I stopped using Darcs a few years ago, but I heard the current generation at least resolved the notorious exponential time slowdowns.
Git's speed is definitely a big selling point. More than that, the ecosystem and services like GitHub are what really sold me on it versus alternatives. But Mercurial has a lot to offer and its simpler user interface, better Windows support and extensions like BFiles make it a much better fit for certain use cases.
I shouldn't have been so hasty to say that Mercurial doesn't store changes. But I'd argue, and you seem to agree, that Mercurial's revlog does not reflect a difference from Git in the basic philosophy of merging and the status and role of versions. In both cases you're basically dealing with genealogically annotated purely functional trees. By comparison, Darcs's theory of patches represents a radical departure. At the very least I'm happy that someone is trying to think deep and different thoughts in this area.
I stopped using Darcs a few years ago, but I heard the current generation at least resolved the notorious exponential time slowdowns.
Git's speed is definitely a big selling point. More than that, the ecosystem and services like GitHub are what really sold me on it versus alternatives. But Mercurial has a lot to offer and its simpler user interface, better Windows support and extensions like BFiles make it a much better fit for certain use cases.
I shouldn't have been so hasty to say that Mercurial doesn't store changes. But I'd argue, and you seem to agree, that Mercurial's revlog does not reflect a difference from Git in the basic philosophy of merging and the status and role of versions. In both cases you're basically dealing with genealogically annotated purely functional trees. By comparison, Darcs's theory of patches represents a radical departure. At the very least I'm happy that someone is trying to think deep and different thoughts in this area.