Auth: no Time: Sat Apr 17 13:15:36 2004 Host: pallas.eruditorum.org From: They only want you when you're 17. When you're 21, you're no fun dude. svk now has support for mirroring remote cvs and p4 repositories Auth: yes Time: Sat Apr 17 13:16:16 2004 Host: MASS-TOOLPIKE.MIT.EDU From: Disengineer What's svk? Auth: no Time: Sat Apr 17 13:16:56 2004 Host: pallas.eruditorum.org From: They only want you when you're 17. When you're 21, you're no fun svk.elixus.org Auth: no Time: Sat Apr 17 13:17:13 2004 Host: pallas.eruditorum.org From: They only want you when you're 17. When you're 21, you're no fun it's a distributed revision control system built on top of svn Auth: no Time: Sat Apr 17 13:18:00 2004 Host: pallas.eruditorum.org From: They only want you when you're 17. When you're 21, you're no fun my favorite feature is that it can do smart, repeated cross-branch merges. Auth: yes Time: Sat Apr 17 13:24:00 2004 Host: ERROR-MESSAGES.MIT.EDU From: solitary, poor, nasty, brutish, and short Presumably there are limitations on how effectively it can push to an upstream CVS repository, because of CVS's lack of interesting metadata storage. Auth: no Time: Sat Apr 17 13:24:10 2004 Host: pallas.eruditorum.org From: They only want you when you're 17. When you're 21, you're no fun I don't know how it's doing that, no Auth: yes Time: Sat Apr 17 13:25:01 2004 Host: ERROR-MESSAGES.MIT.EDU From: solitary, poor, nasty, brutish, and short (E.g. if you and I are both maintaining svk branches against the Athena repository, and we're trading patches back and forth, and we both decide to push upstream to the Athena repository, it doesn't seem like the Athena repository can really keep track of which changes it already has.) Auth: yes Time: Sat Apr 17 13:25:08 2004 Host: gsstark.mtl.istop.com From: My mind's eyelash can you push to two upstreams? effectively giving your svn repository a CVS backup? Auth: no Time: Sat Apr 17 13:25:30 2004 Host: pallas.eruditorum.org From: They only want you when you're 17. When you're 21, you're no fun gsstark: just uses SVN::Mirror for that Auth: no Time: Sat Apr 17 13:25:36 2004 Host: pallas.eruditorum.org From: They only want you when you're 17. When you're 21, you're no fun greg: asking clkao now Auth: no Time: Sat Apr 17 13:25:44 2004 Host: pallas.eruditorum.org From: They only want you when you're 17. When you're 21, you're no fun greg _hudson_: asking clkao now Auth: no Time: Sat Apr 17 13:28:38 2004 Host: pallas.eruditorum.org From: They only want you when you're 17. When you're 21, you're no fun ah. svk does not yet sync metadata back to cvs. he's got something of a plan, but it's not done yet. Auth: no Time: Tue May 4 22:34:27 2004 Host: TIR-NA-NOGTH.MIT.EDU From: Sam Hartman Have people played with creating a useful svk mirror of a cvs repository that is still under active use? Auth: no Time: Tue May 4 22:34:52 2004 Host: pallas.eruditorum.org From: Do I @i(look) like I do things manually? sam: that's functionality that landed in svk about 2 weeks ago. clkao claims it works, but there may still be dragons Auth: no Time: Tue May 4 22:35:28 2004 Host: TIR-NA-NOGTH.MIT.EDU From: Sam Hartman oooooo, shiny. Auth: no Time: Tue May 4 22:36:06 2004 Host: TIR-NA-NOGTH.MIT.EDU From: Sam Hartman Perhaps the Debian Kerberos package is also going to move to svk Auth: no Time: Tue May 4 22:36:27 2004 Host: TIR-NA-NOGTH.MIT.EDU From: Sam Hartman Although of course I'll have to play with all of this before trusting it. Auth: yes Time: Tue May 4 22:44:48 2004 Host: INNOCUOUS.MIT.EDU From: free as in with your paid subscription the svk page says I can continue to use svn tools to work with a repository that I am using through svk. It doesn't make clear how good an idea this is. For example, if I work on a project using an svn server, and a few of the developers use svk and the rest of us don't, would everyone be happy? Auth: no Time: Tue May 4 22:47:25 2004 Host: TIR-NA-NOGTH.MIT.EDU From: Sam Hartman Presumably you are happier if svk users do the merges? Auth: yes Time: Tue May 4 22:50:50 2004 Host: INNOCUOUS.MIT.EDU From: free as in to stay I wasn't clear on that. It seemed like (maybe) the svk users would be forced to do merges, since there wouldn't be anything to merge from the PoV of the regular svn users. Auth: no Time: Tue May 4 22:51:39 2004 Host: pallas.eruditorum.org From: Do I @i(look) like I do things manually? No, anyone can do a merge. but doing a merge with svn won't add the clever bits of metadata that let svk do smart merging Auth: yes Time: Tue May 4 22:51:49 2004 Host: INNOCUOUS.MIT.EDU From: free as in loader Basically I'm wonder if svn/svk supports a model where the cool kids use svk and people who don't want to deal (eg, windows users who are just using it as a network filesystem) use regular svn tools. Auth: no Time: Tue May 4 22:52:41 2004 Host: pallas.eruditorum.org From: Do I @i(look) like I do things manually? Right now, at best practical, I use svk most of the time, autrijus uses svk all the time, linda uses svn all the time and robert uses svn all the time Auth: yes Time: Tue May 4 22:53:13 2004 Host: INNOCUOUS.MIT.EDU From: free as in unoccupied And you just don't have linda do merges? Or is smart merging not all that big a deal? Auth: yes Time: Tue May 4 22:53:29 2004 Host: INNOCUOUS.MIT.EDU From: free as in information But it sounds like this dual-use model makes svn/svk attractive to me. Auth: yes Time: Tue May 4 22:53:49 2004 Host: INNOCUOUS.MIT.EDU From: free as in and easy Since I didn't relish (A) learning bitkeeper and (B) teaching bitkeeper to other people. :) Auth: no Time: Tue May 4 22:53:49 2004 Host: pallas.eruditorum.org From: Do I @i(look) like I do things manually? I don't have linda do merges Auth: no Time: Tue May 4 22:54:18 2004 Host: pallas.eruditorum.org From: Do I @i(look) like I do things manually? note also that for the price of one seat of bitkeeper, you can get an awful lot of feature work done to svk Auth: no Time: Tue May 4 22:54:28 2004 Host: DOGBERT.IHTFP.ORG From: The Warlord BK is pretty easy... and the interface is self-documenting Auth: yes Time: Tue May 4 22:54:31 2004 Host: INNOCUOUS.MIT.EDU From: free as in refills So the svn-only people see branches, and can work on the head of any branch, and they just avoid doing cross branch merges? Auth: yes Time: Tue May 4 22:54:58 2004 Host: INNOCUOUS.MIT.EDU From: free as in hold BK hurts me every time I try to use it, and I consider myself smart. And I don't know what "self-documenting" would even mean. Auth: no Time: Tue May 4 22:55:08 2004 Host: pallas.eruditorum.org From: Do I @i(look) like I do things manually? (I'm underwriting use of custom external merge tools for svk, so that I can do clever things with Filemerge.app during smart merge conflict resolution) Auth: no Time: Tue May 4 22:55:28 2004 Host: DOGBERT.IHTFP.ORG From: The Warlord "bk help" Auth: yes Time: Tue May 4 22:55:36 2004 Host: INNOCUOUS.MIT.EDU From: free as in warren I'd actually love to get a decent class/doc on how to use BK that was targetted at my usage model (tracking development of a product that uses bk), but I haven't found one. Auth: yes Time: Tue May 4 22:55:47 2004 Host: INNOCUOUS.MIT.EDU From: free as in fire bk help is equivalent to man pages. Man pages are not a new invention. Auth: no Time: Tue May 4 22:56:12 2004 Host: pallas.eruditorum.org From: Do I @i(look) like I do things manually? svk is actually pretty good at dealing with resolving changes made on two branches at the same time (as would be the case if an svn user did a cross-branch merge) Auth: no Time: Tue May 4 22:57:33 2004 Host: TIR-NA-NOGTH.MIT.EDU From: Sam Hartman So the summary seems to be that you clearly won't hurt your self more than using svn all the time by mixing svk and svn but that you may lose some features if you use svn instead of svk especially for merges. Auth: no Time: Tue May 4 22:57:47 2004 Host: pallas.eruditorum.org From: Do I @i(look) like I do things manually? that sounds correct, sam. Auth: no Time: Tue May 4 22:58:31 2004 Host: DOGBERT.IHTFP.ORG From: The Warlord bk is easy to explain if someone understands the concepts of repositories, deltas, and changesets. Auth: no Time: Tue May 4 23:03:30 2004 Host: DOGBERT.IHTFP.ORG From: The Warlord any chance we'll have svn/svk in athena-9.3? ;) Auth: yes Time: Tue May 4 23:04:19 2004 Host: ERROR-MESSAGES.MIT.EDU From: solitary, poor, nasty, brutish, and short In the release? No chance. It could go into a locker any time, of course. Auth: yes Time: Tue May 4 23:04:59 2004 Host: ERROR-MESSAGES.MIT.EDU From: solitary, poor, nasty, brutish, and short FSFS support would make svn much more useful on Athena, since we're very fileystem-centric. Auth: yes Time: Thu May 6 15:48:51 2004 Host: TIR-NA-NOGTH.MIT.EDU From: Sam Hartman VCP is a pile of crap. It seems like a useful pile of crap though Auth: no Time: Thu May 6 15:49:11 2004 Host: pallas.eruditorum.org From: Do I @i(look) like I do things manually? don't install VCP unless you _need_ p4 or cvs support Auth: yes Time: Thu May 6 15:49:42 2004 Host: TIR-NA-NOGTH.MIT.EDU From: Sam Hartman Seems sort of hard to evaluate the cvs support without it;) Auth: no Time: Thu May 6 15:49:41 2004 