On Thu, Sep 20, 2012 at 12:13:45AM -0400, Nico Kadel-Garcia wrote: > On Wed, Sep 19, 2012 at 3:34 PM, Alexander Kitaev <kit...@tmatesoft.com> > wrote: > > Hello All, > > > > Our team is proud and delighted to announce SubGit 1.0.0 release! > > New version is available for download at SubGit web site at > > http://subgit.com/ > > > > SubGit is a server-side tool for a smooth, stress-free Svn to Git > > migration. SubGit lets one to set up a bidirectional Subversion to Git > > replication (writable mirror) and thus it allows users to choose > > freely between Subversion and Git version control systems. > > Except, of course when it doesn't. The use of OS specific EOL, which > git does not support, and subversion keywords like $Id$ and $Author$, > which git does not support, would seem to me to be an adventure > begging to happen.
Of course there are problems with trying to do this. But if you're always afraid of running into pitfalls you'll never get anything done. I know you are particularly keen on pointing out pitfalls because you're somebody who often ends up having to work around them professionally, and I sense some love-hate relationship going on here :) We had some very interesting discussions with the Tmate folks when they first announced SubGit at Subversion Day last year: http://www.elegosoft.com/en/company/news/news-in-detail.html?tx_ttnews[pointer]=1&tx_ttnews[tt_news]=114&tx_ttnews[backPid]=195&cHash=f876117be36ef5742b4977d7be07b5da The announcement sparked some interest in German media: http://www.heise.de/developer/meldung/Transparenter-Abgleich-zwischen-Subversion-und-Git-1246894.html My own conviction is that their entire endeavor of rewriting Subversion and related components in Java is outright crazy. But I'm not a Java developer, so maybe I just don't get it. And the quality that Tmate produces is astounding, and the persistence they show with keeping up with Subversion's feature set in SVNKit is highly commendable. Generally, the problem they're tackling with SubGit is a giant compatibility mapping problem between two very different data structures. It is quite interesting to see how much of that problem space they can actually solve, and also how long it has taken them to solve it! One of the developers at Tmate even made this problem part of his thesis. So please, before you dismiss these efforts by pointing out the first minor potential compatibility problem that comes to mind, pay them some respect for their efforts -- they've certainly done their homework around keyword expansion, and lots more.
