Code review comment for lp:~gary/bzr/bug835035

Thanks, John.

The reason I didn't do that is because of this comment (of yours, it seems, from bzr annotate) that you can see in the diff:

             # This repository will call fallback.unlock() when we transition to
             # the unlocked state, so we make sure to increment the lock count

That doesn't happen in the case of bug 835035, but apparently it does in some common case or other that I don't understand (I don't even know what a "fallback" is in this context).

I could simply try making a change like you describe (using a "try/finally" or a "with" within add_fallback_repository) and then run the entire test suite and see if anything breaks. If it doesn't break, I guess I could remove the comment, or supply another that you suggested. Does that sound reasonable, or would you suggest something different?