Haskell acceptance tests and admin database parser changes

[levkk]

Tests

Added a short Haskell integration tests suite.

Admin database

Changed how we parse commands (and created a bit of a mess). We first attempt to detect SET and SHOW using the pg parser. If that fails, we fallback to string interpolation.

This allows us to respond to common commands executed by database drivers, e.g., SHOW server_version (which now returns the parser version we get from pg_raw_parse).

This also allows us to parse and respond to multi-statement SET queries, e.g.:

SET application_name TO 'persistent'; SET client_encoding TO 'utf8';

without those getting mangled by our "string parser". We don't do anything if we see a variable we don't recognize, by the way.

To add to the complexity, we have some commands that would match the show branch but not be valid SHOW commands, e.g., SHOW CLIENTS <column list>. Handling for these is duplicated because apparently that's the best I can come up with today.

[codecov]

Codecov Report

❌ Patch coverage is 98.24561% with 2 lines in your changes missing coverage. Please review.

Files with missing lines	Patch %	Lines
pgdog/src/admin/show_guc.rs	85.71%	2 Missing ⚠️

📢 Thoughts on this report? Let us know!