[00:02:55] --- Russ has left: Disconnected [02:44:55] --- manfred has left [02:49:54] --- manfred has become available [04:35:28] --- jaltman has become available [04:55:34] --- meffie has become available [05:06:40] --- meffie has left [05:18:25] --- meffie has become available [05:26:24] --- jaltman has left: Disconnected [05:30:26] --- jaltman has become available [05:52:10] --- reuteras has left [05:58:37] --- jaltman has left: Disconnected [05:58:53] --- Jeffrey Altman has left: Replaced by new connection [06:02:51] --- deason has become available [07:34:59] --- Simon Wilkinson has left [07:40:02] --- jaltman has become available [07:52:49] --- abo has left [07:53:49] --- abo has become available [08:09:33] --- jaltman has left: Disconnected [08:34:48] --- kula has left [09:02:48] --- jaltman has become available [09:27:50] --- Simon Wilkinson has become available [09:35:09] --- kula has become available [09:43:34] --- Mickey Lane has become available [10:00:43] --- manfred has left [10:35:44] --- Russ has become available [10:45:06] Oh, sweet, Dreamwidth now implements OpenID 2.0. [10:45:14] So I can use my DW OpenID for Gerrit. [10:45:23] Cool... Does it work? [10:45:58] (You probably want to add a new OpenID by doing the 'Add an Identity' thing, rather than just by logging on with the new ID. [10:46:07] Trying it now. [10:46:10] Otherwise I have to unify the accounts manually in the database. [10:46:57] Grumble. Apparently no, not yet. [10:47:44] I wonder if they've actually pused the OpenID 2.0 implementation yet. [10:47:50] I'll wait until after the next code push and try again. [10:48:16] Our local OpenID IdP didn't work because it used self signed certificates. [10:48:42] I wonder if they only added accepting OpenID 2.0 identities and didn't update the provider. [10:49:02] An earlier provider should work, too. Depending on what it does and doesn't implement. [10:49:08] What error do you get when you try? [10:49:30] The same thing as before with LiveJournal and Dreamwidth: I get the "this is an OpenID endpoint, not a user page" page on Dreamwidth. [10:49:36] When I try to log in. [10:49:45] Okay, so that is their implementation... [10:49:51] Yup. [10:50:02] I don't think that's a 1.0 vs 2.0 issue though. I think it's just an implementation bug. [10:55:21] * Russ files a support request with Dreamwidth. [10:55:27] Since the OpenID 2.0 stuff is live. [11:28:19] --- dev-zero@jabber.org has left [11:36:33] Hm, despite the plea on openafs-devel, it doesn't look like we really have that much unreviewed stuff in Gerrit. We have a lot of stuff that either failed verification or has -1 reviews and hasn't been updated since. [11:37:11] one review doesn't count as "reviewed" [11:37:14] We have a large amount of stuff that is reviewed by "the usual suspects" [11:37:21] Oh, okay. I see. [11:38:14] I would prefer to see at least five reviewers for any item but I do not expect it [11:38:53] I think 5 reviewers is not really living in this world. [11:39:18] I'd like to see 2 or 3. Where at least one of them hasn't been involved in crafting the patch. [11:39:27] I don't think there's any open source project in existence that gets five reviewers for each change. [11:39:39] doesn't mean we shouldn't try [11:39:48] but no, i'm not holding my breath for it either [11:41:11] My worry is that if Marc and I (for example) bounce ideas backwards and forwards in email, and Marc then posts a patch, I'm not necessarily a proper "second set of eyes" for that patch. [11:44:17] We really need more reviewers who feel comfortable reviewing kernel code. [11:44:27] If I had an extra four hours in each day, I could become one a lot faster. [11:44:52] I'd really like more regular reviewers who are familiar with non-Linux, non-Darwin code. [11:45:05] We don't really have any review (or test) coverage for Solaris or xBSD. [11:45:20] Good point. [11:45:40] The people most able to review Solaris code are the people who are also submitting most of the Solaris patch sets, but review of shared kernel code would be great. [11:45:54] i was fine for solaris until they decided they