Jan 30 18:44:17 You are now talking on #RPA
Jan 30 18:44:17 --- Topic for #RPA is Ruby Production Archive -- http://rpa-base.rubyforge.org -> preliminary repository with >150 packages and port/package manager --- Best Practices: http://rpa-base.rubyforge.org/wiki/wiki.cgi?GoodPractices -- rpa-base 0.2.3 out
Jan 30 18:44:17 --- Topic for #RPA set by batsman at Sun Nov 21 03:03:15 2004
Jan 30 18:47:53 chris2|ccb enters
Jan 30 18:56:26 chris2|c1b vanishes, with the image of the words “Read error: 60 (Operation timed out)” trailing after.
Jan 30 19:28:02 chris2|c1b enters
Jan 30 19:36:34 chris2|ccb vanishes, with the image of the words “Read error: 60 (Operation timed out)” trailing after.
Jan 30 20:34:27 chris2|c1b vanishes, with the image of the words “Read error: 60 (Operation timed out)” trailing after.
Jan 30 20:42:44 chris2|ccb enters
Jan 30 21:01:13 wyhaines enters
Jan 30 21:32:07 chris2|c1b enters
Jan 30 21:40:39 chris2|ccb vanishes, with the image of the words “Read error: 60 (Operation timed out)” trailing after.
Jan 30 21:43:49 aGorilla enters
Jan 30 21:44:49 chris2|ccb enters
Jan 30 21:49:12 chris2|c1b vanishes, with the image of the words “Read error: 60 (Operation timed out)” trailing after.
Jan 30 21:49:37 aGorilla says “me and my shadow”
Jan 30 22:01:04 aGorilla says “zxy, tell everybody I said hi.  with luck, be back soon.”
Jan 30 22:01:10 aGorilla leaves.
Jan 30 22:38:17 chris2|c1b enters
Jan 30 22:46:49 chris2|ccb vanishes, with the image of the words “Read error: 60 (Operation timed out)” trailing after.
Jan 30 22:48:59 batsman :No such nick/channel
Jan 30 22:49:53 Xentac enters
Jan 30 22:50:27 Aredridel says “Xentac.”
Jan 30 22:50:30 Xentac says “Aredridel: and look who I find here”
Jan 30 22:50:31 Xentac says “;o)”
Jan 30 22:50:34 Aredridel says “Imagine!”
Jan 30 22:50:59 Aredridel says “The best time to catch the most people here is morning, since Lypanov and Chris2 are both european, as is eek when he's on.”
Jan 30 22:51:22 Aredridel says “(Xentac is a friend, and thinking about the various imperfections in revision control systems)”
Jan 30 22:51:35 Xentac says “hehehe, just thinking so far”
Jan 30 22:51:44 Xentac says “too many projects”
Jan 30 22:52:58 Aredridel says “I still think svn gets branching right. It's a copy. That's all!”
Jan 30 22:53:22 Xentac says “yeah, which is really good”
Jan 30 22:53:31 Xentac says “problem is that there's nothing else around branches”
Jan 30 22:53:48 Aredridel says “around?”
Jan 30 22:53:57 Xentac says “well, nothing else to support it”
Jan 30 22:54:03 Xentac says “such as smart merging”
Jan 30 22:54:07 Aredridel says “Ah, right.”
Jan 30 22:54:17 Xentac says “once you make a copy, it's totally independent of everything else”
Jan 30 22:54:20 Xentac says “and you're on your own”
Jan 30 22:54:26 Aredridel says “Yeah. Always feels that way to me in svn.”
Jan 30 22:54:53 Xentac says “but what exactly is needed for smart merging?”
Jan 30 22:55:19 Xentac says “I don't know a lot about it, but I know that gnu arch keeps track of where something was branched from”
Jan 30 22:55:26 Aredridel says “Yeah. As does svn, sorta.”
Jan 30 22:55:43 Xentac says “well sure, copy-wise, but not logic wise”
Jan 30 22:55:49 Xentac says “it's not something that you can easily look up”
Jan 30 22:55:51 Aredridel nods.
Jan 30 22:56:11 Xentac says “but I'm pretty sure you could keep track of branches using properties”
Jan 30 22:56:17 Aredridel says “(Ah, if only eek were here)”
Jan 30 22:56:32 Aredridel says “Yeah, though as implemented, kinda kludgy in svn.”
