This is the source code for a tool that was written in the 2006-2007 timeframe by myself and several others to convert a Visual SourceSafe (VSS) repository into the Subversion "dumpfile" format so that its history can be imported into a Subversion repository.
Active development on this tool has long since been ended but I decided to import it here in case anyone else finds it useful. Feel free to fork it but I can't promise I'll accept pull requests as I don't have the resources to verify or test your code.
I originally wrote this on Windows but it's probably easier to get it up and running on Linux since the former doesn't work as well with Perl. The C program should compile fine on either platform.
A while after initial development on this project ceased, Alexander Gavrilov submitted some additional patches which I was not able to test and merge, but they are available [in the develop branch] (https://github.com/irontoby/vss2svn/tree/develop).
The most up-to-date documentation can be found on the Google Code Wiki, which has also been copied to the wiki here, though some formatting/links are broken.
What makes this tool different from many other similar tools is that it does not require the Microsoft VSS API to be installed to use it. Instead, one rather intrepid developer in our group went and reverse-engineered the VSS physical storage format and wrote a C program to parse those files and output the results into an XML format.
The other half of this project is a Perl program that reads the resulting XML output and assembles it into the Subversion dumpfile format. The dumpfile format is a complete archive of a repository and can be imported using the "svnadmin load" command.
The end result is that this tool often does a better job than even Microsoft's libraries at extracting old history. Unfortunately VSS repositories are so susceptible to corruption that any repo of significant size, age, and/or usage is almost guaranteed to be at least partially corrupted. The farther back in time the history goes, the more likely it is to be less than correct.
So the general upshot is that, while this tool often works "better than most" at getting your true code history out of VSS, some of that history may be gone forever. It's just the nature of the beast. Maybe switching to a new Version Control System is a good chance to refactor your code base anyway? You may want to consider migrating your old history to Subversion for historical reference only, then starting fresh for further active development.
Unfortunately the name "Vss2Svn" turned out to be rather generic and there is at least one other completely unrelated version of a tool that serves the same general purpose.
This tool has had various homes throughout the years. Open source code collaboration wasn't nearly as good as GitHub back then, ya know. Previous homes include:
- http://vss2svn.tigris.org/
- A personal Trac/Subversion site I set up at PumaCode.org (don't go there; it's run by spammers now)
- https://code.google.com/p/vss2svn/
THE SOURCE CODE ON GITHUB SHOULD BE CONSIDERED THE "LATEST" VERSION and supercedes all previous versions of the codebase.