Host: pallas.eruditorum.org From: Do I @i(look) like I do things manually? VCP really assumes you're using it with perforce Auth: yes Time: Thu May 6 15:50:30 2004 Host: TIR-NA-NOGTH.MIT.EDU From: Sam Hartman svk: running cvs log krb5 Hmm, I bet this may take a while. Auth: yes Time: Thu May 6 15:51:55 2004 Host: TIR-NA-NOGTH.MIT.EDU From: Sam Hartman O, perhaps not that long after all: ***BUG REPORT*** $rev_id == undef Please see /dev/null Auth: no Time: Thu May 6 15:56:02 2004 Host: pallas.eruditorum.org From: Do I @i(look) like I do things manually? see the "VCP" perl module in cpan, ctl Auth: yes Time: Thu May 6 15:56:32 2004 Host: TIR-NA-NOGTH.MIT.EDU From: Sam Hartman OK, hmm, does this work for any repository at all? Auth: no Time: Thu May 6 15:57:10 2004 Host: pallas.eruditorum.org From: Do I @i(look) like I do things manually? let me see if clkao is awake. Auth: yes Time: Thu May 6 15:57:19 2004 Host: ERROR-MESSAGES.MIT.EDU From: solitary, poor, nasty, brutish, and short "Please see /dev/null" is a pretty funny error message. Auth: no Time: Thu May 6 15:57:41 2004 Host: pallas.eruditorum.org From: Do I @i(look) like I do things manually? last seen 5 hours ago. Auth: yes Time: Thu May 6 15:58:16 2004 Host: SMALL-GODS.MIT.EDU From: Morgenmuffel It's a bit early for clkao. Auth: yes Time: Thu May 6 15:59:24 2004 Host: TIR-NA-NOGTH.MIT.EDU From: Sam Hartman OK, so on another repository it fails with cvs:/afs/sipb/project/debian/cvs:pam but seems to work with cvs::ext:zb.mit.edu:/afs/sipb/project/debian/cvs:pam Auth: yes Time: Thu May 6 16:00:10 2004 Host: TIR-NA-NOGTH.MIT.EDU From: Sam Hartman Note that I'm just playing around here looking at technologies. Auth: no Time: Thu May 6 16:01:47 2004 Host: pallas.eruditorum.org From: Do I @i(look) like I do things manually? *nod* I do know dave rolsky made it go, but needed patches to use ssh auth Auth: yes Time: Thu May 6 16:08:13 2004 Host: TIR-NA-NOGTH.MIT.EDU From: Sam Hartman Can a depot path be a svn url? I.E. do I actually have to have a local repository? Auth: no Time: Thu May 6 16:08:43 2004 Host: pallas.eruditorum.org From: Do I @i(look) like I do things manually? you'll lose the ability to work offline, but yes.. Auth: no Time: Thu May 6 16:08:59 2004 Host: pallas.eruditorum.org From: Do I @i(look) like I do things manually? svk depotmap will let you set up your depot as an svn repository Auth: yes Time: Thu May 6 16:15:28 2004 Host: TIR-NA-NOGTH.MIT.EDU From: Sam Hartman Hmm, I actually see a lot of things in SVK::XD that want insert file:// in front of a repository path. Auth: no Time: Thu May 6 16:16:10 2004 Host: pallas.eruditorum.org From: Do I @i(look) like I do things manually? ugh. Auth: no Time: Thu May 6 16:16:36 2004 Host: pallas.eruditorum.org From: Do I @i(look) like I do things manually? though that's sort of a good sign for "maybe it's s a SMOP rather than a LMOP Auth: yes Time: Thu May 6 16:20:40 2004 Host: TIR-NA-NOGTH.MIT.EDU From: Sam Hartman Yeah. Although actually I think I really only care about svm for this. Auth: no Time: Thu May 6 16:21:11 2004 Host: pallas.eruditorum.org From: Do I @i(look) like I do things manually? that's fair. are you looking to go from cvs->svn with svm or do something else? Auth: yes Time: Thu May 6 16:24:24 2004 Host: TIR-NA-NOGTH.MIT.EDU From: Sam Hartman Possibly. I'm still evaluating options. One use case I have is the Debian pam package. I need to: * one time convert a cvs repository * continually track the linux-pam cvs repository * do local development All of this should live on a svn.debian.org because other developers need access. On top of that I may want a local svk for my own needs Auth: no Time: Thu May 6 16:28:00 2004 Host: pallas.eruditorum.org From: Do I @i(look) like I do things manually? so. it _should be possible_ for you to do constant mirroring from both svn.debian.org and the linux-pam cvs repository to a local svk where you smart-merge them back to svn.debian.org Auth: yes Time: Thu May 6 16:30:30 2004 Host: TIR-NA-NOGTH.MIT.EDU From: Sam Hartman You lose a lot of history that way; each entire merge is one revision Auth: no Time: Thu May 6 16:31:07 2004 Host: pallas.eruditorum.org From: Do I @i(look) like I do things manually? clkao is working on replay smartmerge. Auth: yes Time: Thu May 6 16:51:24 2004 Host: TIR-NA-NOGTH.MIT.EDU From: Sam Hartman Hmm, it actually seems to use SVN::repos in SVN::Mirror. I wonder if it uses it for anything that it couldn't get through RA Auth: yes Time: Thu May 6 17:34:38 2004 Host: TIR-NA-NOGTH.MIT.EDU From: Sam Hartman OK, I really don't know svn very well, but it seems like there is significant overlap between parts of svn_ra and svn_fs. And the SVN::Mirror code is written to use svn_fs. Auth: no Time: Thu May 6 17:46:16 2004 Host: pallas.eruditorum.org From: Do I @i(look) like I do things manually? I'll ask clkao when I see him Auth: yes Time: Thu May 6 17:46:53 2004 Host: TIR-NA-NOGTH.MIT.EDU From: Sam Hartman If there is a mailing list perhaps it would be more useful for me to write up what I'm trying to do and ask what if any options I have. Auth: no Time: Thu May 6 17:47:44 2004 Host: pallas.eruditorum.org From: Do I @i(look) like I do things manually? svk-dev@lists.openfoundry.org Auth: no Time: Thu May 6 17:48:00 2004 Host: pallas.eruditorum.org From: Do I @i(look) like I do things manually? it's a sympa list and acled for subscriber posting only Auth: no Time: Thu May 6 21:03:40 2004 Host: pallas.eruditorum.org From: Do I @i(look) like I do things manually? smartmerge rocks my world Auth: yes Time: Sat May 8 17:17:22 2004 Host: TIR-NA-NOGTH.MIT.EDU From: Sam Hartman http://www.livejournal.com/users/hartmans/4542.html Is my comments and initial impressions of svk Auth: yes Time: Sun May 9 18:40:30 2004 Host: ERROR-MESSAGES.MIT.EDU From: solitary, poor, nasty, brutish, and short svk now supports "svk cmd --help" and value after single-character option with no whitespace in between, as a result of Sam's livejournal. Auth: yes Time: Sun May 9 18:40:56 2004 Host: INNOCUOUS.MIT.EDU From: free as in trader Free software is good like that. Auth: no Time: Sun May 9 18:40:52 2004 Host: pallas.eruditorum.org From: Do I @i(look) like I do things manually? mm. blame-motivated development Auth: no Time: Sat May 15 13:41:10 2004 Host: pallas.eruditorum.org From: They only want you when you're 17. When you're 21, you're no fun oh good. clkao found and fixed sam's merge bug Auth: yes Time: Sat May 15 17:54:45 2004 Host: TIR-NA-NOGTH.MIT.EDU From: Sam Hartman Such a simple fix;) Auth: no Time: Sat May 15 17:54:46 2004 Host: pallas.eruditorum.org From: They only want you when you're 17. When you're 21, you're no fun hm? Auth: yes Time: Sat May 15 17:57:54 2004 Host: TIR-NA-NOGTH.MIT.EDU From: Sam Hartman Adding a slash seems to have been what fixed my problem Auth: yes Time: Sat May 15 17:58:17 2004 Host: TIR-NA-NOGTH.MIT.EDU From: Sam Hartman Now working on writing the next bug report;) Auth: no Time: Sat May 15 18:01:44 2004 Host: pallas.eruditorum.org From: They only want you when you're 17. When you're 21, you're no fun cool Auth: no Time: Sun May 16 16:33:13 2004 Host: pallas.eruditorum.org From: They only want you when you're 17. When you're 21, you're no fun ooh. it looks like svk+rt will be getting "click to commit" Auth: yes Time: Sun May 16 16:34:07 2004 Host: JOHNSTOWN.andrew.cmu.edu From: Smart is sexy. more incentive to want to ditch cvs Auth: yes Time: Sun May 16 16:34:18 2004 Host: central-air-conditioning.toybox.cambridge.ma.us From: Marc Horowitz what is click to commit? Auth: yes Time: Sun May 16 16:34:24 2004 Host: JOHNSTOWN.andrew.cmu.edu From: Smart is sexy. too bad most of the world will want me not to Auth: no Time: Sun May 16 16:34:44 2004 Host: pallas.eruditorum.org From: They only want you when you're 17. When you're 21, you're no fun someone attaches a patch to a ticket. UI to apply the patch and commit Auth: yes Time: Sun May 16 16:35:23 2004 Host: central-air-conditioning.toybox.cambridge.ma.us From: Marc Horowitz wow, that's trusting. Auth: no Time: Sun May 16 16:35:25 2004 Host: pallas.eruditorum.org From: They only want you when you're 17. When you're 21, you're no fun er, presumably you'd _review_ the patch first Auth: no Time: Sun May 16 16:35:36 2004 Host: pallas.eruditorum.org From: They only want you when you're 17. When you're 21, you're no fun but svk also has gpg patchsigning support Auth: yes Time: Sun May 16 16:36:14 2004 Host: central-air-conditioning.toybox.cambridge.ma.us From: Marc Horowitz unless the patch is trivial, I'd want to do that by applying it and running tests and stuff. I suppose if the apply/commit was branched, it could be a feature I'd want to use. Auth: no Time: Sun May 16 16:36:59 2004 Host: pallas.eruditorum.org From: They only want you when you're 17. When you're 21, you're no fun 90% of the patches I get are trivial and can be verified by inspection Auth: no Time: Sun May 16 16:37:20 2004 Host: pallas.eruditorum.org From: They only want you when you're 17. When you're 21, you're no fun "Yes, that anchor _is_ outisde the blink tag" Auth: yes Time: Sun May 16 16:39:47 2004 Host: TIR-NA-NOGTH.MIT.EDU From: Sam Hartman I wonder how hard it would