Refactor environment to use stack-based scoping

The commit rewrites the environment to use a stack of hash maps for
managing variable scopes. This replaces the old parent-reference
approach with a simpler and more efficient stack-based model.

Key changes: - Rename Environment to EnvironmentStack - Store scopes in
a Vec of HashMaps - Add push/pop scope operations for block handling -
Update interpreter to properly manage scope lifetimes - Clean up error
handling with helper functions

src/main.rs

@@ -6,19 +6,19 @@ mod result;
 
 use crate::{
     backend::{
-        environment::Environment,
+        environment::EnvironmentStack,
         interpreter::{EvaluateInterpreter, Interpreter},
     },
     frontend::{
         lexer::Lexer,
         parser::Parser,
-        source_registry::{SourceFile, SourceId, SourcePosition, SourceRegistry, SourceSlice},
+        source_registry::{SourceId, SourceRegistry, SourceSlice},
     },
     prompt::prompt_lines,
     result::{LoxError, LoxResult},
 };
+use std::env;
 use std::fs;
-use std::{env, error::Error};
 
 fn main() -> LoxResult<()> {
     let args: Vec<String> = env::args().collect();
@@ -83,8 +83,7 @@ impl LoxInterpreter {
                 Ok(asts)
             })
             .and_then(|stmts| {
-                let mut env = Environment::new();
-                let mut interpreter = Interpreter::new(&mut env);
+                let mut interpreter = Interpreter::new();
                 let mut result = None;
                 for (index, stmt) in stmts.iter().enumerate() {
                     println!("executing stmt {}: {:?}", index, stmt);