[00:05:29] --- reuteras has left [00:32:10] --- Russ has left: Disconnected [00:55:01] --- jaltman has left: Replaced by new connection [00:55:13] --- jaltman has become available [01:29:43] --- jaltman has left: Disconnected [01:30:30] --- jaltman has become available [01:49:02] --- abo has left [02:36:53] --- haba has become available [02:52:32] vos still works for me even if I clean out all notices of pdc.kth.se from my CellServDB. It requiteres then the -cell "cellname" methinks [02:53:05] ThisCell should be used to find the default cell. I wonder if that only works for cells that are in CellServDB. [02:55:09] I have not tested many ThisCell cells I did not have in CellServDB but that should be tested as well ;-) [02:56:27] Works with pdc.kth.se in ThisCell and nothing in CellServDB. [02:56:58] Disclaimer: At least how my laptop is set up [03:35:06] Where should i look for the source RPM to the RPMs I find in www.openafs.org/pages/release/1.4.12/index.opensuse-11.2.html or .../dl/openafs/1.4.12/opensuse-11.2/* [03:35:36] I don't know who builds the OpenSUSE RPMs. Perhaps speak to Derrick. [03:37:05] (they're definitely built from a different src.rpm than the Fedora/RedHat ones) [03:37:36] The result looks that way, yes. [03:40:35] I don't find any matching SRPMs either for the rest (and my find . -noleaf -name \*.src.rpm is really slow). [03:41:28] There ist something called .../openafs/1.4.12/openafs-1.4.12-1.1.1.src.rpm but that seems not to be what I am looking for, [03:41:43] That's the RHEL/Fedora SRPM. [03:41:54] RHEL is 1.1.2 [03:42:03] only FC11 is 1.1.1 [03:42:55] Or maybe. FC11 is 1.1 [03:43:41] But there are MANY numbers to match. A SRPM along the RPMs would be helpful. [03:48:03] On kerberos news, the last thing I did on friday was a opensuse 11.2 RPM for openssh with your patches. I had to remove some old GSSAPI version support and support for forwarding LOCALE or something like that. I was just tooo lazy to take care of the conflicts. See ftp://ftp.pdc.kth.se/pub/krb/contrib/opensuse/11.2 [04:31:09] --- jaltman has left: Disconnected [04:35:57] --- Jeffrey Altman has become available [04:38:24] --- kula has become available [05:12:00] --- jaltman has become available [05:31:18] oensuse rpms came from christof hanke [05:31:36] and istr the srpm is edited just enough to build. [05:31:58] The build output is so different, I find that hard to believe. [06:24:18] * haba sent email to openafs-info about the SRPMs. So will see... [06:49:24] --- haba has left [06:49:47] --- haba has become available [07:08:26] --- deason has become available [07:16:08] --- meffie has become available [07:16:52] > obvious that it "works" from userspace too oops. duh. [07:17:10] shouldn't it be possible to use without a cellservdb entry, or a cellservdb at all, though? [07:18:11] * haba has not tested if it works completely without CellServDB [07:20:49] according to that mohammed guy it doesn't [07:22:43] i think it wants cellservdb regardless. no entry should work. [08:13:25] --- jaltman has left: Disconnected [08:13:46] --- jaltman has become available [08:17:45] --- jaltman has left: Lost connection [08:18:20] --- jaltman has become available [08:59:24] --- kaj has left [09:32:35] --- deason has left [09:40:09] --- deason has become available [09:52:57] --- haba has left [09:57:13] --- kaj has become available [10:17:07] --- jaltman has left: Disconnected [11:01:33] --- jaltman has become available [11:29:24] --- Russ has become available [11:31:52] --- meffie has left [11:59:36] --- jaltman has left: Replaced by new connection [11:59:37] --- jaltman has become available [12:11:41] --- haba has become available [12:17:12] --- meffie has become available [12:32:17] Hmmm. No AC_USE_SYSTEM_EXTENSIONS on RHEL5 [12:38:43] what autoconf version is that? [12:39:04] 2.59 [12:39:19] It shouldn't actually matter, because it's only developers who'll be running regen.sh [12:42:16] I assume it gives a non-dev an unhelpful error message, though; we should use AC_PREREQ [12:42:28] > it's only developers who'll be running regen.sh Not necessarily, but you can mostly make that assumption. However, I don't think it's reasonable to require anyone who wants to be a "developer" to have the latest bleedy autoconf so we can use esoteric features. [12:43:14] looks like AC_U_S_E is 2.60; that doesn't seem very bleeding new [12:43:28] Yeah, I agree. AC_U_S_E should be OK. [12:44:34] Though I'm not sure it's such a good idea. Its use seems inherently nonportable, since you can't know what extensions it's going to enable, and thus how it's going to end up polluting your namespace [12:45:11] When Russ suggested using it, he implied that there were certain guarantees about what it would enable, and more importantly, that it wouldn't disable anything. [12:46:44] there was something specific I wanted it for, but now I can't find it.... [12:47:12] preadv or something like that. You needed the X_OPEN stuff, IIRC. [12:47:16] istr I wanted to #define something to get more consistent/posix prototypes