Normalize PostgreSQL identifiers in pgdog.toml

[murex971]

PgDog treats configured identifiers as literal, case-sensitive strings. It should follow PostgreSQL semantics

• lowercase unquoted identifiers
• preserve quoted identifiers
  ensuring configuration matches parsed SQL consistently.

example:

• MyFunction() → myfunction
• "MyFunction"() → MyFunction

as per comment

[nsavvide]

Hey, I would love to contribute to this project as a first-timer :)). Will take it up right now.

I am planning a normalize_identifier helper in pgdog-config applied at load time in config::set() so all downstream matches compares the folded values. Also, which config fields count here? I have found sharded_tables.{name, column, schema} and the multi-tenant column. Noting that #1236 introduces the same normalization for write_functions, we should have a public helper to use in many other places.

[nsavvide]

Hey, I would love to contribute to this project as a first-timer :)). Will take it up right now.

I am planning a normalize_identifier helper in pgdog-config applied at load time in config::set() so all downstream matches compares the folded values. Also, which config fields count here? I have found sharded_tables.{name, column, schema} and the multi-tenant column. Noting that #1236 introduces the same normalization for write_functions, we should have a public helper to use in many other places.

Just tested the usecase, and indeed, pgdog does not normalize the identifiers in the config yet. A helper will be effective here.

[nsavvide]

^ PR has been made. Please review. As a whole, the logic change is not much and should be easy enough to understand. Here you go :))