be to do cvs mirroring in the other direction. Auth: no Time: Sun May 16 16:39:54 2004 Host: pallas.eruditorum.org From: They only want you when you're 17. When you're 21, you're no fun clkao is working on it Auth: yes Time: Sun May 16 16:40:22 2004 Host: TIR-NA-NOGTH.MIT.EDU From: Sam Hartman So you could migrate things to svk even though you know a lot of people who still use cvs Auth: no Time: Sun May 16 16:40:25 2004 Host: pallas.eruditorum.org From: They only want you when you're 17. When you're 21, you're no fun oh. you mean a cvs readonly mirror? Auth: yes Time: Sun May 16 16:41:09 2004 Host: TIR-NA-NOGTH.MIT.EDU From: Sam Hartman Yes. The svk mirror of a cvs repository works but isn't really useful for any of the repsitories I care about Auth: no Time: Sun May 16 16:41:23 2004 Host: pallas.eruditorum.org From: They only want you when you're 17. When you're 21, you're no fun there are solutions out there for doing svn->cvs as a read-only mirror Auth: no Time: Sun May 16 16:41:35 2004 Host: pallas.eruditorum.org From: They only want you when you're 17. When you're 21, you're no fun at one point, iirc, ghudson was talking about a cvs pserver for svn too Auth: no Time: Sun May 16 16:41:50 2004 Host: hydra.hamachi.org From: Bill Sommerfeld then you add a wrapper script which redirected "cvs commit" into the equivalent svk.... Auth: no Time: Sun May 16 16:45:14 2004 Host: pallas.eruditorum.org From: They only want you when you're 17. When you're 21, you're no fun fear Auth: yes Time: Sun May 16 16:47:04 2004 Host: EGYPTIAN-GODS.MIT.EDU From: solitary, poor, nasty, brutish, and short I haven't thought about the pserver thing in a while; I suspect the CVS protocol is screwy enough to discourage me. Auth: yes Time: Sun May 16 16:47:35 2004 Host: EGYPTIAN-GODS.MIT.EDU From: solitary, poor, nasty, brutish, and short I have, however, seen a post-commit hook which translates an svn commit into a CVS commit. I assume that's the sort of thing Jesse is talking about. Auth: yes Time: Sun May 16 16:52:41 2004 Host: kirameki-neko.thok.org From: The meek will not inherit the Earth. The lawyers challenged God's Will and had it overturned. any protocol with messages like "I LOVE YOU" and "I HATE YOU" pretty much counts as screwy, yes Auth: yes Time: Sun May 16 16:54:05 2004 Host: JOHNSTOWN.andrew.cmu.edu From: Smart is sexy. *cough*zephyr*cough* Auth: no Time: Mon May 17 00:24:47 2004 Host: TIR-NA-NOGTH.MIT.EDU From: Sam Hartman Hmm, dealing with the vendor branch after doing cvs2svn seems less successful than one might hope. svk smerge gives 1.198 conflicts svk cmerge skips any additions and crashes. svk merge leaves me without a ticket, but will at least give a baseline for how well things should be expceted to work Auth: no Time: Mon May 17 00:25:54 2004 Host: pallas.eruditorum.org From: They only want you when you're 17. When you're 21, you're no fun you know about svk smerge -C right? Auth: no Time: Mon May 17 00:26:06 2004 Host: TIR-NA-NOGTH.MIT.EDU From: Sam Hartman *ponders manual creation of a merge ticket* Auth: no Time: Mon May 17 00:26:39 2004 Host: TIR-NA-NOGTH.MIT.EDU From: Sam Hartman 79 conflicts. Worse than ideal, but better than 1100 Auth: no Time: Mon May 17 00:26:53 2004 Host: pallas.eruditorum.org From: They only want you when you're 17. When you're 21, you're no fun also, are you running svk -head? Auth: no Time: Mon May 17 00:28:01 2004 Host: TIR-NA-NOGTH.MIT.EDU From: Sam Hartman I'm running close to -head. it is close enough to head to include the fix for my bug. I know about svk smerge -C; that will tell me how many conflicts I have. Hmm, I guess it will also give me the merge ticket, won't it Auth: no Time: Mon May 17 21:44:24 2004 Host: pallas.eruditorum.org From: 22. Is there a trampoline on the premises? Oh wow. merges suck so much less when your version control system can just pop up your chosen merge tool Auth: yes Time: Mon May 17 21:44:47 2004 Host: hydra.hamachi.org From: Bill Sommerfeld heh. Auth: no Time: Mon May 17 21:44:50 2004 Host: STRATTON-SIX-NINETY-FIVE.MIT.EDU From: Is there justice in Perl? What is your chosen merge tool? Auth: no Time: Mon May 17 21:45:01 2004 Host: pallas.eruditorum.org From: 22. Is there a trampoline on the premises? FileMerge.app Auth: no Time: Mon May 17 21:45:16 2004 Host: pallas.eruditorum.org From: 22. Is there a trampoline on the premises? but svk supports anything you can call from a shel lscript ;) Auth: no Time: Mon May 17 21:45:20 2004 Host: STRATTON-SIX-NINETY-FIVE.MIT.EDU From: Imperialist pig. Give us some wheaties. Such a think really exists? Auth: yes Time: Mon May 17 21:45:40 2004 Host: MULTICS.MIT.EDU From: Jacob Morzinski *blink blink* Never heard of it. I was thinking of emacs's ediff/merge tool. Auth: yes Time: Mon May 17 21:45:41 2004 Host: hydra.hamachi.org From: Bill Sommerfeld (CVS is really the outlyer on that front. nearly every other VCS's i've ever used requires manual intervention for merges..) Auth: no Time: Mon May 17 21:45:44 2004 Host: STRATTON-SIX-NINETY-FIVE.MIT.EDU From: God doesn't love Eponine. God loves Cosette. Well sure. But I don't have any merge tool, unless you count "load up the file with <<<< and >>>> in emacs". Auth: no Time: Mon May 17 21:46:29 2004 Host: pallas.eruditorum.org From: 22. Is there a trampoline on the premises? well, this is only in case of conflict, but cross-branch meres create conflicts Auth: no Time: Mon May 17 21:47:08 2004 Host: pallas.eruditorum.org From: 22. Is there a trampoline on the premises? FileMerge.app is apple's gui merge tool. It's pretty and fast. and simple as hell to use. I couldn't even remember the keystrokes for ediff Auth: yes Time: Mon May 17 21:47:48 2004 Host: hydra.hamachi.org From: Bill Sommerfeld presumably you get passed files containing ancestor/A/B ? Auth: no Time: Mon May 17 21:48:24 2004 Host: TIR-NA-NOGTH.MIT.EDU From: Sam Hartman IS there a way to make svk interact with emerge? Or a related but different question have your filemerge.app changes been integrated yet? Auth: no Time: Mon May 17 21:48:25 2004 Host: pallas.eruditorum.org From: 22. Is there a trampoline on the premises? sam: yes. Auth: no Time: Mon May 17 21:49:17 2004 Host: TIR-NA-NOGTH.MIT.EDU From: Sam Hartman To both questions? Auth: no Time: Mon May 17 21:49:25 2004 Host: pallas.eruditorum.org From: 22. Is there a trampoline on the premises? I'm in the process of doccing this, but if you have the -HEAD of svk, set SVKMERGE to a shell script. Mine looks like this: #!/bin/sh # This helper script will call apple's "FileMerge" tool with the correct arguments for svk. # $1 Working label configure.ac (YOURS) # $2 Working file /tmp/svk-mergeP0xhK.tmp # $3 Ancestor label configure.ac (BASE) # $4 Ancestor file /tmp/svk-merge9adFA.tmp # $5 Foreign label configure.ac (THEIRS) # $6 Foreign file /tmp/svk-mergeUgPfz.tmp # $7 Target file svk-mergeFa4Rp echo "Left: $1 Right: $3 Ancestor: $5 Target file: $7" /Developer/Applications/Utilities/FileMerge.app/Contents/MacOS/FileMerge -left $2 -right $4 -ancestor $6 -merge $7 Auth: no Time: Mon May 17 21:49:45 2004 Host: pallas.eruditorum.org From: 22. Is there a trampoline on the premises? http://www.clkao.org/tmp/svk-merge-emacs may also prove usful Auth: no Time: Mon May 17 21:50:30 2004 Host: pallas.eruditorum.org From: 22. Is there a trampoline on the premises? Disclosure: this feature was my showstopper. So Best Practical paid for it to be developed. Auth: yes Time: Mon May 17 21:52:34 2004 Host: EGYPTIAN-GODS.MIT.EDU From: solitary, poor, nasty, brutish, and short The merge tool thing? Auth: no Time: Mon May 17 21:52:27 2004 Host: pallas.eruditorum.org From: 22. Is there a trampoline on the premises? yeah Auth: no Time: Mon May 17 21:53:18 2004 Host: TIR-NA-NOGTH.MIT.EDU From: Sam Hartman Ah, my show stopper is trying to avoid conflicts completely.;) Auth: yes Time: Mon May 17 21:53:45 2004 Host: hydra.hamachi.org From: Bill Sommerfeld mmm, unobtainium. Auth: no Time: Mon May 17 21:53:57 2004 Host: pallas.eruditorum.org From: 22. Is there a trampoline on the premises? I'm slightly pragmatic ;) Auth: no Time: Mon May 17 21:54:51 2004 Host: TIR-NA-NOGTH.MIT.EDU From: Sam Hartman OK. Well what I really mean is that there should actually be a conflict that the scm system could not have avoided before I get given a conflict. Auth: no Time: Mon May 17 21:55:33 2004 Host: pallas.eruditorum.org From: 22. Is there a trampoline on the premises? yeah. and I'm working through my big conflict scenario with clkao now. He's got a test-case and agrees that I've found a hard problem. Auth: no Time: Mon May 17 21:55:56 2004 Host: pallas.eruditorum.org From: 22. Is there a trampoline on the premises? right now, the smart-merges on a local branch that flow back up to the master don't have all the right context Auth: yes Time: Mon May 17 21:57:45 2004 Host: hydra.hamachi.org From: Bill Sommerfeld teamware has two levels of conflict. 1) same file touched on both branches requiring merge. 