Does it connect to your actual database as well to get the list of tables, columns, etc? Or just leverage the same code base to be able to parse the raw text without any data dictionary information into a parse tree?
it leverages your code-base, rather than connecting to your database. it uses to libpg_query (https://github.com/pganalyze/libpg_query) to construct the syntax tree which can then be used to for the LSP features
It will definitely be supported at some point. We will have some features that require a db connection, eg autocompletion or code actions such as „Execute the current statement“. But always as an opt in.
