Postgres Language Server: implementing the Parser

"Postgres Language Server: Implementing the Parser"

• The Postgres Language Server aims to improve the experience of writing SQL in editors like VSCode.
• The core component of a language server is the parser, which takes input source code and creates a semantic model of the code.
• Implementing a parser for Postgres is challenging due to the complex and constantly evolving syntax of Postgres.
• The parser needs to handle both expressions and statements, with statements being particularly difficult to parse.
• The current parser API only parses the entire input and returns an error if there is any syntax error, making it challenging to handle errors in specific statements.
• To implement the parser, we need to work with the AST and a stream of tokens.
• The LL parser we use is simple and relies on distinct keywords to determine the start of each statement.
• Ambiguity in statement identification is acceptable for our use case, as we only need to know if there is a statement and not the exact statement type.
• The parser should generate a concrete syntax tree (CST) with proper ranges for each node in the AST.