On 10/11/2011 11:11 PM, Buddy Burden wrote: > Guys, > > So, I found a bug in a CPAN module that hadn't been updated in some > time. After I submitted a bug in RT, I checked the author's other > modules and his RT tickets: no activity in years. So I sent the > author an email, and said, hey, if you don't want to mess with this > module any more, I'd be happy to take it over for you. And, voila, > I'm now the proud(?) maintainer of Data::Random. > > So I need to create a repo for the code, and it would be nice to start > with the previous versions, right? Happily, there is gitpan, so I can > pretty easily access all the CPAN versions of the code at least. Now, > I've read Schwern's blog[1] on how to merge gitpan history into an > existing repo, so I could essentially create a blank repo, do the > merge as he describes, and go from there. But it seems to me that's > unnecessarily complex, since I'm starting from scratch. What I'm > wondering is, couldn't I just fork the gitpan repo for Data::Random[2] > and have that become my repo? Obviously I would never plan on merging > it back, but that shouldn't matter, I don't think. > > Does anyone see any downsides to this plan, or have a better suggestion? TIA!
Forking from gitpan is a fine plan! I have even done it for several of my own modules when they needed some updating after years of being untouched and I was to lazy to go recovering them from my old archived subversion repository. Most modules are just not so complex to make having the full history available such a critical thing.