2) automatic merge failed and needs "help". there's human interaction for both of these (in 1 you just ratify that the merge tool DTRT). automated merge tools can differ in their ability to cope with changes to RCS/SCCS tags, insertions into the same place in the file, moved lines, etc., Auth: yes Time: Mon May 17 22:00:54 2004 Host: EGYPTIAN-GODS.MIT.EDU From: solitary, poor, nasty, brutish, and short I assume the problem in this place is that the merge sources are screwed up somehow; svk is accidentally trying to merge the same change twice, or something. Auth: yes Time: Mon May 17 22:00:59 2004 Host: EGYPTIAN-GODS.MIT.EDU From: solitary, poor, nasty, brutish, and short s/place/case/ Auth: no Time: Mon May 17 22:01:03 2004 Host: pallas.eruditorum.org From: 22. Is there a trampoline on the premises? greg: right Auth: no Time: Mon May 17 22:10:18 2004 Host: TIR-NA-NOGTH.MIT.EDU From: Sam Hartman Interesting, Jesse. I was not able to reproduce that with my tests. Auth: no Time: Mon May 17 22:10:46 2004 Host: pallas.eruditorum.org From: 22. Is there a trampoline on the premises? this is tests with two different hosts Auth: no Time: Mon May 17 22:11:04 2004 Host: pallas.eruditorum.org From: 22. Is there a trampoline on the premises? doing smerges on local branches on the mac, I then lose when i do the same semerge on the linux box Auth: no Time: Mon May 17 22:11:32 2004 Host: TIR-NA-NOGTH.MIT.EDU From: Sam Hartman Hmm, and svk seems to have messed up a copy fairly significantly. I wish I knew how/why and could reproduce it. Auth: yes Time: Mon May 17 22:12:36 2004 Host: hydra.hamachi.org From: Bill Sommerfeld svk uses the bk/teamware model of having a complete local repository? Auth: no Time: Mon May 17 22:12:42 2004 Host: TIR-NA-NOGTH.MIT.EDU From: Sam Hartman I didn't try with two hosts but did try with something that should have reasonably closely emulated that. Auth: no Time: Mon May 17 22:13:05 2004 Host: TIR-NA-NOGTH.MIT.EDU From: Sam Hartman Bill, sort of. Approximately yes. Auth: no Time: Mon May 17 22:13:47 2004 Host: TIR-NA-NOGTH.MIT.EDU From: Sam Hartman In particular you can only mirror part of a repository (restricted both in space and time) but you can then only work with that. Auth: yes Time: Mon May 17 22:16:06 2004 Host: hydra.hamachi.org From: Bill Sommerfeld well, that sounds exactly like teamware -- for solaris, i can bringover all of usr/src or just usr/src/cmd/ls , but what bringover constructs has the complete history of what was brought over.. Auth: no Time: Mon May 17 22:16:57 2004 Host: TIR-NA-NOGTH.MIT.EDU From: Sam Hartman svk also lets you grab an incomplete history if you so choose Auth: no Time: Mon May 17 22:24:24 2004 Host: TIR-NA-NOGTH.MIT.EDU From: Sam Hartman There are some cases where svk makes really strange decisions about what to skip in a merge and I think it is broken. Auth: no Time: Mon May 17 22:24:26 2004 Host: pallas.eruditorum.org From: 22. Is there a trampoline on the premises? Agreed Auth: no Time: Mon May 17 22:24:48 2004 Host: pallas.eruditorum.org From: 22. Is there a trampoline on the premises? clkao claims it's using an svn library. Auth: yes Time: Mon May 17 22:25:51 2004 Host: EGYPTIAN-GODS.MIT.EDU From: solitary, poor, nasty, brutish, and short What do yo umean by "decisions about what to skip?" Auth: no Time: Mon May 17 22:26:19 2004 Host: TIR-NA-NOGTH.MIT.EDU From: Sam Hartman Uh. I don't think that's it though. I think it is the merge editor. I should go understand that code at some point. Auth: yes Time: Mon May 17 22:27:43 2004 Host: EGYPTIAN-GODS.MIT.EDU From: solitary, poor, nasty, brutish, and short There is an svn library which implements the equivalent of diff3. Presumably svk is using it for some purpose. Since I don't understand what Sam said, I don't know if it could be at fault. Auth: no Time: Mon May 17 22:30:44 2004 Host: TIR-NA-NOGTH.MIT.EDU From: Sam Hartman Some times when doing a merge svk will decide to skip directories it is especially likely to do this when it is trying to add a directory. But I just saw it skip a merge ofr src/lib/crypto/aes from one copy of krb5 to another for no apparent reason Auth: yes Time: Tue May 18 11:40:00 2004 Host: TIR-NA-NOGTH.MIT.EDU From: Sam Hartman Hmm, I have a case where svk seems to barf with a transaction out of date error syncing across a revision with lots of copies. Auth: yes Time: Tue May 18 18:18:17 2004 Host: TIR-NA-NOGTH.MIT.EDU From: Sam Hartman My laptop doesn't seem to have enough memory to use svk. This is unfortunate. Auth: no Time: Tue May 18 18:18:33 2004 Host: pallas.eruditorum.org From: 22. Is there a trampoline on the premises? really? Auth: yes Time: Tue May 18 18:20:17 2004 Host: TIR-NA-NOGTH.MIT.EDU From: Sam Hartman I have 384M. The perl for svk syncs tends to get really really huge. svm in general seems to have lots of inefficiencies in time, space (disk and memory) and network bandwith especially when dealing with copy-heavy repositories Auth: yes Time: Tue May 18 18:21:04 2004 Host: TIR-NA-NOGTH.MIT.EDU From: Sam Hartman I've been trying to mirror the repository for my Debian krb5 package all day and have gotten through most of it. Auth: no Time: Tue May 18 18:21:07 2004 Host: pallas.eruditorum.org From: 22. Is there a trampoline on the premises? there's probably a leak Auth: no Time: Tue May 18 18:21:14 2004 Host: pallas.eruditorum.org From: 22. Is there a trampoline on the premises? what version of svn? Auth: yes Time: Tue May 18 18:21:52 2004 Host: TIR-NA-NOGTH.MIT.EDU From: Sam Hartman It's relatively copy heavy (tag for every upload and for every upstream krb5 release) Auth: no Time: Tue May 18 18:22:07 2004 Host: pallas.eruditorum.org From: 22. Is there a trampoline on the premises? isn't that supposed to be cheap in svn/svk? Auth: yes Time: Tue May 18 18:22:25 2004 Host: TIR-NA-NOGTH.MIT.EDU From: Sam Hartman 0.14, but it seems like the version of SVN::Mirror is going to matter much more for this. Auth: yes Time: Tue May 18 18:23:02 2004 Host: TIR-NA-NOGTH.MIT.EDU From: Sam Hartman Jesse, seems to be cheap in svk/svn but very much not svm at least in terms of network bandwith Auth: yes Time: Tue May 18 18:24:03 2004 Host: TIR-NA-NOGTH.MIT.EDU From: Sam Hartman libsvn-mirror-perl 0.35-1 from Debian Auth: no Time: Tue May 18 18:23:58 2004 Host: pallas.eruditorum.org From: 22. Is there a trampoline on the premises? what version of svn Auth: yes Time: Tue May 18 18:24:50 2004 Host: TIR-NA-NOGTH.MIT.EDU From: Sam Hartman I don't know enough about svn_ra to kno whether this has to be expensive or not. Auth: yes Time: Tue May 18 18:25:40 2004 Host: TIR-NA-NOGTH.MIT.EDU From: Sam Hartman Presumably 1.02 Yeah, libsvn0 1.02-5 Auth: no Time: Tue May 18 18:25:31 2004 Host: pallas.eruditorum.org From: 22. Is there a trampoline on the premises? *nod* Auth: no Time: Tue May 18 18:25:49 2004 Host: pallas.eruditorum.org From: 22. Is there a trampoline on the premises? is the krb5 repository publicly accessible via anoncvs? Auth: yes Time: Tue May 18 18:27:03 2004 Host: TIR-NA-NOGTH.MIT.EDU From: Sam Hartman Yes, but this is the krb5 debian repository, which is svn not cvs. The krb5 debian repository is not currently public although I could make it accessible by svn or http. Auth: no Time: Tue May 18 18:28:17 2004 Host: pallas.eruditorum.org From: 22. Is there a trampoline on the premises? if you can set up a demo case, clkao is most likely to cope. Auth: yes Time: Tue May 18 18:29:43 2004 Host: TIR-NA-NOGTH.MIT.EDU From: Sam Hartman This makes sense in retrospect, but it turns out you have a significant space penalty if you skip the initial revisions of the repository. IN particular if you skip to a revision where a bunch of copies already exist and don't include the revision that creates the copy, it seems to actually duplicate space for everything. Not surprising but goes against my initial intuition that skipping revisions would save space. Auth: yes Time: Tue May 18 18:30:42 2004 Host: TIR-NA-NOGTH.MIT.EDU From: Sam Hartman Well, let me first ask if this is expected behavior; it may just be that you have to read all the data for each copy. Auth: no Time: Tue May 18 18:30:32 2004 Host: pallas.eruditorum.org From: 22. Is there a trampoline on the premises? I think that skipping revs makes the most sense if you only want to get the head, but not tags. Auth: no Time: Tue May 18 18:30:42 2004 Host: pallas.eruditorum.org From: 22. Is there a trampoline on the premises? But I don't know anything for sure Auth: yes Time: Tue May 18 18:32:51 2004 Host: TIR-NA-NOGTH.MIT.EDU From: Sam Hartman That seems to be my experience. Auth: yes Time: Wed May 19 14:47:34 2004 Host: TIR-NA-NOGTH.MIT.EDU From: Sam Hartman So, when I mentioned that svk wanted more memory than I had in my laptop, I didn't realize quite how memory starved it was being. Adding 400M of swap much improves performance. Auth: yes Time: Wed May 19 14:55:50 2004 Host: TIR-NA-NOGTH.MIT.EDU From: Sam Hartman I think I want to try head. It's like some pool is not getting released somewhere. (I seem to recall recent pool handling changes) It could also be a memory leak, but if so memory usage does not seem to depend too much on how many revisions have been committed. Auth: yes Time: Sat Jun 5 16:26:35 2004 Host: TIR-NA-NOGTH.MIT.EDU From: Sam Hartman Are the temporary directories created by svk cmerge intended to stick around? Auth: no Time: Sat Jun 5 16:29:45 2004 Host: pallas.eruditorum.org From: They only want you when