# Rerere subcommand conversion raises question about option order

_Junio Hamano asks whether moving git rerere to standard subcommand parsing silently breaks scripts that put options before the subcommand._

Section: Discussion. Edition of 2026-11-15. Written by an AI editor from the thread "[PATCH 00/12] builtin/rerere: convert to use subcommands" (108 messages): https://gitlist.dev/t/66677

Patrick Steinhardt posted a 12-patch series converting builtin/rerere to use subcommands. He said he was put off by the custom subcommand parsing in git-rerere(1) and wants it to behave like other Git commands with subcommands. The series is based on the 9th batch of integration with ps/rerere-remove-global-state merged in.

Junio Hamano asked whether the new syntax is a subtle UI regression. Git used to accept "git rerere --skip-locked gc". With the new parsing, the option has to be spelled "git rerere gc --skip-locked".

Hamano said he does not care much about --skip-locked itself. He also said he generally prefers options after non-options as a CLI principle. His worry is that Git has a large user base, so a change that might go unnoticed in a smaller project could affect a good number of people.

He said he does not know how often people run "git rerere" with these subcommands directly. The excerpts do not show Steinhardt's reply or whether the series will keep the old option placement. The question of compatibility with the old ordering appears to be open.
