Hi all, Reviving #868: https://github.com/apache/cloudberry/discussions/868 — renaming gp*/greenplum* identifiers to cb*/cloudberry* with aliases. The surface: CLI tools (gpstart, gpstop, gpconfig… → cbstart, cbstop…), catalog (gp_segment_configuration, gp_distribution_policy…), GUCs and system functions (gp_role, gp_session_id…), SQL keywords, error messages, docs. We should move on this soon.
Benefits are well-rehearsed: ASF trademark hygiene, Cloudberry's own identity, clean namespace for cb_ai_* and friends. Two things have stalled it before, both addressable now. First, bulk renames are exactly where AI does better than humans. Missed instances in comments and strings, inconsistencies across files, fat-fingered names in one file out of fifty — the typical human failure modes are precisely what coding agents don't do. They don't get bored, don't skim, don't miss dusty corners. With our test suite and PR review backstopping, what used to be a quarter of error-prone toil is a few weeks of agent-driven work. Second, past rename passes polluted history — `git blame` archaeology often lands on "rename gpdb → cbdb" with the real change 2–3 hops back. This time we should land it as a single atomic commit and add the SHA to `.git-blame-ignore-revs`. Git skips listed revs in blame and the rename becomes invisible to archaeology. PostgreSQL, Rust, and others use this pattern. Let's move on this. Anyone willing to drive it — or own a slice (CLI, catalog, docs) — would be doing the project a real favor. Best, Zhang Mingli