you're 17. When you're 21, you're no fun imo, no, but clkao may feel differently Auth: yes Time: Sat Jun 5 16:31:57 2004 Host: TIR-NA-NOGTH.MIT.EDU From: Sam Hartman OK, but the current code does not try to remove them? Auth: no Time: Sat Jun 5 16:31:58 2004 Host: pallas.eruditorum.org From: They only want you when you're 17. When you're 21, you're no fun Don't know off the top of my head. it probably uses File::Temp and doesn't pass CLEANUP => 1 Auth: yes Time: Sat Jun 5 16:35:37 2004 Host: TIR-NA-NOGTH.MIT.EDU From: Sam Hartman No, I'm talking about svn_fs dirs created in the repository Auth: yes Time: Sat Jun 5 16:39:27 2004 Host: TIR-NA-NOGTH.MIT.EDU From: Sam Hartman I also just got very screwed by svk delete not respecting mirrored paths. Auth: no Time: Sat Jun 5 16:41:31 2004 Host: pallas.eruditorum.org From: They only want you when you're 17. When you're 21, you're no fun what svk are you running? Auth: yes Time: Sat Jun 5 16:42:44 2004 Host: TIR-NA-NOGTH.MIT.EDU From: Sam Hartman I think 0.14. Auth: yes Time: Sat Jun 5 16:42:59 2004 Host: TIR-NA-NOGTH.MIT.EDU From: Sam Hartman On that machine; something newer on another. Auth: no Time: Sat Jun 5 16:42:46 2004 Host: pallas.eruditorum.org From: They only want you when you're 17. When you're 21, you're no fun I know clkao has put energy into making that sort of thing safer lately. Auth: yes Time: Sat Jun 5 16:46:10 2004 Host: TIR-NA-NOGTH.MIT.EDU From: Sam Hartman I'm reloading a dump to avoid the unfortunate revision. Auth: no Time: Sat Jun 5 16:47:20 2004 Host: pallas.eruditorum.org From: They only want you when you're 17. When you're 21, you're no fun I,I "svn undo" Auth: yes Time: Wed Jun 9 21:49:44 2004 Host: TIR-NA-NOGTH.MIT.EDU From: Sam Hartman Weeks later I'm still mostly happy. ALthough I'm sad that Debian's libsvn0 is too old for bleeding-edge svk. Auth: no Time: Wed Jun 9 21:49:36 2004 Host: pallas.eruditorum.org From: They only want you when you're 17. When you're 21, you're no fun the only thing you'll lose on is the patchset stuff Auth: yes Time: Wed Jun 9 21:51:56 2004 Host: TIR-NA-NOGTH.MIT.EDU From: Sam Hartman No, I think I want some of the changes to how merge bases are found, I want the support for external merge programs, etc. Auth: no Time: Tue Jul 6 19:50:53 2004 Host: TIR-NA-NOGTH.MIT.EDU From: Sam Hartman Hmm, first time I've been unable to commit something. Auth: yes Time: Mon Jul 19 17:23:47 2004 Host: TIR-NA-NOGTH.MIT.EDU From: Sam Hartman Is the svk patch api actually useful yet? In particular, can I actually do anything with a patch? Auth: yes Time: Thu Jul 22 14:29:46 2004 Host: TIR-NA-NOGTH.MIT.EDU From: Sam Hartman What's been up with the svk repository server lately? Or perhaps what's been down would be a better question. Auth: yes Time: Tue Jul 27 16:20:08 2004 Host: TIR-NA-NOGTH.MIT.EDU From: Sam Hartman Hmm, svk sync is failing with a transaction out of date error Auth: yes Time: Tue Jul 27 16:22:08 2004 Host: TIR-NA-NOGTH.MIT.EDU From: Sam Hartman ANd people who use autoload in perl should almost all be shot. Auth: yes Time: Thu Jul 29 18:25:30 2004 Host: TIR-NA-NOGTH.MIT.EDU From: Sam Hartman Grrh. All my simplifying assumptions fail. Auth: yes Time: Fri Jul 30 18:06:01 2004 Host: TIR-NA-NOGTH.MIT.EDU From: Sam Hartman clkao++ Reading longish mail to skv-dev Auth: yes Time: Thu Aug 12 14:29:36 2004 Host: TIR-NA-NOGTH.MIT.EDU From: Sam Hartman Does anyone here use svk with svn 1.0.x? Auth: yes Time: Tue Oct 19 12:37:51 2004 Host: TIR-NA-NOGTH.MIT.EDU From: Sam Hartman SVK 0.23 seems not quite ready for prime time based on list traffic Auth: yes Time: Fri Oct 22 14:58:04 2004 Host: TIR-NA-NOGTH.MIT.EDU From: Sam Hartman Suck. svk seems to have done some amazingly bad jobs of finding merge bases. I wish I had time to debug Auth: yes Time: Fri Oct 22 15:02:41 2004 Host: TIR-NA-NOGTH.MIT.EDU From: Sam Hartman OK, and it works correctly if I do the merge that way. *ponder* Auth: yes Time: Thu Jan 13 20:27:20 2005 Host: TIR-NA-NOGTH.MIT.EDU From: Sam Hartman Hmm, there seem to be a large number of revisions that do not end up creating svk revisions. This is true even though I'm mirroring the entire repository Auth: yes Time: Wed Feb 23 13:12:07 2005 Host: TIR-NA-NOGTH.MIT.EDU From: Sam Hartman While trying to do an import I get: Mixed newlines found in "/tmp/openafs/openafs-1.3.79/src/WINNT/doc/install/Documentation/de_DE/README .txt" at /usr/share/perl5/SVK/Util.pm line 380. That's all well in good; I'm perfectly happy to believe my sources have bogus new lines. But that's not a reason to fail the import. Or at least anything that makes me change the sources before import seems wrong. Suggestions on how to convince svk not to care? Auth: no Time: Wed Feb 23 13:17:02 2005 Host: diesel.bestpractical.com From: Now part of the problem Dunno. #svk? Auth: yes Time: Wed Feb 23 13:14:52 2005 Host: TIR-NA-NOGTH.MIT.EDU From: Sam Hartman That was in fact where I was headed next;) Auth: yes Time: Mon May 2 03:25:08 2005 Host: TIR-NA-NOGTH.MIT.EDU From: Sam Hartman Ick. svk's ability to do smart copies when doing mergebacks is not what you'd hope it is. Auth: yes Time: Wed May 11 00:30:33 2005 Host: whooper.wustl.edu From: Comfort the disturbed and disturb the comfortable. Cool, svk just hit 1.0 Auth: no Time: Sun Jul 10 18:53:02 2005 Host: TIR-NA-NOGTH.MIT.EDU From: Sam Hartman It's annoying that if you copy something on a local branch and then push it back you don't actually get a copy. Auth: yes Time: Mon Dec 5 00:19:43 2005 Host: fooserver.marchorowitz.com From: Marc Horowitz is there an svk tutorial which doesn't suck? Auth: no Time: Mon Dec 5 00:20:05 2005 Host: CLIODEV.IHTFP.ORG From: The Warlord is there an svk installation method that doesn't suck? Auth: yes Time: Mon Dec 5 00:20:18 2005 Host: fooserver.marchorowitz.com From: Marc Horowitz "apt-get install svk" Auth: yes Time: Mon Dec 5 00:21:58 2005 Host: fooserver.marchorowitz.com From: Marc Horowitz I mirrored a remote depot. I synced, and made some changes. Now, I want to commit back to the depot. what do I do? Auth: no Time: Mon Dec 5 00:22:18 2005 Host: CLIODEV.IHTFP.ORG From: The Warlord svk push? Auth: yes Time: Mon Dec 5 00:25:38 2005 Host: fooserver.marchorowitz.com From: Marc Horowitz the docs say that's the smae as svk smerge -If . I don't want -I. svk smerge -f . says > svk smerge -f . Cannot find the path which '.' copied from. Exit 1 Auth: no Time: Mon Dec 5 00:25:40 2005 Host: CYBERTRON.MIT.EDU From: "The dreams in which I'm dying are the best I've ever had" -- Mad World yes, svk push Auth: no Time: Mon Dec 5 00:26:02 2005 Host: CYBERTRON.MIT.EDU From: roomqs are the alpha, lambda, and omega what did you do to mirror the remote depot/check out? Auth: yes Time: Mon Dec 5 00:28:13 2005 Host: fooserver.marchorowitz.com From: Marc Horowitz svk mirror svn+ssh://svn.blahblah.org/repos/myproject //myproject svk sync //myproject svk co //myproject myproject Auth: no Time: Mon Dec 5 00:29:14 2005 Host: CYBERTRON.MIT.EDU From: roomqs are the alpha, lambda, and omega so you modify myproject/foo and do a svk push and you get the previous error? Auth: yes Time: Mon Dec 5 00:29:27 2005 Host: fooserver.marchorowitz.com From: Marc Horowitz well, I did svk smerge -f . Auth: no Time: Mon Dec 5 00:30:32 2005 Host: CYBERTRON.MIT.EDU From: "Getting it on is like programming in Linux" well, you could also use svk push -l Auth: yes Time: Mon Dec 5 00:30:52 2005 Host: fooserver.marchorowitz.com From: Marc Horowitz > svk push -l Cannot find the path which '' copied from. Auth: yes Time: Mon Dec 5 00:31:09 2005 Host: fooserver.marchorowitz.com From: Marc Horowitz I did that from myproject/trunk Auth: no Time: Mon Dec 5 00:31:16 2005 Host: CYBERTRON.MIT.EDU From: light sabers, the strap on? what does svk co --list show? Auth: yes Time: Mon Dec 5 00:32:12 2005 Host: fooserver.marchorowitz.com From: Marc Horowitz Depot Path Path ======================================================================== //myproject /home/marc/project/myproject I was in /home/marc/project/myproject when I did the svk co Auth: yes Time: Mon Dec 5 00:32:23 2005 Host: fooserver.marchorowitz.com From: Marc Horowitz oops. I was in /home/marc/project Auth: no Time: Mon Dec 5 00:36:38 2005 Host: CYBERTRON.MIT.EDU From: "Highly special with a capital 'K'" hrm, I'm not sure. I've never gotten that before Auth: yes Time: Mon Dec 5 00:38:18 2005 Host: fooserver.marchorowitz.com From: Marc Horowitz oh, I think I see. the path is /home/marc, but /home is a symlink, so pwd isn't giving back what svk expects. where does svk co --list come from, so I can fix it? Auth: yes Time: Mon Dec 5 00:39:18 2005 Host: fooserver.marchorowitz.com From: Marc Horowitz oh, no. that's not it. I misread it. Auth: yes Time: Mon Dec 5 00:42:47 2005 Host: fooserver.marchorowitz.com From: Marc Horowitz svk commit --import seems to be doing something. How does that differ from svk push ? Auth: yes Time: Mon Dec 5 00:48:32 2005 Host: fooserver.marchorowitz.com From: Marc Horowitz ok, svk commit --import did *everything*. I wish there was a document somewhere which explained why, or what I do if I don't want the change synced upstream, like if I'm disconnected. Auth: yes Time: Mon Dec 5 01:33:55 