or something [12:47:21] oh, right, yeah, thatwas it [12:47:32] I've attempted to write code without namespace pollution. It's possible, but way more trouble than it's worth. [12:47:43] I've been using AC_USE_SYSTEM_EXTENSIONS in all my packages for quite some time. [12:48:06] Attempting to get all the various optional bits enabled on both AIX and Linux without it can be quite annoying and lead to weird bugs. [12:48:41] Is there a log of #openafs? [12:48:47] The POSIX_SOURCE and XOPEN_SOURCE defines are nasty, in that on some systems (such as Linux), they'll also *turn off* a bunch of other stuff in addition to making other things available. [12:48:53] http://jabber.openafs.org/ [12:49:01] IRC? Yes. See the welcome message when you join the room. [12:49:09] Oh, sorry, IRC. [12:51:44] Yeah; glibc has some default behaviors if you don't use any *_SOURCE, but if you do, you need to specify everything you want [12:52:21] And I usually end up needing GNU_SOURCE anyway, eventually, at which point you get all the pollution there is to be had. [12:53:03] Since I want stuff like the system asprintf if it's available, since it's usually way more efficient than cobbling together my own. [12:55:54] --- jaltman has left: Disconnected [12:56:19] --- jaltman has become available [13:47:48] Can vos setaddrs be used with any version of the VLDB server? [13:48:04] When were UUIDs introduced on the VLDB server side? Or have they been there basically forever? [13:51:36] --- meffie has left [13:52:04] around 3.3, IIRC [13:52:12] I suspect they've been there for OpenAFS forever. Introducing UUIDs would have meant changing the VLDB format, and I don't think we've ever done that. [13:52:32] That would imply that vos setaddrs can probably be used against any version of the VLDB, yes? [13:52:50] Since the only thing that seemed "new" was referencing the server by UUID. [13:53:39] Hm, why do I have a server UUID in my vldb with no corresponding addresses? [13:53:45] I think the fileservers have used the UUID to register for ~forever, in OpenAFS terms. [14:12:34] --- haba has left [14:56:03] 1.5.74-1 uploaded to Debian experimental. [14:57:05] Hmmm. I suppose I should try and do RPMs at some point. Added to the list. [14:59:11] Now that there are no APSL files (thank you, Derrick!) and Debian isn't carrying any local patches, it's nice and quick to update the packages. Testing and building takes a lot longer than the update process. It's lovely. [15:20:15] --- mdionne has become available [15:25:15] --- jaltman has left: Replaced by new connection [15:25:19] --- jaltman has become available [15:43:03] --- deason has left [16:10:13] --- jaltman has left: Disconnected [16:17:33] --- dwbotsch has left [16:17:36] --- dwbotsch has become available [16:25:01] --- dwbotsch has left [16:25:46] --- dwbotsch has become available [17:09:32] --- mdionne has left [17:10:27] --- mdionne has become available [17:11:51] --- mdionne has left [17:18:00] --- shadow@gmail.com/owl9B0B49DD has left [17:22:09] --- shadow@gmail.com/owl33D1E86D has become available [17:42:01] --- jaltman has become available [18:19:28] --- deason has become available [18:33:15] --- kula has left [18:39:49] --- kula has become available [19:06:47] --- mdionne has become available [19:29:59] --- mdionne has left [20:24:13] 1.5.74-1 uploaded to Ubuntu PPA . [20:24:14] --- pod has left: Lost connection [20:25:19] Russ, is afsd.fuse in the package yet? I don’t see it. [20:26:40] doubt it [20:28:00] No, and the dh_install rule didn't complain, which would imply that it's not installed? [20:28:13] did it build? [20:28:19] Oh. Duh. [20:28:20] Right. [20:28:24] I should add that dependency. [20:28:24] ok then [20:29:08] Hm, should that be included in the regular openafs-client package? [20:29:16] Or should we do something more complicated? [20:29:22] I suppose just putting it into openafs-client is fine for experimental. [20:29:37] for now, yes [20:34:21] --- steven.jenkins has left [20:35:02] --- steven.jenkins has become available [21:13:13] authclient.c: In function 'ka_ChangePassword': authclient.c:788: error: incompatible type for argument 6 of 'ubik_KAM_SetPassword' kauth.h:438: note: expected 'EncryptionKey' but argument is of type 'int' authclient.c:788: error: too many arguments to function 'ubik_KAM_SetPassword' [21:14:57] And afsd.fuse gets built twice but is never installed. [21:15:01] by make install [21:15:38] Ah, yes, there are indeed no install rules. [21:20:59] * Russ doesn't understand the ka_ChangePassword code. Why were we using a different calling sequence only for s390? [21:22:08] Is rxgen actually generating different prototypes? [21:27:07] --- Born Fool has become available [21:54:21] --- phalenor has left [22:04:18] --- phalenor has become available [22:08:11] --- deason has left [22:09:22] --- reuteras has become available [22:24:16] --- kaj has left [22:35:20] --- Born Fool has left [23:16:35] --- steven.jenkins has left [23:17:01] --- steven.jenkins has become available [23:27:05] --- kaj has become available