter, or the other Peter? :) [10:46:22] I certainly wasn't there. [10:46:28] Is there a slide deck? [10:47:57] slide deck? no. [10:48:23] he expressed it by standing up and being animated [10:48:34] "by being peter" [10:56:15] --- Kevin Sumner has become available [10:57:38] --- Simon Wilkinson has become available [12:17:01] --- Russ has left: Disconnected [12:19:06] --- Russ has become available [13:00:17] --- Simon Wilkinson has left [13:01:37] --- Simon Wilkinson has become available [13:02:28] --- Russ has left: Disconnected [13:14:29] --- Russ has become available [13:25:23] > "by being peter" I remember Monterey ;) [13:31:42] --- Simon Wilkinson has left [13:31:42] --- Simon Wilkinson has become available [13:47:19] --- kaj has left [13:52:06] --- Simon Wilkinson has left [13:52:08] --- Simon Wilkinson has become available [13:57:45] * Russ wonders if this person's Cosign problem is that Cosign doesn't build with Heimdal, or if he's just confused about needing MIT Kerberos. [13:57:53] * Russ doesn't know enough about Cosign to really be answering these questions. [13:58:23] I dunno. We've only tried building cosign with MIT. [13:58:37] Although, if they don't use krb5-config, I suspect the chance of Heimdal compatibility is small. [14:10:57] Yeah, probably so. [14:15:38] --- haba has left [14:35:16] --- Kevin Sumner has left [14:39:14] --- steven.jenkins has left [14:40:02] --- steven.jenkins has become available [14:44:59] --- Simon Wilkinson has left [14:52:52] --- jaltman has left: Replaced by new connection [14:52:53] --- jaltman has become available [14:58:47] --- shadow@gmail.com/owlB07515D6 has left [15:03:11] --- jaltman has left: Replaced by new connection [15:03:12] --- jaltman has become available [15:28:15] --- deason has left [16:33:22] --- asedeno has left [16:33:34] --- asedeno has become available [16:34:28] --- steven.jenkins has left [16:35:17] --- steven.jenkins has become available [16:42:45] --- RedBear has left [16:42:59] --- dwbotsch has become available [18:06:40] --- phalenor has left [18:16:43] --- phalenor has become available [18:50:32] --- jaltman has left: Disconnected [18:50:40] --- jaltman has become available [19:07:41] --- jaltman has left: Replaced by new connection [19:07:44] --- jaltman has become available [19:11:52] --- jaltman has left: Disconnected [19:12:04] --- jaltman has become available [20:04:58] --- deason has become available [21:26:25] --- deason has left [21:46:10] --- shadow@gmail.com/owlBD8333E1 has become available [21:48:25] --- jaltman has left: Disconnected [23:43:38] --- dev-zero@jabber.org has become available [23:43:54] --- dev-zero@jabber.org has left [23:44:05] --- dev-zero@jabber.org has become available [23:54:56] --- abo has become available