Michael G Schwern <schwern@xxxxxxxxx> wrote: > On 2012.7.26 10:18 PM, Junio C Hamano wrote: > > Again, I agree with you that passing $prefix as one of the arguments > > to ->new is the right thing to do in the final state after applying > > the whole series. I don't know if later steps in your patch series > > will do so, but it _might_ make more sense to update ->new and its > > callers to do so without doing anything else first, so that you do > > not have to call out to the ::opt_prefix() when you split things > > out. > > I don't personally plan on doing any more about it, no. It isn't needed for > SVN 1.7, there's very little real code change (which you could see by looking > at my remote instead of waiting to be fed patches...) and its a very, very > minor problem in the grand scheme. I agree, its not worth it right now. > The first step toward that would be to change git-svn so it can be loaded as a > library using the standard "main() unless caller" trick. Then Git::SVN unit > tests can require git-svn as a library without executing it and get some tests > written with a minimum of Git::SVN code change. > None of which I plan to get into just now. That's fine. The modules were an afterthought and not intended at the time for standalone use, so it'd take a bit of work. I doubt the modules will be useful elsewhere, but will make code easier to maintain in the future. I also value functional/integration tests far more than unit tests. -- To unsubscribe from this list: send the line "unsubscribe git" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html