Jan 30 22:56:39 Aredridel says “But conceptually, yeah, it would be a property.”
Jan 30 22:56:40 Xentac says “eek knows more about branching? ;o)”
Jan 30 22:56:56 Xentac says “what's kludgy? using properties or coding them?”
Jan 30 22:57:08 Aredridel says “Using properties as svn implements them.”
Jan 30 22:57:19 Aredridel says “But if it were represented as a property, that makes sense.”
Jan 30 22:58:06 Aredridel says “Hm.”
Jan 30 22:58:08 Xentac says “you could keep track of lots of merge data through properties”
Jan 30 22:58:12 Aredridel says “Yeah.”
Jan 30 22:58:19 Aredridel says “Which revision of which branch came from where...”
Jan 30 22:58:26 Xentac says “yup”
Jan 30 22:58:40 Xentac says “then some star-merge-esque logic could sync between repositories”
Jan 30 22:58:47 Aredridel says “Though the idea of just letting the developer specify what to merge would be nice, and then have the repository cough up the difference.”
Jan 30 22:59:01 Xentac says “sure”
Jan 30 22:59:22 Xentac says “but common usage would be more like sync-all-the-recent-changes-from-repo-X”
Jan 30 22:59:36 Aredridel says “With cherry-picking when possible.”
Jan 30 22:59:41 Xentac says “of course”
Jan 30 22:59:53 Aredridel says “Mmm. That's exactly what's needed for using svn as a config file manager.”
Jan 30 23:00:07 Aredridel says “sync this-server-configs with master-config-copy.”
Jan 30 23:00:10 Xentac says “smart usage of properties?”
Jan 30 23:00:13 Xentac says “ah”
Jan 30 23:00:14 Xentac says “hehehe”
Jan 30 23:00:15 Xentac says “yeah”
Jan 30 23:00:25 Aredridel says “I struggle using svn.”
Jan 30 23:00:42 Xentac says “to manage config files?”
Jan 30 23:00:47 Aredridel says “Yeah.”
Jan 30 23:01:00 Xentac says “I don't manage enough configs to need version control ;)”
Jan 30 23:01:03 Aredridel says “For me, all the branches are functional, not frozen like release branches, but I want a master-slave setup between repos.”
Jan 30 23:01:06 Aredridel says “Hehe.”
Jan 30 23:01:25 Xentac says “so that you can make global changes as well as local changes”
Jan 30 23:01:29 Aredridel says “So I can go update the master httpd config, and have several servers get the changes merged in, but no customizations lost.”
Jan 30 23:01:31 Aredridel says “Right.”
Jan 30 23:01:45 Xentac says “yeah, that's a perfect use”
Jan 30 23:02:04 Aredridel says “And highlights a shortcoming in The Way It's Done.”
Jan 30 23:02:36 Xentac says “you mean the way subversion is done?”
Jan 30 23:02:40 Aredridel says “Yeah.”
Jan 30 23:02:59 Aredridel says “Or just what's easily possible now.”
Jan 30 23:03:08 Xentac says “yeah”
Jan 30 23:03:22 Xentac says “I think that the distributed version control paradigm is a decent one”
Jan 30 23:03:26 Aredridel says “I also want to see, at any moment, how divergent any two trees are.”
Jan 30 23:03:41 Aredridel says “Not just branches that started the same, but how different any two trees are.”
Jan 30 23:03:44 Aredridel says “That's a harder task.”
Jan 30 23:03:45 Xentac says “but I don't like the theory of patches, I still like having partial repos”
Jan 30 23:03:51 Xentac says “hmmm”
Jan 30 23:03:57 Xentac says “would you want to track changes or just differences?”
Jan 30 23:04:09 Aredridel says “Well, given common history, changes.”
Jan 30 23:04:12 Aredridel says “Without, differences.”
Jan 30 23:04:17 Xentac says “yeah”
Jan 30 23:04:33 Xentac says “it wouldn't be too bad... but yeah, it is harder”
Jan 30 23:04:35 Aredridel says “I'd be happy to inform at compare time that "foo is loosely based on x-1.6"”
Jan 30 23:04:50 Xentac says “:)”
Jan 30 23:04:51 Aredridel says “Well, the theory of patches seems to be to be the flipside...”
