[00:12:25] --- manfred furuholmen has become available [00:30:07] --- Russ has left: Disconnected [00:30:37] --- abo has become available [01:15:54] --- manfred furuholmen has left [02:35:28] --- haba has left [04:48:29] --- dev-zero@jabber.org has become available [04:48:35] --- dev-zero@jabber.org has left: offline [05:04:31] Another day, another git repo. http://openafs-git.stanford.edu/git/openafs-test-20090527.git/ (git web at http://openafs-git.stanford.edu/?p=openafs-test-20090527.git;a=summary ) [05:05:41] This addresses all of yesterdays issues, and adds a couple of known new ones - we're not expanding commiter (rather than author) identities, and non-ASCII characters in author names aren't being correctly stored... [05:28:08] --- Jeffrey Altman has left: Replaced by new connection [05:33:00] --- Jeffrey Altman has become available [05:55:53] --- cclausen has become available [06:16:15] --- mmeffie has become available [06:20:49] --- dev-zero@jabber.org has become available [06:20:54] --- dev-zero@jabber.org has left: offline [06:39:05] --- reuteras has left [07:21:32] Well, given that previously you weren't expanding committer _or_ author identites (right?), and that both ASCII and non-ASCII characters in author names weren't being stored, that sounds like an improvement. [07:22:09] Yeh. It is an improvement, just needs at least one more iteration. [07:22:50] The log concatenation code also needs a bit of work too - at moment you get the delta name at the start of each concatenated log fragment, which is a little boring. [07:49:38] 2742 Derrick Brashear 9 Derrick Brashear 8 shadow 2 Derrick Brashear 1 Derrick Brashear 1 Derrick Brashear [07:50:47] shadow , and demenetia.org should be fixed, but I'm not sure we can legitimately map all of the others together. [07:51:00] Derrick might have reasons for committing from different places. [07:51:42] For example, say some of my contributions were personal, and others paid for by my employer. I might want to use simon@sxw.org.uk and sxw@inf.ed.ac.uk (respectively) to note that. [07:52:48] I, too, do. [07:53:21] the sjenkins@sinenomine.net & steven@endpoint.com were paid for by the respective companies. The steven.jenkins@gmail.com were done on my own time. [07:59:07] Are you planning on talking about the git conversion as part of the talk you're already giving? [07:59:43] I was planning on giving an overview of how to drive both git and gerrit as part of the 'contributing to OpenAFS' talk. [08:00:02] I wasn't planning on talking in that much detail about how the conversion has actually been done. [08:00:51] OK. [08:02:30] I guess while we're at it, I should be "Alejandro R. Sedeño" [08:03:43] I'll fix that spelling in the next run. [08:04:05] Hrm. I've been reading it that way, but I think this is the first confirmation I've seen that that's the correct spelling. [08:04:58] I've been slowly migrating to using it as character representation has become more uniform. Thanks, Unicode. [08:05:02] Hm. Who from MIT/SIPB is coming to the workshop? [08:05:06] I'm also going to lower case the domain portion of all email addresses ... [08:05:43] I am. I know others will be in the area and we might see them outside of the workshop. [08:08:19] > Derrick might have reasons for committing from different places. [08:08:34] i do "new version" commits as shadow@openafs.org, since they're administrative [08:09:33] otherwise, it's usually shadow@dementia.org for everything i do, despite the fact that most mail from me is from gmail now [08:11:28] > i do "new version" commits as shadow@openafs.org Or rather, you did, exactly once, if Simon's count is to be believed. [08:32:24] --- matt has become available [09:20:15] “Steven N . Hirsch” has an extra space. [09:24:48] It was probably that way in the email [09:25:08] Is it possible to fix up things like that after generating the repository? [09:25:18] i assume so [09:27:49] --- stevenjenkins has left [09:28:06] --- stevenjenkins has become available [09:37:15] --- mmeffie has left [10:35:05] Whether you can change things after the repo is generated depends on whether they're included in the hash of the commit or not. [10:35:50] If they're in the hash, changing things will create new commit IDs, and new branch histories, and everyone would have to rebase to the new tree. [10:36:06] I think it's best to plan to be able to change nothing once we flip the switch. [10:36:36] ideally [10:37:21] Basically, if you change an object, you change the hash of every object which descends from it. So, changing your history is painful. [10:40:09] From what I can see online, the author and committer identifiers are part of the commit object, and so part of its hash. So, to answer jhutz's question, you can't change them without a flag day. [10:40:26] Oh, I meant after the mechanical run but before we flip the switch. Like, if one author name is wrong ini one place, just fix it instead of doing another 5-hour run [10:41:16] Yeh, we could modify that commit, and rebase the rest of the tree on top of it. That's entirely doable at that stage. [10:42:19] --- Russ has become available [10:51:58] --- dev-zero@jabber.org has become available [11:02:31] --- haba has become available [12:27:36] rebase modifies all the commit dates, and is generally a pain when there are lots of branches to move; you actually want filter-branch for this. [12:28:44] Personally, I think I'd rather just eat the reimport time, than start messing around in the middle of the commit history. But that's good to know, anyway, for future reference. [12:35:28] I'm in the process of doing another import run which should address all of today's issues. Providing nothing else major comes up, I'm going to base gerrit and what I show at the workshop on that run. [12:35:52] ok [13:22:39] --- cclausen has left [14:45:58] --- cclausen has become available [14:52:50] --- haba has left [17:51:36] --- edgester has become available [19:00:27] --- edgester has left [21:18:22] --- haba has become available [22:09:45] --- dev-zero@jabber.org has left: offline [22:10:55] --- haba has left [22:21:49] --- Russ has left: Disconnected [23:36:24] --- reuteras has become available [23:59:08] --- dev-zero@jabber.org has become available [23:59:14] --- dev-zero@jabber.org has left: offline