2005 Host: diesel.bestpractical.com From: Back, It's What Baby Got marc: so, "push" probably doesn't mean what you think it means Auth: yes Time: Mon Dec 5 01:34:38 2005 Host: diesel.bestpractical.com From: Back, It's What Baby Got the concept for mirrored paths in SVK is that you really really want them to be as accurate a reflection of whatever they're mirroring as possible. so committing to them should always try to post the change back to the original server. Auth: yes Time: Mon Dec 5 01:35:41 2005 Host: diesel.bestpractical.com From: Back, It's What Baby Got if you'd like to do disconnected work, just make a local branch (a copy) with something like 'svk cp //myproject //local/myproject' then (after committing your changes in //local/myproject with svk ci), svk push //local/myproject well push the changes in the local branch back to //myproject, the mirrored repository Auth: no Time: Mon Dec 5 01:41:14 2005 Host: CYBERTRON.MIT.EDU From: "honey-nut md5!" oh, right. I forgot he wasn't making a local branch. marc: you want svk ci Auth: yes Time: Mon Dec 5 01:45:32 2005 Host: diesel.bestpractical.com From: Back, It's What Baby Got "push and pull are for branches" "commit (ci) and update (up) are for working copies" more or less. Auth: yes Time: Tue Jan 24 17:44:38 2006 Host: TIR-NA-NOGTH.MIT.EDU From: Sam Hartman Is there a way for me to set the contents of a property from a file? Auth: yes Time: Tue Jan 24 17:51:57 2006 Host: diesel.bestpractical.com From: Back, It's What Baby Got with svk ps? Auth: yes Time: Tue Jan 24 17:52:19 2006 Host: diesel.bestpractical.com From: Back, It's What Baby Got (and your shell) Auth: yes Time: Tue Jan 24 17:54:10 2006 Host: TIR-NA-NOGTH.MIT.EDU From: Sam Hartman How would I do that with svk ps Auth: yes Time: Tue Jan 24 18:02:41 2006 Host: diesel.bestpractical.com From: Back, It's What Baby Got What shell do you use? Auth: yes Time: Tue Jan 24 18:03:11 2006 Host: TIR-NA-NOGTH.MIT.EDU From: Sam Hartman bash Auth: yes Time: Tue Jan 24 18:04:28 2006 Host: diesel.bestpractical.com From: Back, It's What Baby Got I'd suspect that svk ps propname <(cat valueIsHere) file would work Auth: yes Time: Fri Jun 9 20:16:29 2006 Host: CLIODEV.IHTFP.ORG From: The Warlord Hmm, the first time you 'svk sync' it can take a long time. Auth: yes Time: Fri Jun 9 20:19:03 2006 Host: CLIODEV.IHTFP.ORG From: The Warlord yeah.. there are only 14k changesets in the repo.. But of course each changeset is getting renumbered in the svk repo. Auth: yes Time: Fri Jun 9 20:23:22 2006 Host: CLIODEV.IHTFP.ORG From: The Warlord Is there some easy way to map the svk revision# to the upstream revision#? Auth: no Time: Mon Jun 12 18:03:51 2006 Host: VINEGAR-POT.MIT.EDU From: America in the Age of Wizardry A big downside of not having .svn directories is that you can't just mv a working copy. Auth: yes Time: Mon Jun 12 18:04:20 2006 Host: diesel.bestpractical.com From: Back, It's What Baby Got Yeah, you need to do svk co --relocate Auth: yes Time: Mon Jun 12 18:04:32 2006 Host: CLIODEV.IHTFP.ORG From: The Warlord it also makes it hard to detect if you're in an SVN/SVK working directory. Auth: yes Time: Mon Jun 12 18:04:41 2006 Host: diesel.bestpractical.com From: Back, It's What Baby Got i'm not convinced that's a big problem in practice. sharing a checkout is. Auth: no Time: Mon Jun 12 18:07:24 2006 Host: VINEGAR-POT.MIT.EDU From: "Always think twice before doing what a @i[man] asks you to do..." -- Aimee L. Smith svk co --relocate seems to be sufficient for normal use, though I don't know how I could have guessed it was a checkout subcommand. Auth: yes Time: Mon Jun 12 18:07:56 2006 Host: SUPOX.MIT.EDU From: moose. I think being able to grep through local copies without having to | grep -v .svn is worth it Auth: no Time: Mon Jun 12 18:08:24 2006 Host: VINEGAR-POT.MIT.EDU From: zephyr hermit I never have that problem, since I use grep foo **/* instead of grep -r foo . . Auth: yes Time: Mon Jun 12 18:08:46 2006 Host: HOROBI.MIT.EDU From: Jacob Morzinski I assume you have an insane shell, if ** has meaning? Auth: no Time: Mon Jun 12 18:09:03 2006 Host: VINEGAR-POT.MIT.EDU From: endocrine automaton I use zsh, which is hardly cutting edge technology. Auth: yes Time: Mon Jun 12 18:09:09 2006 Host: CLIODEV.IHTFP.ORG From: The Warlord Hmm. I might need to hack the gnucash sources.. they assume that 'svnversion' will get the svn revision#.. And it expects to find a .svn directory Auth: yes Time: Thu Jun 22 19:52:14 2006 Host: CLIODEV.IHTFP.ORG From: The Warlord Hmm, Is there an SVK equivalent to 'svnversion'? Auth: yes Time: Thu Jul 13 10:47:17 2006 Host: h067e-net84db.lab.risq.net From: The Warlord Is there some way to get svk to resync unversioned properties when they are updated in a mirrored repository? Auth: no Time: Thu Jul 13 10:50:02 2006 Host: diesel.bestpractical.com From: Back, It's What Baby Got That's a good question Auth: yes Time: Thu Jul 13 10:51:05 2006 Host: h067e-net84db.lab.risq.net From: The Warlord in particular, I know four particular revisions where svn:log was updated in the svn repository. I'd like to pull down those changes into my svk depot... Assume I even know which revisions were changed. Auth: yes Time: Wed Oct 25 15:55:13 2006 Host: 204.107.200.20 From: The Warlord Hmm, how to I delete a commit from my repository? Somehow (probably during a pull?) a changeset got committed without a log message and I need to remove that changeset from my repository. Is there an svk admin (or svnadmin) function to remove it? Auth: yes Time: Wed Oct 25 15:56:15 2006 Host: HOROBI.MIT.EDU From: Jacob Morzinski Last I heard, there is no svn admin. There is a svn dump, and then you can run scripts over your dump to excise parts of the database, then you can undump back into svn. Auth: yes Time: Wed Oct 25 15:57:52 2006 Host: 204.107.200.20 From: The Warlord svn dump will take hours. I can't believe there's no way to just remove a commit Auth: no Time: Wed Oct 25 16:02:05 2006 Host: VINEGAR-POT.MIT.EDU From: useful There is a way to change the log message. There's no way to remove a previous svn commit. Auth: no Time: Tue Nov 21 15:57:27 2006 Host: luminous.suchdamage.org From: Sam Hartman svk 1.8's behavior dealing with merges with directory-level conflicts is not optimal. Auth: no Time: Fri Feb 2 15:49:47 2007 Host: VINEGAR-POT.MIT.EDU From: The first step towards knowledge is to know that we are ignorant. Enroll at MIT today. How does svk interact with svn repository hook scripts? Auth: yes Time: Fri Feb 23 02:57:40 2007 Host: QUICHE-LORRAINE.MIT.EDU From: Ken T Takusagawa Damn. svk is not on athena or a locker. Auth: yes Time: Fri Feb 23 08:50:41 2007 Host: LUNATIQUE.MIT.EDU From: To get a vector perpendicular to itself and the road! *stark* Poke glasser, I know he was trying to install it at one point, and had it mostly working. Auth: yes Time: Fri Feb 23 08:54:34 2007 Host: diesel.bestpractical.com From: Word cannot open the Microsoft. yo. Auth: yes Time: Fri Feb 23 08:54:46 2007 Host: diesel.bestpractical.com From: Word cannot open the Microsoft. It's in ~glasser :) Auth: yes Time: Fri Feb 23 08:59:19 2007 Host: diesel.bestpractical.com From: Word cannot open the Microsoft. I should put it in a locker at some point, but I don't really use Athena heavily enough these days Auth: no Time: Thu May 31 15:06:36 2007 Host: VINEGAR-POT.MIT.EDU From: Professor Ashfield doesn't wear shirts made out of hemp. He wears shirts made out of LSD. OK, I've read the svk intro and some command help texts. I still am not quite understanding how I'm supposed to push and pull changes from a branch that lives on the svn server to/from the trunk that lives on the svn server. I don't really want a local branch right now, since local state is unreliable. Auth: no Time: Thu May 31 16:37:19 2007 Host: VINEGAR-POT.MIT.EDU From: Crisis = Danger + Machinery svk makes 2-line changes to files in branches into a gigantic project. I hope that it helps reduce the gigantic project that merges currently are. Auth: no Time: Thu May 31 22:48:13 2007 Host: VINEGAR-POT.MIT.EDU From: In 1978, for the first time it was thought necessary to construct a special board beneath the chess table to prevent the players from kicking one another. Can svk pull multiple levels of branches down, or only one level? I have a //mirror/proj/trunk -> //mirror/proj/branch -> //local/branch -> working copy and svk pull pulls from //mirror/proj/branch but not //mirror/proj/trunk. Auth: no Time: Thu May 31 22:50:10 2007 Host: VINEGAR-POT.MIT.EDU From: little fantasy creature Ah, I figured it out, I need to do svk pull //mirror/proj/branch . Auth: no Time: Fri Jul 13 13:15:00 2007 Host: VINEGAR-POT.MIT.EDU From: Don't stop now! You can @i[win] this flamewar. It seems that $Revision$ keywords get my local mirror's revision number. Is it plausible to get some keyword that reflects the main repository's revision number? Auth: yes Time: Fri Jul 13 13:23:45 2007 Host: VINEGAR-POT.MIT.EDU From: No, I don't know when. [thesis draft due in 7 days, 3 hours, 36 minutes] There is one. Auth: yes Time: Fri Jul 13 13:24:31 2007 Host: VINEGAR-POT.MIT.EDU From: No, I shouldn't be watching TV. [thesis draft due in 7 days, 3 hours, 35 