{"thread":{"id":"8907","subject":"git-svn set-tree","startedAt":"2007-07-07T00:03:47Z","lastAt":"2007-07-10T22:45:53Z","messageCount":3,"participants":["Tjernlund","Eric Wong","Joakim Tjernlund"],"isPatch":false,"patchVersion":null,"patchTotal":null},"messages":[{"id":"46665","messageId":"001501c7c02a$4bc130a0$0e67a8c0@Jocke","threadId":"8907","inReplyTo":null,"subject":"git-svn set-tree","fromName":"Tjernlund","fromEmail":"tjernlund@tjernlund.se","sentAt":"2007-07-07T00:03:47Z","receivedAt":"2007-07-07T00:03:47Z","isPatch":false,"sender":{"key":"tjernlund@tjernlund.se","avatar":null},"body":"I have noticed that if I do a git-svn set-tree, remotes/git-svn retains the parent\nfrom the branch where set-tree was performed.\n\nIf a coworker wants recreate my tree by using git-svn init && git-svn fetch\nhe looses the parent I have in my tree.\n\nI wonder if not git-svn set-tree can record the parent information in the\nsvn repos log, so that git-svn init/fetch can recreate the parent relationship?\n\n Jocke\n"},{"id":"46832","messageId":"20070709054541.GA2301@mayonaise","threadId":"8907","inReplyTo":"001501c7c02a$4bc130a0$0e67a8c0@Jocke","subject":"Re: git-svn set-tree","fromName":"Eric Wong","fromEmail":"normalperson@yhbt.net","sentAt":"2007-07-09T05:45:53Z","receivedAt":"2007-07-09T05:45:53Z","isPatch":false,"sender":{"key":"e@80x24.org","avatar":null},"body":"Tjernlund <tjernlund@tjernlund.se> wrote:\n> I have noticed that if I do a git-svn set-tree, remotes/git-svn\n> retains the parent from the branch where set-tree was performed.\n> \n> If a coworker wants recreate my tree by using git-svn init && git-svn\n> fetch he looses the parent I have in my tree.\n> \n> I wonder if not git-svn set-tree can record the parent information in\n> the svn repos log, so that git-svn init/fetch can recreate the parent\n> relationship?\n\nWe could at yet another non-standardized property into SVN to handle\nmerges.  Currently there are at least two properties used in the SVN/SVK\nworld to represent merges (Sam Vilain can give you the fun details of\neach one!).\n\nI'm afraid adding a third incompatible yet similair property for git-svn\nwould just confuse people.\n\nI've become very much against crazy stuff like set-tree which ends up\ncreating a M:N history mapping between git and svn.  1:1 is the simplest\nand easiest.  I'm more than willing to sacrifice multi-parent histories\nin git for easier compatibility with other systems.\n\nHeck, linear history is just easier to deal with and probably preferable\nin most/many cases.  I'm sure that the rising popularity of git-rebase,\nquilt, stgit, guilt, mq and other like tools is a testament to that.\n\n-- \nEric Wong\n"},{"id":"46989","messageId":"000901c7c344$12b431a0$02ac10ac@Jocke","threadId":"8907","inReplyTo":"20070709054541.GA2301@mayonaise","subject":"RE: git-svn set-tree","fromName":"Joakim Tjernlund","fromEmail":"joakim.tjernlund@transmode.se","sentAt":"2007-07-10T22:45:53Z","receivedAt":"2007-07-10T22:45:53Z","isPatch":false,"sender":{"key":"joakim.tjernlund@transmode.se","avatar":null},"body":" \n\n> -----Original Message-----\n> From: Eric Wong [mailto:normalperson@yhbt.net] \n> Sent: den 9 juli 2007 07:46\n> To: Tjernlund\n> Cc: git@vger.kernel.org\n> Subject: Re: git-svn set-tree\n> \n> Tjernlund <tjernlund@tjernlund.se> wrote:\n> > I have noticed that if I do a git-svn set-tree, remotes/git-svn\n> > retains the parent from the branch where set-tree was performed.\n> > \n> > If a coworker wants recreate my tree by using git-svn init \n> && git-svn\n> > fetch he looses the parent I have in my tree.\n> > \n> > I wonder if not git-svn set-tree can record the parent \n> information in\n> > the svn repos log, so that git-svn init/fetch can recreate \n> the parent\n> > relationship?\n> \n> We could at yet another non-standardized property into SVN to handle\n> merges.  Currently there are at least two properties used in \n> the SVN/SVK\n> world to represent merges (Sam Vilain can give you the fun details of\n> each one!).\n\nI have read his page, quite informative.\n\n> \n> I'm afraid adding a third incompatible yet similair property \n> for git-svn\n> would just confuse people. \n> \n> I've become very much against crazy stuff like set-tree which ends up\n> creating a M:N history mapping between git and svn.  1:1 is \n> the simplest\n> and easiest.  I'm more than willing to sacrifice multi-parent \n> histories\n> in git for easier compatibility with other systems.\n\nset-tree is needed for starting a SVN tree from a git tree, I don't\nknow of any other way to do that. I got both linux and u-boot\ngit trees in which I do custom mods. I then use git-svn \nto maintain a svn tree which I start with set-tree to commit\nthe initial tree, then I dcommit my local mods. \n\nMaybe maybe an option to git-svn init/clone where you can specify the git\nparent for a certain svn revision? \n\n> \n> Heck, linear history is just easier to deal with and probably \n> preferable\n> in most/many cases.  I'm sure that the rising popularity of \n> git-rebase,\n> quilt, stgit, guilt, mq and other like tools is a testament to that.\n>\n\ngit rebase is a bit annoying as you loose the old tree if you don't take precations.\nAlso, I am not sure what will be committed to SVN if I rebase my local mods on top\nof linus latest.\n\n Jocke\n \n> -- \n> Eric Wong\n> \n"}]}