On Tue, 6 Sep 2011, Robert P. Goldman wrote:

P.S. I'd be interested in feedback about the use I made of git here. I *could* have squashed the series of commits that ended up with the patch you see here, but in the end, I decided that it might be helpful for people to see in the history the blind alley that I pursued, because it seemed like a simpler fix to begin with.

db16d0..8f5e2d I frequently like my history to show stuff like this. Some people prefer a clean, "always works" history. One approach is to treat the development work as a branch and merge it back into master with the summary message you would use in the squash. I think this caters to both styles. - Daniel