minutes] Actually, I lied, there isn't. Patches welcome. SVK::XD::get_keyword_layer Auth: no Time: Fri Jul 13 13:34:41 2007 Host: VINEGAR-POT.MIT.EDU From: where everything's mine if I can keep my mouth shut tight I mean, it doesn't actually matter that much since I could always check out from the svn trunk if I cared. But I'd have to remember to do that. Auth: yes Time: Fri Jul 13 13:35:17 2007 Host: VINEGAR-POT.MIT.EDU From: Yes, I'm moving to California. [thesis draft due in 7 days, 3 hours, 24 minutes] "svk info" will also tell you Auth: no Time: Mon Jul 30 22:33:03 2007 Host: VINEGAR-POT.MIT.EDU From: I'd probably get just as many presents if I'd set the Pope on fire in the past year. I made a mess by 'svk push'ing some changes from a branch to the trunk before I wanted them pushed. I can't figure out how to revert this and unscramble the omelette, though. Auth: yes Time: Mon Jul 30 22:54:08 2007 Host: BRUTE-FORCE.MIT.EDU From: @@mantra{must only use powers for good} Not sure if it's correct, but couldn't use svn to reverse the merge, then pull that change back to svk? Auth: no Time: Mon Jul 30 22:55:28 2007 Host: luminous.suchdamage.org From: Sam Hartman Did you push as one change or as many? Auth: no Time: Mon Jul 30 22:56:37 2007 Host: luminous.suchdamage.org From: Sam Hartman Basically let's say you pushed 100-105 and that no one else committed during your commit. you can svk merge -r 105:100 -m "Back out 100-105" /path_to_trunk You then svk propedit the trunk's svk:merge to indicate the merge never happened Auth: no Time: Mon Jul 30 22:57:47 2007 Host: luminous.suchdamage.org From: Sam Hartman Or just delete the trunk and svk cp it from before the merge back into existence Auth: no Time: Mon Jul 30 23:18:52 2007 Host: VINEGAR-POT.MIT.EDU From: Oscar Mayer's Great New "Sack O'Sauce in a Can O'Meat"! Sorry for not responding --- I just noticed I wasn't subbed to -c sipb before. I figured out the svk reverse merge thing myself, but didn't bother changing merge tickets. It seems to hve recovered OK anyway, though I'm sure I've messed up svn blame and logs. Auth: no Time: Mon Aug 27 19:20:33 2007 Host: VINEGAR-POT.MIT.EDU From: In Denmark, a company is now building the world's first power plant that generates electricity by burning cattle. Is it 'normal' to accumulate a list of 'copied from' 6 items long, or should I be pruning? Auth: no Time: Wed Aug 29 18:01:08 2007 Host: VINEGAR-POT.MIT.EDU From: "Bituminous coal." These were his last words on earth. Is it possible to mark a branch as pull-only? I don't want to accidentally push changes to the trunk. Auth: yes Time: Thu Sep 6 00:04:57 2007 Host: pulse-plasma-gun.nerd-militia.org From: Marc Horowitz is there an svk mode for emacs? Auth: no Time: Thu Sep 6 00:05:27 2007 Host: linode.innocuous.org From: Boston. Orlando 9/17-18?? Vienna 9/23-27. London 9/28-29. Montreal 10/21-25? Not that I was able to find. Auth: yes Time: Sun Oct 14 13:34:36 2007 Host: central-air-conditioning.toybox.cambridge.ma.us From: Marc Horowitz > svk commit -m "don't quote pathnames twice" Commit into mirrored path: merging back directly. Merging back to mirror source svn+ssh://svn.nerd-militia.org/home/marc/.svnrepos/gnus/trunk. Merge back committed as revision 24. Syncing svn+ssh://svn.nerd-militia.org/home/marc/.svnrepos/gnus/trunk Retrieving log information from 16 to 24 Committed revision 41 from revision 24. Killed by signal 15. Killed by signal 15. what's getting SIGTERMed?. Auth: yes Time: Sun Oct 14 13:34:44 2007 Host: LUNATIQUE.MIT.EDU From: ssh Auth: yes Time: Sun Oct 14 13:35:04 2007 Host: central-air-conditioning.toybox.cambridge.ma.us From: Marc Horowitz svk is killing it, and ssh just recently began complaining about it? Auth: yes Time: Sun Oct 14 13:35:13 2007 Host: LUNATIQUE.MIT.EDU From: svn is killing it Auth: yes Time: Sun Oct 14 13:35:19 2007 Host: central-air-conditioning.toybox.cambridge.ma.us From: Marc Horowitz so it's harmless? Auth: yes Time: Sun Oct 14 13:35:21 2007 Host: LUNATIQUE.MIT.EDU From: This is svn's behavior, not svk's. For some weird reason, it kill -TERM's its ssh tunnel when it's done with it. Auth: yes Time: Sun Oct 14 13:35:23 2007 Host: LUNATIQUE.MIT.EDU From: correct Auth: yes Time: Sun Oct 14 13:35:45 2007 Host: central-air-conditioning.toybox.cambridge.ma.us From: Marc Horowitz can I get ssh to shut up? Auth: yes Time: Sun Oct 14 13:36:01 2007 Host: central-air-conditioning.toybox.cambridge.ma.us From: Marc Horowitz I see I'm not the first person to say wtf here :-) Auth: yes Time: Sun Oct 14 13:36:21 2007 Host: curl-up-paw-under-nose.thok.org From: The reward for a job well done... is another, harder job... it's not harmless. Auth: yes Time: Sun Oct 14 13:36:54 2007 Host: central-air-conditioning.toybox.cambridge.ma.us From: Marc Horowitz upgrading to a new machine is such a pain. I want to give a wedgie to everyone who has allowed software to regress. Auth: yes Time: Sun Oct 14 13:37:07 2007 Host: curl-up-paw-under-nose.thok.org From: Cry and you cry alone. Smile, and all the world will wonder what what you're up to. Note that the change is that ssh actually *prints* that it's dying; it always did.... *without restoring the non-blocking setting of the tty* Auth: yes Time: Sun Oct 14 13:37:22 2007 Host: central-air-conditioning.toybox.cambridge.ma.us From: Marc Horowitz the shell should take care of that. Auth: yes Time: Sun Oct 14 13:37:32 2007 Host: curl-up-paw-under-nose.thok.org From: You know I've always liked that word, GARGANTUAN. I so rarely have an opportunity to use it in a sentence. you'd like to think that. Auth: yes Time: Sun Oct 14 13:37:45 2007 Host: central-air-conditioning.toybox.cambridge.ma.us From: Marc Horowitz tcsh is excessively paranoid about this. I don't know about bash. Auth: yes Time: Sun Oct 14 13:37:55 2007 Host: central-air-conditioning.toybox.cambridge.ma.us From: Marc Horowitz if my shell was in the wrong mode, how would I notice? Auth: yes Time: Sun Oct 14 13:38:36 2007 Host: curl-up-paw-under-nose.thok.org From: dal mouimento annuo della Terra wierdshit, like tar xvf crashing on ewouldblock Auth: yes Time: Sun Oct 14 13:39:05 2007 Host: VINEGAR-POT.MIT.EDU From: not worth transcribing Yeah, we are pretty sure that this is an ssh bug. There is a patch somewhere. "Last time I asked the patch was in one of eichin's open tabs" Auth: yes Time: Sun Oct 14 13:39:07 2007 Host: central-air-conditioning.toybox.cambridge.ma.us From: Marc Horowitz is there a workaround? Auth: yes Time: Sun Oct 14 13:39:56 2007 Host: curl-up-paw-under-nose.thok.org From: The gene pool could use a little chlorine. I added a python script to our build system, after it runs svn for anything, it fixes the tty. Auth: yes Time: Sun Oct 14 13:40:07 2007 Host: central-air-conditioning.toybox.cambridge.ma.us From: Marc Horowitz lovely. Auth: yes Time: Sun Oct 14 13:40:38 2007 Host: curl-up-paw-under-nose.thok.org From: I like these calm little moments before the storm. It reminds me of... Beethoven. actually what was still in my browser was the patch that introduced ssh setting non-blocking in the first place, so I could ask the author what he was thinking (since the change was completely undocumented.) Auth: yes Time: Sun Oct 14 13:41:03 2007 Host: curl-up-paw-under-nose.thok.org From: emacs - thermonuclear word processor but I haven't gotten around to further stalking. Auth: yes Time: Sun Oct 14 13:41:35 2007 Host: central-air-conditioning.toybox.cambridge.ma.us From: Marc Horowitz I have seen it claimed that you can only select on a nonblocking fd. I don't believe it, but maybe some obscure system has this restriction. Auth: yes Time: Sun Oct 14 13:41:59 2007 Host: central-air-conditioning.toybox.cambridge.ma.us From: Marc Horowitz hopefully, unrelated: > svk push Cannot find the path which '' copied from. Auth: yes Time: Sun Oct 14 13:42:10 2007 Host: curl-up-paw-under-nose.thok.org From: ... are there going to be meetings? "Of course!" @b(No meetings.) *blam* I don't believe it either, and even if it were true, it's a bug that the commit doesn't say that :-) Auth: yes Time: Sun Oct 14 13:42:24 2007 Host: central-air-conditioning.toybox.cambridge.ma.us From: Marc Horowitz sure, if it's that obscure :-) Auth: yes Time: Sun Oct 14 13:43:57 2007 Host: central-air-conditioning.toybox.cambridge.ma.us From: Marc Horowitz what's causing push to choke? Auth: yes Time: Sun Oct 14 14:00:24 2007 Host: central-air-conditioning.toybox.cambridge.ma.us From: Marc Horowitz No clues? svn smerge doesn't work either, and I need to get this changes onto another machines... Auth: yes Time: Sun Oct 14 14:03:59 2007 Host: VINEGAR-POT.MIT.EDU From: not worth transcribing what does 'svk info' say? Auth: yes Time: Sun Oct 14 14:07:26 2007 Host: central-air-conditioning.toybox.cambridge.ma.us From: Marc Horowitz > svk info Checkout Path: /u1/home/marc/sandbox/gnus Depot Path: //mirror/gnus Revision: 41 Last Changed Rev.: 41 Mirrored From: svn+ssh://svn.nerd-militia.org/home/marc/.svnrepos/gnus/trunk, Rev. 24 Auth: yes Time: Sun Oct 14 14:10:19 2007 Host: VINEGAR-POT.MIT.EDU From: not worth transcribing um, what do you express push to do? Auth: yes Time: Sun Oct 14 14:10:36 2007 Host: VINEGAR-POT.MIT.EDU