> If we support a wrapper we should support it for all pg_ctl usage IMO.
As i understand it - you propose to patch pg_ctl.c & regress.c instead of
PostgresNode.pm & regress.c?
This is a deeper invasion to pg_ctl. There will be a conflict between the
environment variable and the pg_ctl -p parameter, and possibly induced
bugs.
I suggest microscopic changes for iterative recompilation/test/debug
without installation.
> For the sake of others with similar needs I attach my current
> wrapper/launcher script. To use it you have to:
IMHO, the method what you proposed (wrapper/launcher) - is more suitable for
complex testing.
I agree that the my proposed way is incomplete.