Jan 30 23:05:05 Aredridel says “In a perfect world, patchsets and revisioning models are identical.”
Jan 30 23:05:30 Xentac nods.
Jan 30 23:05:43 Aredridel says “In practice, finding what's changed between arbitrary trees is art, not science.”
Jan 30 23:06:07 Aredridel says “With common branch points, you get the info you need to construct a plausible best patchset. Not always the best, though.”
Jan 30 23:06:35 Xentac says “yeah”
Jan 30 23:06:35 Aredridel says “(Imagine the pathological case, where you refactor everything multiple times, re-arranging huge chunks of code, renaming variables en masse...)”
Jan 30 23:06:54 Aredridel says “(finding an optimal, least-possible-difference ... insanely hard.)”
Jan 30 23:07:01 Xentac says “at that point, diff doesn't help you at all”
Jan 30 23:07:05 Aredridel says “Right.”
Jan 30 23:07:16 Aredridel says “It's getting toward AI-complete.”
Jan 30 23:08:07 Xentac says “now I'm all interested in this problem”
Jan 30 23:08:24 Xentac says “Aredridel: you have a way of making me interested in problems that I think up :P”
Jan 30 23:08:28 Aredridel laughs
Jan 30 23:08:29 Aredridel says “I try.”
Jan 30 23:08:38 Xentac says “maybe we're just interested in the same things”
Jan 30 23:08:47 Xentac says “or maybe you just argee with me too much ;o)”
Jan 30 23:08:51 Aredridel says “The utility of patch-oriented revision control is that one can see what set of changed introduced error, and potentially reverse just that change, and also eases merging because there's a /chance/ that a patch will apply to a somewhat different tree.”
Jan 30 23:09:03 Aredridel says “Some of each, I think ;-)”
Jan 30 23:09:32 Aredridel says “The revision control problem is one I think about a lot, and somewhat unorthodoxly, since I don't tend to have problems when managing source code.”
Jan 30 23:09:40 Aredridel says “(my projects are small)”
Jan 30 23:09:44 Xentac says “yeah, but I still think generating patches when you need them will cover those cases”
Jan 30 23:09:59 Aredridel says “But I -do- think about versioning and branching of things like filesystems, config files, less orthodox stuff.”
Jan 30 23:10:03 Xentac says “I've been thinking more about version control for managing archlinux”
Jan 30 23:10:08 Aredridel nods.
Jan 30 23:10:29 Xentac says “that, are remote/disconnected development”
Jan 30 23:10:34 Aredridel says “Yep.”
Jan 30 23:10:59 Aredridel says “That's important.”
Jan 30 23:11:06 Aredridel says “I wish I could work disconnectedly better.”
Jan 30 23:11:10 Aredridel says “I'd be more productive at times.”
Jan 30 23:11:12 Xentac says “yeah”
Jan 30 23:11:19 Xentac says “when I was at school, it would have been nice”
Jan 30 23:11:26 Xentac says “but now, I'm mostly at work and home”
Jan 30 23:11:31 Aredridel says “And I when I want to spend a day in the hills.”
Jan 30 23:11:37 Xentac says “maybe if I travelled more”
Jan 30 23:12:11 Xentac says “I've been thinking about distributed version control in business too”
Jan 30 23:12:23 Aredridel says “In business?”
Jan 30 23:12:42 Xentac says “sometimes at work, we get code from customer's version control systems (cause we write/maintain it, but they want version history)”
Jan 30 23:12:42 Aredridel says “I've also been working with distributed filesystems.”
Jan 30 23:12:51 Xentac says “we used to just make copies of it, then give it back to them”
Jan 30 23:12:57 Aredridel says “Worked with Coda, looking at Lustre, using unison...”
Jan 30 23:13:15 Aredridel nods.
Jan 30 23:13:24 Xentac says “but if your customers used the same distributed version control system, you could just branch it”
Jan 30 23:13:28 Aredridel says “Right.”
Jan 30 23:13:30 Xentac says “then merge changes back”
Jan 30 23:13:31 Aredridel says “And ... replicate it.”
Jan 30 23:13:34 Xentac says “which would be sweet”
Jan 30 23:13:37 Aredridel says “Yeah.”
