Re: Add --system-identifier / -s option to pg_resetwal

Thank you for the review!

Attached is v3 addressing your feedback:

- Changed label to "Database system identifier" for consistency
- Use PRIu64 instead of UINT64_FORMAT for gettext
- Moved option entry after "char-signedness" in getopt array
- Fixed test to use pg_controldata instead of pg_control_system()

On Wed, Jun 4, 2025 at 4:56 PM Fujii Masao <masao(dot)fujii(at)oss(dot)nttdata(dot)com>
wrote:

> One question about this feature: why do we need to allow
> explicitly setting the system identifier? If the goal is
> simply to ensure it's different from the original,
> wouldn't it be sufficient to let pg_resetwal generate
> a new (hopefully unique) value using existing logic,
> like what's done in GuessControlValues() or BootStrapXLOG()?

Auto-generation would cover most cases. But explicit setting is a superset
-- users who want random can generate one themselves. Explicit control
costs nothing extra and gives full flexibility for edge cases (eg,
reproducible test environments, coordinating multiple clones). pg_resetwal
users are expected to be experts, so giving more control can cover more
possible scenarios.

Nik