From: not worth transcribing looks like you're in a checkout of trunk... Auth: yes Time: Sun Oct 14 14:10:50 2007 Host: central-air-conditioning.toybox.cambridge.ma.us From: Marc Horowitz merge the changes I just committed to the mirror? Auth: yes Time: Sun Oct 14 14:11:05 2007 Host: VINEGAR-POT.MIT.EDU From: not worth transcribing but you're in a direct checkout of the mirror, not a locl branch Auth: yes Time: Sun Oct 14 14:11:47 2007 Host: VINEGAR-POT.MIT.EDU From: not worth transcribing i would expect anything you've committed there to already be in ~/.svnrepos/gnus/trunk Auth: yes Time: Sun Oct 14 14:12:21 2007 Host: VINEGAR-POT.MIT.EDU From: not worth transcribing push/pull/smerge move changes around paths in a depot, but committing back to a mirror just happens when you commit to a mirrored path Auth: yes Time: Sun Oct 14 14:12:29 2007 Host: central-air-conditioning.toybox.cambridge.ma.us From: Marc Horowitz oh. huh. How did that happen. in fact, the bits are in the repository. thanks :-) Auth: yes Time: Sun Oct 14 14:23:19 2007 Host: tempest.toybox.cambridge.ma.us From: Marc Horowitz when I mirror an upstream repository, does it copy the whole thing? or do something less extreme? Auth: yes Time: Sun Oct 14 14:30:14 2007 Host: VINEGAR-POT.MIT.EDU From: not worth transcribing It copies whatever part you tell it to mirror Auth: yes Time: Sun Oct 14 14:30:35 2007 Host: VINEGAR-POT.MIT.EDU From: not worth transcribing (eg, you can mirror just trunk) Auth: yes Time: Sun Oct 14 14:31:06 2007 Host: tempest.toybox.cambridge.ma.us From: Marc Horowitz sure. but let's say it's the entire linux kernel for all history or something. do I need to copy the entire kernel history to make a simple patch? Auth: yes Time: Sun Oct 14 14:31:14 2007 Host: VINEGAR-POT.MIT.EDU From: not worth transcribing also, you can tell it not to mirror history (ie, start with HEAD) but I don't generally recommend that; it kind of defeats a lot of the point of svk. (and absolutely don't do that for the root of a repo, just for a single branch) Auth: yes Time: Sun Oct 14 14:32:07 2007 Host: tempest.toybox.cambridge.ma.us From: Marc Horowitz oh, you can give it a range. that seems completely reasonable. Auth: yes Time: Sun Oct 14 14:32:24 2007 Host: VINEGAR-POT.MIT.EDU From: not worth transcribing since when does linux use svn? :) Auth: yes Time: Sun Oct 14 14:32:37 2007 Host: tempest.toybox.cambridge.ma.us From: Marc Horowitz it was a theoretical question :-) Auth: yes Time: Sun Oct 14 14:32:53 2007 Host: tempest.toybox.cambridge.ma.us From: Marc Horowitz none of my personal projects are nearly large enough for it to be an issue :-) Auth: yes Time: Sun Oct 14 14:33:26 2007 Host: VINEGAR-POT.MIT.EDU From: not worth transcribing yeah, but just watch out. recall that copies (including branches) are stored c-o-w in an svn repo, but if you say "sync all of repo X but just -rHEAD", you can't take advantage of that (copies have to be from older revisions), so you end up with one revision which contains all the branches *explicitly* Auth: yes Time: Sun Oct 14 14:35:35 2007 Host: tempest.toybox.cambridge.ma.us From: Marc Horowitz I'm not sure what you mean by "explicitly" Auth: yes Time: Sun Oct 14 14:53:37 2007 Host: LUNATIQUE.MIT.EDU From: It will contain a full copy of all the data in each of the branches, rather than sharing information for files and directories that haven't changed between the trunk and the branches Auth: yes Time: Sun Oct 14 14:55:20 2007 Host: tempest.toybox.cambridge.ma.us From: Marc Horowitz ok. but if I only mirror one branch, that doesn't matter as much, right? and if I were to mirror the last month, say, then any branches created since then would be shared? Auth: no Time: Sat Nov 10 21:23:26 2007 Host: YIFF.MIT.EDU From: Joshua Pollack Oops, and at the end it segfaults Warning: unsupported node type /factrest/rev2/new-unpacked/dev/zero. Merge back committed as revision 2. Syncing file:///u1/frtestrepo Retrieving log information from 2 to 2 Segmentation fault Auth: no Time: Mon Dec 31 15:39:21 2007 Host: CYBERTRON.MIT.EDU From: "(set! *disallow-illegal-definitions?* #t)" --syntax.scm does anyone know how I can bootstrap an svk mirror from a svn dump where the mirror paths differ? Auth: yes Time: Tue Jan 29 19:32:45 2008 Host: tempest.toybox.cambridge.ma.us From: Marc Horowitz I did an svk mirror. then, svk sync runs for a while, and craps out: Retrieving log information from 9006 to 9929 RA layer request failed: REPORT request failed on 'https://svn.sqprod.com/svn/azrael/trunk': REPORT of 'https://svn.sqprod.com/svn/azrael/trunk': Could not read chunk delimiter: Secure connection truncated (https://svn.sqprod.com) at /System/Library/Perl/Extras/5.8.8/SVK/Mirror/Backend/SVNRaPipe.pm line 214 the apache logs say: Tue Jan 29 22:16:20 2008] [error] [client 98.216.109.205] Problem replaying revision [404, #160013] [Tue Jan 29 22:16:20 2008] [error] [client 98.216.109.205] File not found: revision 9042, path '/azrael/trunk/web/db' [404, #160013] what do I need to do to make this work? Auth: yes Time: Sat Feb 2 16:43:15 2008 Host: tempest.toybox.cambridge.ma.us From: Marc Horowitz maybe more people are online now. how can this happen, and is there an easy way to fix it: $ svk sync //mirror/azrael Syncing https://svn.sqprod.com/svn/azrael Retrieving log information from 9944 to 9944 Committed revision 1976 from revision 9944. $ svn proplist https://svn.sqprod.com/svn/azrael/trunk/client-common/message/msggen.pl Properties on 'https://svn.sqprod.com/svn/azrael/trunk/client-common/message/msggen.pl': svn:executable $ svk proplist //mirror/azrael/trunk/client-common/message/msggen.pl $ Auth: no Time: Wed Apr 16 17:39:19 2008 Host: VINEGAR-POT.MIT.EDU From: Trespassers will be shot; survivors will be shot again. Is this expected behavior? brainchild:~% SVKROOT=/home/ctl/bc/.svk svk depotmap --list Repository /home/ctl/.svk/local does not exist, create? (y/n) Auth: no Time: Sun Jan 25 14:05:25 2009 Host: VINEGAR-POT.MIT.EDU From: Values will be adjusted, and enterprising people will pick up from less competent people. "svk di -r'{2009-01-26}'" is giving me the changes since today. Today is the 25th. What is going on here? Auth: no Time: Sun Jan 25 14:05:39 2009 Host: VINEGAR-POT.MIT.EDU From: The result is that our possibilities of wealth may run to waste for a time --— perhaps for a long time. s/since today/since this morning/ Auth: yes Time: Sun Jan 25 14:06:32 2009 Host: VINEGAR-POT.MIT.EDU From: what is a graboid and are there any warnings Time zones? Auth: yes Time: Sun Jan 25 14:06:50 2009 Host: VINEGAR-POT.MIT.EDU From: what is a graboid and are there any warnings Or, well. Is it giving you more than one? Auth: yes Time: Sun Jan 25 14:07:08 2009 Host: VINEGAR-POT.MIT.EDU From: what is a graboid and are there any warnings {DATE} resolves to "the revision that was HEAD at the moment of DATE" Auth: yes Time: Sun Jan 25 14:07:20 2009 Host: VINEGAR-POT.MIT.EDU From: what is a graboid and are there any warnings which here means "the revision that will be HEAD at midnight", eg, HEAD Auth: yes Time: Sun Jan 25 14:07:37 2009 Host: VINEGAR-POT.MIT.EDU From: what is a graboid and are there any warnings (ignore the "more than one", I thought you had said log) Auth: no Time: Sun Jan 25 14:08:28 2009 Host: VINEGAR-POT.MIT.EDU From: Values will be adjusted, and enterprising people will pick up from less competent people. I'm guessing this is some GMT thing. Auth: no Time: Sun Jan 25 14:10:52 2009 Host: VINEGAR-POT.MIT.EDU From: High costs of living and high living will come down. OK, yes, my observations are consistent with "it becomes the 25th at 7pm on the 24th". Auth: yes Time: Mon Nov 12 13:36:18 2012 Host: SYSTEM-LOW-SIPB.MIT.EDU From: very subtle joke I have heard that svk is "no longer actively developed". Is that true, and does that mean that it should not be recommended as a VCS option? Auth: yes Time: Mon Nov 12 13:38:04 2012 Host: XVM-FOUR-117.MIT.EDU From: Adam Glasgall AIUI, no one should be usiung svk. Auth: yes Time: Mon Nov 12 13:39:10 2012 Host: TEAM-ROCKET.MIT.EDU From: https://mitcruft.org/ - show your school pride with an HMAC http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=554335 Auth: yes Time: Mon Nov 12 13:39:45 2012 Host: TEAM-ROCKET.MIT.EDU From: https://mitcruft.org/ - show your school pride with an HMAC http://lists.bestpractical.com/pipermail/svk-devel/2009-May/001224.html Auth: yes Time: Mon Nov 12 13:40:46 2012 Host: TEAM-ROCKET.MIT.EDU From: https://mitcruft.org/ - show your school pride with an HMAC "... this marks the end of SVK's 6 year life as a standalone development tool ... SVK's core technology will live on as part of Pushmi, our enterprise read/write replication system for Subversion." "for at least the next 18 months, we're committed to making sure that SVK will continue to work with the current release of Subversion", as of May 2009. Auth: yes Time: Mon Nov 12 13:47:51 2012 Host: chmrr.net From: moose. svk is no longer maintained, yes. "you can get it from its git repo" Auth: yes Time: Mon Nov 12 13:49:16 2012 Host: chmrr.net From: moose. It will oboviously still do what it used to do, but is not guaranteed to work with recent versions of SVN.