Jan 30 23:13:47 Aredridel says “I want merges to preserve history, too, as an option.”
Jan 30 23:14:05 Aredridel says “Kind of meta-revision-control at that point.”
Jan 30 23:14:06 Xentac says “well, how does arch do it?”
Jan 30 23:14:25 Xentac says “I'm pretty sure you can reference merged repos”
Jan 30 23:14:34 Aredridel says “I think with arch, you need access to all the archives to get the complete history.”
Jan 30 23:14:36 Xentac says “though you're accessing the remote repo”
Jan 30 23:14:41 Aredridel nods
Jan 30 23:15:00 Aredridel says “So it's seamlessish, but without independent copies.”
Jan 30 23:15:00 Xentac says “preserving history almost sounds more like what bitkeeper does...”
Jan 30 23:15:06 Aredridel says “Mmm, yeah.”
Jan 30 23:15:09 Xentac says “right”
Jan 30 23:15:35 Xentac says “I'm not sure if subversion can handle inserted changes (as opposed to appended)”
Jan 30 23:15:45 Xentac says “how exactly would you use the current system to handle something like that?”
Jan 30 23:16:07 Aredridel says “subversion can't.”
Jan 30 23:16:10 Aredridel says “Hm... ”
Jan 30 23:16:18 Aredridel says “I don't think any current system can.”
Jan 30 23:16:52 Xentac says “again, I don't know how bitkeeper works, but I hear it has very sophisticated merging logic”
Jan 30 23:17:24 Xentac says “you know, I just realized I should probably keep a backup of my blog”
Jan 30 23:17:36 Xentac says “and that I'm pretty sure I could do it well with subversion”
Jan 30 23:18:11 Aredridel says “Yeah. . . .blog is pretty simple-case. Even if you had it distributed across servers, it could merge relatively easily.”
Jan 30 23:18:19 Xentac says “yeah”
Jan 30 23:18:22 Xentac says “that was really a side note”
Jan 30 23:18:23 Xentac says “;)”
Jan 30 23:19:15 Aredridel laughs.
Jan 30 23:23:01 Xentac says “you know, I didn't even realize that svn had a merge command already”
Jan 30 23:23:04 Xentac says “well, more likely, I forgot”
Jan 30 23:23:40 Aredridel says “Hehe.”
Jan 30 23:23:53 Aredridel says “My problem is that it's hard to cherry-pick with.”
Jan 30 23:24:03 Aredridel says “It happily merges in too much with no conflicts.”
Jan 30 23:24:16 Xentac says “yeah, merge is the stupid merge”
Jan 30 23:24:39 Xentac says “but I think using properties would help merge work smarter”
Jan 30 23:25:10 Aredridel says “Yeah... what more info would help? Branch point is already taken into account in svn, but not distributedly.”
Jan 30 23:25:43 Xentac says “already merged revisions”
Jan 30 23:26:08 Aredridel says “Ah, right.”
Jan 30 23:26:09 Xentac says “I don't know how easy it is/would be to figure out branch points based on the copies”
Jan 30 23:26:26 Xentac says “so I was thinking of an explicit branch point property too”
Jan 30 23:26:32 Aredridel says “It's implicit in SVN's design. Harder if distributed. You have to be explicit.”
Jan 30 23:26:43 Xentac says “that would fix the distributed part of branching”
Jan 30 23:26:56 Aredridel nods.
Jan 30 23:27:35 Aredridel says “I like the patchset way better when merging.”
Jan 30 23:27:44 Aredridel says “But the revisions theory more when using it otherwise.”
Jan 30 23:27:53 Xentac says “right”
Jan 30 23:27:54 Aredridel says “They feel so much like the same thing.”
Jan 30 23:28:07 Xentac says “but I think you really can go from one to the other”
Jan 30 23:28:21 Aredridel says “I think so too.”
Jan 30 23:28:30 Xentac says “and I just like subversion's file system better”
Jan 30 23:28:31 Aredridel says “I think I'd love to let the developer be explicit about either or both.”
Jan 30 23:28:43 Aredridel says “Mmm, me too. Better than arch, anyway. I like darcs' too.”
Jan 30 23:28:52 Xentac says “about what way they want their repo to be?”
Jan 30 23:29:05 Aredridel says “No — about which way to view what's going on.”
