Martin von Zweigbergk [off-list ref] writes:
This is probably quite off topic for the thread, but I'm curious why
you think it was a bad idea to have octopus merges in git.git's
history (there seem to be 37 of them).
Octoupi in our history, at least the older ones, solve no real life
problem; it only gives us "now we have something cool-looking that
other people's version control systems never had", which is of
dubious value.
And their presense makes bisection less efficient than it could be
around them, which is a real downside.