From 27267e3daa09cacaf017bfa34ca7e0b1e6989b0b Mon Sep 17 00:00:00 2001
From: "J. Hinchliffe" <jjh413@student.bham.ac.uk>
Date: Wed, 25 Jun 2025 17:03:48 +0100
Subject: [PATCH] assembler: use smart pointer for modules since sourceinfo
 gets copy

---
 assembler/src/model/module_registry.rs | 10 ++--
 assembler/src/source/tokeniser.rs      | 68 ++++++++++++++++++--------
 2 files changed, 52 insertions(+), 26 deletions(-)

diff --git a/assembler/src/model/module_registry.rs b/assembler/src/model/module_registry.rs
index 72ed50a..21e4094 100644
--- a/assembler/src/model/module_registry.rs
+++ b/assembler/src/model/module_registry.rs
@@ -1,13 +1,13 @@
 //! This module contains the code for the module registry. This is a singleton storing all
 //! the modules being assembled.
 
-use std::collections::HashMap;
+use std::{collections::HashMap, sync::Arc};
 
 use super::module::{Module, ModuleId};
 
 /// Stores all the [`Module`]'s to be assembled.
 pub struct ModuleRegistry {
-    modules: HashMap<ModuleId, Module>,
+    modules: HashMap<ModuleId, Arc<Module>>,
 }
 
 impl Default for ModuleRegistry {
@@ -26,19 +26,19 @@ impl ModuleRegistry {
 
     /// Gets a [`Module`] by ID.
     #[must_use]
-    pub fn get(&self, module_id: &ModuleId) -> Option<&Module> {
+    pub fn get(&self, module_id: &ModuleId) -> Option<&Arc<Module>> {
         self.modules.get(module_id)
     }
 
     /// Adds a [`Module`] and returns its [`ModuleId`].
-    pub fn add(&mut self, module: Module) -> ModuleId {
+    pub fn add(&mut self, module: Arc<Module>) -> ModuleId {
         let id = module.id;
         self.modules.insert(id, module);
         id
     }
 
     /// Returns an iterator of modules.
-    pub fn modules(&self) -> impl Iterator<Item = &Module> {
+    pub fn modules(&self) -> impl Iterator<Item = &Arc<Module>> {
         self.modules.values()
     }
 }
diff --git a/assembler/src/source/tokeniser.rs b/assembler/src/source/tokeniser.rs
index 1f3c588..b996c88 100644
--- a/assembler/src/source/tokeniser.rs
+++ b/assembler/src/source/tokeniser.rs
@@ -1,7 +1,10 @@
 //! This file contains the [`Tokeniser`], which consumes a [`Vec`] of input bytes and
 //! outputs a [`Vec<Token>`].
 
-use std::path::{Path, PathBuf};
+use std::{
+    path::{Path, PathBuf},
+    sync::Arc,
+};
 
 use regex::Regex;
 
@@ -10,7 +13,10 @@ use crate::{
     error::{AssembleError, AssembleErrorKind, IoError, IoErrorKind},
     model::module::Module,
     source::{
-        lines::lines_with_spans, load_source_bytes, token::Token,
+        lines::lines_with_spans,
+        load_source_bytes,
+        source_info::SourceInfo,
+        token::{Token, TokenType},
     },
 };
 
@@ -69,37 +75,57 @@ impl Tokeniser {
 
     // Note that modules are tokenised in their own threads, possibly in parallel.
     pub fn tokenise(self, ctx: &AssemblerContext) -> Result<Vec<Token>, AssembleError> {
-        let lines = lines_with_spans(&self.data);
-
-        let Some(module_name) = self.path.file_name().and_then(|f| f.to_str()) else {
-            return Err(AssembleError::new_other_error(AssembleErrorKind::Io(
-                IoError::new(
-                    IoErrorKind::InvalidData,
-                    Some(
-                        "filename couldn't be extracted, is it valid UTF-8?".to_string(),
-                    ),
-                ),
-            )));
-        };
+        let module_name = self.extract_module_name()?;
+        let _file_path = self.path.to_string_lossy().to_string();
 
         // Create a module for the source file being processed.
-        let module = Module::new(module_name.to_string(), &self.path);
+        let module = Arc::new(Module::new(module_name, &self.path));
 
         {
             let mut module_registry = ctx.module_registry.write()?;
-            module_registry.add(module);
+            module_registry.add(module.clone());
         }
 
+        let mut token_stream = Vec::new();
+        let lines = lines_with_spans(&self.data);
+
         // Technically ignores newlines since line will be trimmed. We just append a
         // Newline token for each line.
         for line_result in lines {
-            let line = line_result?;
+            let line_span = line_result?;
 
-            eprintln!("{}", line.line_number);
+            // Skip empty lines and comments
+            let trimmed = line_span.content.trim();
+
+            // Add newline token on blank lines.
+            if trimmed.is_empty() {
+                token_stream.push(Token::new(
+                    TokenType::Newline,
+                    SourceInfo::new(line_span.line_number, module.clone(), 0..1),
+                ));
+                continue;
+            }
+
+            eprintln!("{}", line_span.line_number);
         }
 
-        Err(AssembleError::new_other_error(
-            AssembleErrorKind::Unimplemented("tokeniser not written yet!"),
-        ))
+        Ok(token_stream)
+    }
+
+    fn extract_module_name(&self) -> Result<String, AssembleError> {
+        let module_name = self
+            .path
+            .file_name()
+            .and_then(|f| Some(f.to_string_lossy().to_string()))
+            .ok_or_else(|| {
+                AssembleError::new_other_error(AssembleErrorKind::Io(IoError::new(
+                    IoErrorKind::InvalidData,
+                    Some(
+                        "filename couldn't be extracted, is it valid UTF-8?".to_string(),
+                    ),
+                )))
+            })?;
+
+        Ok(module_name)
     }
 }