Jan 30 23:29:14 Xentac says “oh”
Jan 30 23:29:25 Aredridel says “I'd love to commit, several times, then pick out patches, merge -patches-, and then push patches at another devel.”
Jan 30 23:29:57 Aredridel says “So there's an ~explicit way to view revisions as patches and vice versa.”
Jan 30 23:30:00 Xentac says “so it really comes down to what the difference between a revision and a patch/changeset is, right?”
Jan 30 23:30:04 Aredridel says “Right.”
Jan 30 23:30:25 Aredridel says “Really, it seems like a patch is a less context sensitive delta.”
Jan 30 23:30:32 Xentac says “would you want there to be a one-to-one mapping between revisions and patches?”
Jan 30 23:30:47 Aredridel says “Not neccesarily.”
Jan 30 23:31:01 Xentac says “hmmm...”
Jan 30 23:31:02 Aredridel says “I can imagine asking for the patch between revisions, then reverting half.”
Jan 30 23:31:06 Aredridel says “Manually.”
Jan 30 23:31:17 Aredridel says “And more often, seeing several revisions as a single patch.”
Jan 30 23:31:18 Xentac says “say that again”
Jan 30 23:31:33 Xentac says “ok, let me say it another way”
Jan 30 23:31:43 Xentac says “would you want revisions to be the smallest unit of a patch?”
Jan 30 23:31:54 Aredridel says “alter two functions. Commit. Ask for the difference. Half uncommit, with some context-sensitive intelligence.”
Jan 30 23:32:01 Aredridel says “Hm. Maybe.”
Jan 30 23:32:10 Aredridel says “That would be a concession that seems reasonable.”
Jan 30 23:32:18 Aredridel says “Also encourages commit often and commit early.”
Jan 30 23:32:35 Aredridel says “So if you want to be able to split patches later, mark it now by commiting a revision.”
Jan 30 23:32:42 Xentac says “as long as the developer was using their commits as if they were using the patchset theory, you'd be fine”
Jan 30 23:32:46 Aredridel says “PErhaps to a sub repository, one that's not in the main server's repo.”
Jan 30 23:32:52 Xentac says “yeah”
Jan 30 23:32:53 Aredridel says “Right.”
Jan 30 23:32:59 Xentac says “if branching becomes that easy, why not?”
Jan 30 23:33:05 Aredridel says “Totally. ”
Jan 30 23:33:34 Aredridel says “I'd love a branch-retaining-history option, where I can copy, say "all since foo-1.6" to my laptop and go road-warrior mode.”
Jan 30 23:33:45 Aredridel says “Then come back, and merge.”
Jan 30 23:34:21 Xentac says “hmmm”
Jan 30 23:34:39 Aredridel says “That's what svn -feels- lacking in, because checkouts have no history.”
Jan 30 23:34:45 Xentac says “this is going to sound kind of silly, but you'd only want the history for historical purposes, right?”
Jan 30 23:34:59 Xentac says “for looking back at”
Jan 30 23:35:05 Xentac says “not really anything functional”
Jan 30 23:35:10 Aredridel says “Nah, I'd want to be able to locally branch, then see what I did differently in each branch. ”
Jan 30 23:35:25 Aredridel says “Hm. I essentially want to be able to check out from a checkout.”
Jan 30 23:35:41 Aredridel says “Or clone a checkout.”
Jan 30 23:35:46 Xentac says “hmmm...”
Jan 30 23:35:58 Aredridel says “"or it could go this way...." and go hack.”
Jan 30 23:36:05 Xentac says “but wouldn't that mean you'd need a lot of local history?”
Jan 30 23:36:12 Aredridel says “Yeah. Local, but not long history.”
Jan 30 23:36:33 Aredridel says “And, in the main repo, I'm only going to want a more polished set of changes. Not so much of my experimenting.”
Jan 30 23:36:44 Xentac says “hmmm”
Jan 30 23:36:57 Aredridel says “Even if I checked in all my local toying including history, I'd want to be able to mark clean patch sets from that.”
Jan 30 23:37:05 Xentac says “so you'd be able to specify which revisions to combine into one revision in the master”
Jan 30 23:37:08 Aredridel says “Right.”
Jan 30 23:37:14 Xentac says “neat idea”
Jan 30 23:37:14 Aredridel says “Back to theory of patches.”
