[00:03:30] --- Jeffrey Altman has left: Replaced by new connection [00:04:21] --- brantgurga has left [00:13:30] --- abo has become available [00:52:06] --- dev-zero@jabber.org has become available [02:21:06] --- brantgurga has become available [03:10:49] --- brantgurga has left [03:33:08] --- brantgurga has become available [03:38:09] --- brantgurga has left [03:50:52] --- brantgurga has become available [04:17:47] --- brantgurga has left [04:20:20] --- cclausen has left [04:31:46] --- dev-zero@jabber.org has left [04:45:00] --- Jeffrey Altman has become available [05:02:45] --- brantgurga has become available [05:10:49] --- brantgurga has left [05:32:02] --- Jeffrey Altman has left: Replaced by new connection [06:17:23] --- cclausen has become available [06:31:38] --- Jeffrey Altman has become available [07:00:51] --- dev-zero@jabber.org has become available [07:02:18] --- deason has become available [07:58:58] --- reuteras has left [09:09:24] --- brantgurga has become available [09:32:01] --- brantgurga has left [09:32:05] --- brantgurga has become available [09:40:30] --- mmeffie has become available [10:10:01] --- Russ has become available [10:12:24] --- dev-zero@jabber.org has left [10:23:51] --- brantgurga has left [10:47:20] --- mmeffie has left [11:21:29] --- dev-zero@jabber.org has become available [11:26:26] --- brantgurga has become available [12:02:16] --- brantgurga has left [12:43:52] --- dev-zero@jabber.org has left [12:56:50] --- dev-zero@jabber.org has become available [14:34:36] --- mdionne has become available [15:32:58] --- dev-zero@jabber.org has left [15:35:20] --- cclausen has left [15:51:02] --- deason has left [16:01:45] --- deason has become available [16:19:05] mdionne: Nice to see you found the code review stuff ... [16:19:11] Still ironing out some niggles ... [16:22:09] I cannot believe Marcus just complained that the number of warnings in openafs is "large" [16:23:10] Well, it probably is large by most standards. [16:23:40] But if now == large, then to describe 2 years ago, you'd probably need to define a whole new language. [16:23:57] And I don't know quite where he's getting his numbers from, either ... [16:45:02] Simon: yes I was curious to give it a try - Derrick said he had no objection - I realize everything might not be 100% done. Looks pretty easy to use so far. [16:45:50] I'm just busy reviewing your parallel make patch. [16:45:56] Sadly, make dest is still hosed... [16:46:59] you mean hosed when trying to build in parallel? I don't usually try that - I'll check it out. [16:47:13] Yeh. make -j8 dest doesn't work. [16:47:20] make -j8, on the other hand works like a charm. [16:47:42] (I tried -j32, but my Mac had a heart attack mid way through) [16:48:51] I'm sure there's more dependencies missing out there - a matter of luck to stumble upon them [16:49:16] Yeh. The patch definitely makes things better, though. [16:49:39] If you have failing build logs, that would be helpful to weed out more issues [16:50:20] I will do, shortly. [16:51:28] I'll get you a whole log, but with make dest it seems to be make[4]: *** No rule to make target `single_compdir_libafs'. Stop. [16:52:51] Full log is /afs/inf.ed.ac.uk/user/s/sxw/Public/failing-make-dest.log.gz [16:54:29] Ok, will have a look, thanks. [16:55:06] No problem. [16:56:56] BTW minor typo in your "tiny acorns" PDF - the gerrit port is 29418 and not 24918 [17:03:08] Gah. I keep doing that. [17:03:15] oh and thanks for setting all this up (and the git repo)! will certainly make life easier [17:03:18] I'll see about getting that fixed. [17:04:03] No problem. I've primarily done it because I've lost far too many hours to wrestling with quilt, and trying to figure out which CVS sandbox I still hand uncommitted patches in. [17:04:48] (enlightened self-interest, I think it's called :) ) [17:05:08] quilt is a lovely workaround for not having a real VCS, but alas that's all it is. [17:05:43] Yeh. But everytime I forgot to quilt add before editing a file, another little piece of me died. [17:08:42] hmm, make -j32 dest seems happy enough here. I'll have it run in a loop for a while to see if I can trigger some failures. [17:09:22] It was pretty repeatable for me. This is a Mac OS X build, rather than Linux, so it might be something Darwin specific. [17:09:40] --- edgester has become available [17:11:09] Simon, I read your announcement on July 4. Is openafs git in production? [17:11:52] It's not yet been officially announced. [17:11:59] Some of the tools still have rough edges. [17:12:06] ah [17:12:23] The current situation is that commits into 1.5.x are on hold, and we're planning on running with the repository at git.openafs.org/ [17:12:38] ok [17:12:55] (1.4.x will, if we need another release candidate, be developed in parallel between git and CVS, and cut over once 1.4.11 is finally out. [17:12:58] I saw some git emails on the cvs-commits list [17:13:10] Yes. You'll probably see some more :) [17:13:17] :) [17:14:14] Expect a real announcement, and lots of git documentation, over the next few days. [17:16:28] is it OK to develop against git now? [17:17:33] I think it's highly unlikely at this point that we'll rollback to CVS, but until the official announcement has been made, there's always that possibility. I don't make the decisions round here :) [17:18:02] * edgester nods [17:18:13] I logged into gerrit OK [17:18:42] Cool. [17:19:12] Now I just need to figure out how to submit into gerrit [17:19:21] twitter said git announcement coming soon [17:19:24] Hang on. I'll send you a work in progress. [17:19:50] Is it ok to use an existing ssh key, or should I generate a new one? [17:20:17] You can use an existing one if you like. [17:20:36] ok, I was just wondering is there is a best practice [17:20:58] It's really down to the way that you manage your ssh keys. [17:21:27] I have a separate gerrit key, because that key lives on a load more systems than my standard personal key does. [17:21:33] --- stevenjenkins has left [17:21:59] --- abo has left [17:22:31] --- abo has become available [17:22:47] I use my personal key almost everywhere. [17:24:47] Then just use that for gerrit, too. [17:24:49] http://gerrit.openafs.org/4 is ready for review as a test (very simple fix that's been in the Debian package for a while). [17:26:57] I've added my +1 ... [17:27:19] (Either you, or another gatekeeper will need to +2 it - I'm not sufficiently endowed) [17:27:29] --- stevenjenkins has become available [17:28:34] If I understand the process correctly, I should note that I've verified the change because I've built it and am using it currently. How do I do that? I'm not seeing the UI widget -- do I have to add myself as a reviewer first? [17:28:46] Or am I not supposed to verify my own changes? [17:28:48] Click on 'Publish Comments' [17:29:07] Oh, for some reason that sounded to me like a "finalize comments I've already added" button. [17:29:09] Anyone can do anything at the moment - there are no technical restrictions on reviewing or verifying your own code. [17:29:32] Simon: does adding a dependency on setup help here (libafs/Makefile.common.in): "single_compdir single_instdir single_destdir: setup" [17:29:41] Yeh. Everyone says that on the gerrit list too. I ssupect it will change at some point. [17:30:12] Yeah, that works as expected. Sweet. [17:30:15] I really like this interface. [17:30:20] Particularly since it's very fast for me. :) [17:30:43] The little bits of Flash to copy things to the clipboard look weird with flashblock. [17:30:51] It's actually not that bad for me here, either. It uses GWT, so most of the UI is rendered in your browser. [17:31:21] mdionne: Trying ... [17:34:18] Ah, I can turn off the Flash thing. Sweet. [17:34:21] I'm really impressed by this tool. [17:34:21] mdionne: Seems to have. [17:35:24] There's a load more nice stuff in 2.0.16 which came out to day (Shawn does pretty much weekly/bi-weekly releases). But it's a good idea to let android hammer on the new code for a couple of days before deploying it - lets them have the pain. [17:35:43] * Russ is clearly going to eventually want to package this for Debian. [17:36:00] That will be less fun. I kept my installation notes. [17:37:43] Essentially, you'd need jetty to be packaged properly (Fedora claims to have packages, but they're missing bits, which really breaks things), and packages for the JDBC postgres jar, c3p0, and the Bouncy Castle Crypto jars. [17:38:13] Then packaging gerrit would be fairly straightforward :) [17:39:17] already have jetty plus jetty-extra, so I'm guessing I'm okay there. [17:39:36] libpg-java sounds promising. [17:39:39] Ah. I thought that wasn't there when we looked