didn't want to be solaris anymore [11:46:53] At some point, we're going to need to make some pretty big, cross platform changes to the Unix CM. For example, it would be really good if we could move to a page-based view of filesystem memory on those platforms that have that model internally. [11:46:58] Is there any reason why the Dreamwidth folks can't create a Gerrit account to test the OpenID linkage with Dreamwidth? I don't think there's any harm, but wanted to check. [11:47:06] No harm at all. [11:48:18] for the systems we can virtualize, i'd like VMs we could use to test. even if it's an archived copy we can pull out, taint and then toss. that doesn't help us on irix, aix or hpux tho [11:48:40] the irix machine i have is a fossil and not mine to donate, but i can ping redacted again [11:50:02] I'd just really like there to be someone I can go "I'm thinking of doing this - how bad will that be on platform X". I know who those people are for Linux, Darwin, xBSD. But no idea who to talk to about HPUX, IRIX, AIX or Solaris. [11:50:28] We should probably also be thinking about what our exit criteria is for platforms -- how we decide when they go the way of DUX. [11:50:42] This primarily affects HPUX and IRIX, since the others I know are actively used. [11:51:48] I'm all in favor of being very conservative about dropping platforms; we should just have some criteria in mind. [11:51:58] Which may be "it doesn't build, no one provides any hardware, and no one is fixing it." [11:53:00] Do we have any code specific to Linux 2.2 and earlier still in the tree, OOC? [11:53:02] irix i know of one for sure. hpux well, i assume nothing's changed in 14 months [11:53:12] The other option is that we look at ways of removing those platforms from any major changes we make - so they can continue using the old code. [11:53:17] 2.2 worked not that long ago [11:53:44] We still have ifdefs for 2.0, 2.2 and 2.4, in the LINUX24 directory, and in the main tree. [11:53:47] Simon: Yeah, like we did with 2.4. I like that move. [11:54:14] Anything that gets the mainline, actively developed code to have fewer #ifdef mazes makes it easier for everyone to help maintain the code. [11:54:43] To that end, I have a patch that completely reworks NewVCache so all of the platform specific code goes out into the platform directories. [11:55:32] Yay! [11:55:57] just need to test it now ... [12:13:41] --- jaltman has left: Disconnected [12:20:38] --- manfred has become available [13:51:46] --- manfred has left [14:23:43] --- jaltman has become available [14:49:32] --- Mickey Lane has left [15:48:52] --- deason has left [15:49:45] --- jaltman has left: Disconnected [16:12:44] --- mdionne has become available [16:16:12] do we know if anyone depends on the fact that you can currently "join" a PAG by adding the appropriate groups to your group list ? i.e. is it OK if this ability goes away? [16:25:40] I certainly hope no one relies on that. [16:26:28] Yes, I believe we already decided it can/will go away when we move to pure keyring PAG's on Linux [16:30:40] That's what I thought but wanted to make sure. I'll have an updated version of change 730 that makes PagInCred look only at keyrings. [17:09:13] would we still expect a group to show up as an indicator that we're in a PAG, or could that go as well? [17:09:29] I believe we decided that would go eventually, too [17:35:49] --- Russ has left: Disconnected [17:53:47] --- Russ has become available [18:03:55] --- mdionne has left [18:09:57] Having the group go away will bother more people, but I think in the long run it should go. [18:10:20] I would kind of like to add a command-line tool that tells you whether or not you're in a PAG, and the libkafs function that I mentioned on openafs-devel, before we drop the GIDs. [18:21:09] --- deason has become available [19:53:16] --- jaltman has become available [20:35:31] --- meffie has left [21:32:12] --- deason has left [21:56:07] --- Russ has left: Replaced by new connection [21:56:07] --- Russ has become available [22:02:18] --- reuteras has become available [22:39:37] --- jaltman has left: Disconnected