Jan 30 23:37:19 Xentac says “I like it”
Jan 30 23:37:33 Xentac says “but, I don't think that'd be hard”
Jan 30 23:37:34 Aredridel says “Which is another view on revisions here and there.”
Jan 30 23:37:41 Aredridel says “Yeah. .. the trick is in the tools to merge.”
Jan 30 23:37:52 Aredridel says “And to edit a changelog without too much pain.”
Jan 30 23:38:00 Xentac says “right”
Jan 30 23:38:11 Aredridel says “once your patches get to a certian point, reading diffs gets really tiring.”
Jan 30 23:38:30 Xentac says “worst case, create another repo, merge in revisions one at a time, commit that as a seperate revision, then merge to master”
Jan 30 23:38:45 Aredridel says “Right.”
Jan 30 23:38:47 Xentac says “that's what it comes down to”
Jan 30 23:38:49 Aredridel says “Hm.”
Jan 30 23:39:12 Aredridel says “That makes total sense. making a new repo is like checking out.”
Jan 30 23:39:26 Aredridel says “So you make a fresh ~checkout to merge into, then merge that.”
Jan 30 23:39:31 Aredridel says “That sounds a -lot- like darcs.”
Jan 30 23:39:36 Xentac says “yeah”
Jan 30 23:39:41 Aredridel says “'cept that storing as revisions seems faster.”
Jan 30 23:39:51 Xentac says “seems?”
Jan 30 23:40:07 Aredridel says “Yeah, except that darcs feels like a sea of patches sometimes.”
Jan 30 23:40:17 Xentac says “right”
Jan 30 23:40:40 Aredridel says “(which is uncomfortable to me)”
Jan 30 23:41:11 Xentac says “yeah, I don't really like the feeling of everything being a patch”
Jan 30 23:41:22 Xentac says “I don't mind it for transitioning (merging)”
Jan 30 23:41:24 Aredridel nods.
Jan 30 23:41:24 Xentac says “but not for everything”
Jan 30 23:41:26 Aredridel says “Yeah. ”
Jan 30 23:41:32 Aredridel says “Actually, any time you transmit, it's a patch.”
Jan 30 23:41:47 Aredridel says “Perhaps it's a patch you -know- has the right context (shared history lets you be sure)”
Jan 30 23:41:50 Aredridel says “But a patch still.”
Jan 30 23:41:58 Aredridel says “But when you commit, it can be stored as a true delta.”
Jan 30 23:41:59 Xentac says “right”
Jan 30 23:42:10 Xentac says “true delta..?”
Jan 30 23:42:10 Aredridel says “context sensitive to the bit if it so suits.”
Jan 30 23:42:24 Aredridel says “Patch being a "Weak" delta, one with less context sensitivity.”
Jan 30 23:42:35 Xentac says “ah”
Jan 30 23:42:43 Aredridel says “(and a higher chance of applying somewhere it shouldn't.)”
Jan 30 23:42:51 Xentac says “yeah”
Jan 30 23:46:56 Xentac says “so... you like my idea? ;)”
Jan 30 23:47:20 Aredridel says “Quite.”
Jan 30 23:47:41 Aredridel says “Sounds like major adaptation in svn, or underlying storage shift in darcs.”
Jan 30 23:47:49 Aredridel says “Both big, but both doable.”
Jan 30 23:48:02 Aredridel says “I'd love a way to have them both talk. A way to synch history.”
Jan 30 23:48:11 Xentac says “hehehe”
Jan 30 23:48:15 Xentac says “sync between darcs and svn, eh?”
Jan 30 23:48:15 Aredridel says “So one can use, locally, whatever RCS is nicest.”
Jan 30 23:48:28 Aredridel says “And have a way to send patchsets, and optionally patchsets since a branchpoint.”
Jan 30 23:48:29 Xentac says “you'd need someone who knew both systems *really* well”
Jan 30 23:48:30 Aredridel says “Yeah.”
Jan 30 23:48:39 Aredridel says “Well, or a clear protocol for patches.”
Jan 30 23:48:58 Xentac says “you can't have shared history then”
Jan 30 23:49:06 Aredridel says “Could.”
