3 ms·
> Git has arguably been falling behind in terms of features for quite a while now It’s all I’ve ever really used, and I’ve never seriously considered alternati
by confounded 8y ago
> Git has arguably been falling behind in terms of features for quite a while now
It’s all I’ve ever really used, and I’ve never seriously considered alternatives. What features in other VCSs are interesting?
- gboudrias 8y agoI'm not expert enough to answer that, but Fossil and its "Built-in Web Interface" seems really interesting. I remember Mercurial seemed interesting back in the day, but as far as I can tell it mostly didn't take off because it's too similar to Git. All in all, I'd just like a distributed scm that doesn't feel like it was designed to make Linux with 10,000 other people. It is... frustratingly sophisticated at times.
- chriswarbo 8y agoMercurial is very similar to git, but its CLI is considered to be easier than gits (if potentially less powerful). Another difference is that mercurial heavily discourages editing of history, which some git users seem to like doing. Git is also faster than mercurial. GNU arch was an early DVCS but didn't seem to gain much traction. It was eventually superceded by Bazaar, which is much slower than git but was better at tracking history e.g. across renames. I think git has improved in that area, and bazaar has been getting less popular. Fossil is built on an sqlite database and has a built-in Web UI and issue tracker, which is fine for those who want that but violates the "do one thing" principle (e.g. if you wanted to use some other issue tracker). Darcs keeps track of patches and their dependencies, unlike git and mercurial (which track snapshots and their diffs). This approach supposedly makes life easier, and makes the darcs CLI quite usable. One problem is that some of the merge algorithms can take an exponential amount of time to complete, which has been getting addressed in recent years but many users seem to jump ship to git. Pijul is similar to darcs, but breaks compatibility in order to solve the exponential merge problems. It looks very promising, but is still very new.