during the Workshop. [17:39:49] It's not in stable. [17:39:52] It's only in unstable. [17:39:55] Ah. Right. [17:40:05] Although that's right, now that you remind me, it's some ancient version. [17:40:09] Well, I could fix that if I care enough. [17:40:21] c3p0 also already packaged for Debian. [17:40:37] We're about to run gerrit at work, and that means I've been spending time looking at the Fedora situation. [17:40:38] As is Bouncy Castle Crypto. [17:40:41] Cool. [17:40:52] So basically fix jetty and this shouldn't be too hard. [17:41:01] Yes - pretty much. [17:41:30] The critical thing is being able to remove all of the demo stuff that comes with jetty. If any of that is there, then gerrit can't deploy at the the top level, which upsets it. [17:42:17] I have this hankering to add GSSAPI support to the underlying SSHD, which I'm trying to resist. [17:46:17] * Russ laughs! [17:47:52] * Russ pushes various bits from the Debian patch set into Gerrit. [17:48:16] Looking at Marc's change 3, the "Patch History" option which has appeared since I last used gerrit rocks. It makes it trivial to see when a patch is modified, what has actually been changed ... [17:55:46] http://gerrit.openafs.org/6 will rank very high on the amusement scale. [17:55:59] Oh, hm, I should add a FIXES to that. Well, that gives me an opportunity to try the patch modification thing. [18:00:06] Is there a way to have the side-by side diff include all the files? [18:01:00] --- kula has left [18:02:04] --- summatusmentis has left [18:02:42] --- summatusmentis has become available [18:11:38] Is there supposed to be a head and a master branch in the git repo? I thought HEAD==master [18:12:54] In Git, HEAD means whatever you currently have checked out, or what will be initially checked out when you clone the repository. [18:13:31] A lowercase "head" is just a synonym of "branch". [18:27:08] --- kula has become available [18:31:17] I'm getting a connection refused error when using git push [18:33:31] Do you have the ssh configuration specifying the separate port? [18:34:43] Host gerrit.openafs.org PubkeyAuthentication yes Port 24918 User jason [18:35:26] hold on [18:36:09] ugh, had the wrong port number [18:38:07] woot. I just submitted a trivial change to gerrit. [18:39:07] Looks good! Let me give you nit-picky feedback on the commit message, and then you can test the resubmit part if you want. [18:39:19] ok [18:39:35] much ado over a two character commit [18:39:38] ;) [18:39:44] yeah. :) [18:41:47] ok, so how do I edit my commit message? [18:42:21] git commit --amend in your branch where you made the original commit [18:46:28] amended and pushed [18:48:14] Submitted. [18:48:36] Ah, I don't need to add a comment on approval. [18:48:50] * Russ manually deltaized it with git update-ref. [18:49:22] thanks [18:50:13] Hm. That didn't send mail. [18:51:06] Not sure why not. [18:53:08] Made it into the repository, though. [19:06:58] cool [19:07:12] now how do I merge and delete my branch [19:08:15] Since the commit was pushed, you don't need to merge it. You can just git checkout some other branch and git branch -d to delete it. [19:09:49] There, that's most of the Debian-specific patches. Still need to figure out what to do with the bosserver permission checking difference and the fixes for the message catalog installation location for fstrace. [19:10:17] I had to run git branch -D. It said my branch wasn't an ancestor of the current head [19:10:20] And then the only patches left are the ugly PAM module hack and the compiler name and flag fixes which require more intrusive work to do properly. [19:10:38] anyways, anyways, time for bed. Good night [19:10:40] Oh, yes, you'll have that problem because the SHA1 changes because Gerrit cherry-picks. [19:10:49] ah, ok [19:10:57] --- edgester has left [19:23:24] --- mdionne has left [20:32:09] --- dev-zero@jabber.org has become available [20:58:46] use topic branches if you are submitting to gerrit [21:05:00] I think he is -- the question was what to do with the topic branch after your patch has been accepted. [21:07:08] junk it [21:43:29] --- deason has left [21:48:29] --- Russ has left: Disconnected [22:04:17] --- dev-zero@jabber.org has left [22:37:09] --- dev-zero@jabber.org has become available [22:44:07] --- reuteras has become available [23:14:52] --- kula has left