Jan 30 23:49:24 Xentac says “unless you were making full copies”
Jan 30 23:49:32 Xentac says “well... full copies of what you wanted”
Jan 30 23:49:45 Aredridel says “Right. Disconnected development implies that anyway.”
Jan 30 23:49:52 Xentac says “true”
Jan 30 23:50:29 Xentac says “I bet that's probably doable too... though it'd take a lot of work to get a working protocol or implement the other's protocol in the project”
Jan 30 23:50:54 Aredridel nods.
Jan 30 23:51:04 Aredridel says “I don't think any protocol at the moment is entirely adequare.”
Jan 30 23:51:27 Aredridel says “Since darcs is already patchsetish, it could be adapted more easily.”
Jan 30 23:51:31 Aredridel says “But that's only synching one level.”
Jan 30 23:51:42 Aredridel says “merging in history is harder. Perhaps unneeded.”
Jan 30 23:55:22 Xentac says “it'd be a nice feature to have, but I don't know if it'd be a showstopper”
Jan 30 23:55:30 Aredridel says “Hm.”
Jan 30 23:55:57 Xentac says “anyway, version control interoperation wasn't really in the scope of my original idea ;)”
Jan 30 23:56:15 Aredridel laughs.
Jan 30 23:58:19 Xentac says “I wonder who I should talk to now about it...”
Jan 30 23:58:35 Aredridel says “Hm.”
Jan 30 23:58:49 Aredridel says “Yeah... it's a big project to tackle. . . perhaps svk ?”
Jan 30 23:58:52 Aredridel says “Though perl. Ick.”
Jan 30 23:59:12 Xentac says “well, I still think it could be implemented directly in svn”
Jan 30 23:59:46 Aredridel says “Yeah, it could. It'd have to start outside and get somewhat popular first for the svn guys to accept it, politically.”
Jan 30 23:59:54 Xentac says “yeah”
Jan 31 00:00:08 Xentac says “but I'm pretty sure I couldn't implement it by myself ;)”
Jan 31 00:00:12 Aredridel says “Mmm, yeah..”
Jan 31 00:00:15 Xentac says “hmmm”
Jan 31 00:00:28 Xentac says “I just thought of something else you could do with this that you couldn't do with a patchset based system”
Jan 31 00:00:35 Aredridel says “yeah?”
Jan 31 00:00:42 Xentac says “have multiple branches into seperate directories in one repo”
Jan 31 00:00:53 Aredridel says “I don't follow.”
Jan 31 00:01:10 Xentac says “I'm not sure of the usefullness”
Jan 31 00:01:14 Xentac says “... but”
Jan 31 00:01:27 Xentac says “you could create a branch of repo X”
Jan 31 00:01:39 Xentac says “then in a directory Y in X (X/Y) branch another repo Z”
Jan 31 00:02:19 Xentac says “and you can also do subdirectory branching”
Jan 31 00:02:36 Xentac says “so you could create branches of directories within repos, instead of the whole repo”
Jan 31 00:02:39 Aredridel says “Hm, yeah.”
Jan 31 00:02:47 Aredridel says “That's something I do in svn and love.”
Jan 31 00:02:52 Xentac says “yeah”
Jan 31 00:02:59 Xentac says “me too”
Jan 31 00:03:10 Xentac says “it's great cause I have a big arch-tools repo”
Jan 31 00:03:16 Xentac says “and I can point people to projects within it”
Jan 31 00:03:21 Aredridel says “Right!”
Jan 31 00:05:01 Aredridel says “Hm. Could be done patchset-based, but ick. Painful.”
Jan 31 00:05:13 Aredridel says “I should go sleep.”
Jan 31 00:05:18 Xentac says “yeah... none of them try to support it (that I know of)”
Jan 31 00:05:21 Xentac says “ok”
Jan 31 00:05:23 Xentac says “have a good one”
Jan 31 00:05:29 Xentac says “I should probably go to sleep too”
Jan 31 00:05:36 Xentac says “6:30 comes early these days”
Jan 31 00:05:39 Aredridel says “That it does.”
Jan 31 00:06:01 Aredridel says “Night, all. (someone remind me to send some of this log to eek and lypanov, eh?)”
Jan 31 00:07:55 --- Disconnected ().
**** ENDING LOGGING AT Mon Jan 31 00:07:55 2005