I've never gotten the point of squash-merge. The normal "merge into release branch, never fast-forward" workflow retains more useful information. Every release

I've never gotten the point of squash-merge. The normal "merge into release branch, never fast-forward" workflow retains more useful information. Every release commit has 2 parents: the previous release commit, and the former head of the feature branch it got its changes from. If you want a chain of releases, you go down the path of commits labeled "RELEASE". If you want information on the changes, go down the other path. I regularly use "git blame" on lines of code to try and figure out what the specific reasoning was on a line of code. I'm going to have a very different interpretation depending on if the resulting commit says "implement feature xyz" vs "fix issue abc on feature xyz". Squash murders that added context. At best you're optimizing for the wrong thing, and worst it's outright vandalism.