Extremely slow after ~50% of repo are analyzed

[kkmuffme]

I have a repo with ~70000 commits.
--analyze is rather quick (minutes) until it reaches about 35000 commits, and from that point on it takes ages (= days)

a) the commits I want to change are all in the same branch or sub-branches of it and less than 2 days old.

It would be great if there were an option to limit analyze to e.g. a specific commit and commits after it?

b) any idea what could cause this slowness?

[newren]

My first guess would be that in the first 35000 commits, you didn't delete or rename files, but afterward you did. Likely in the form of entire directories being moved or deleted, causing rename detection to be very expensive.

If you don't mind losing out on renames, you could change the line which reads

    cmd = ('git rev-list --topo-order --reverse {}'.format(' '.join(args.refs)) +
           ' | git diff-tree --stdin --always --root --format=%H%n%P%n%cd' +
           ' --date=short -M -t -c --raw --combined-all-paths')

in the source to instead be

    cmd = ('git rev-list --topo-order --reverse {}'.format(' '.join(args.refs)) +
           ' | git diff-tree --stdin --always --root --format=%H%n%P%n%cd' +
           ' --date=short -t -c --raw --combined-all-paths')

i.e. to remove the -M. Does that make it run faster for you?