feat(lexer): Add token memory management infrastructure

- Add token_info_t structure with type, value, and source location tracking
- Add source_location_t for precise error reporting (line, column, filename)
- Add token_pool_t for freelist-based memory reuse pattern
- Add token_buffer_t for 8-token circular buffer lookahead capability
- Increase MAX_TYPES from 128 to 256 for additional type definitions
- Increase MAX_IR_INSTR from 60000 to 80000 for self-hosting capacity

These structures provide the foundation for significant memory savings
through token reuse and enable advanced parsing with multi-token lookahead.
The changes maintain full backward compatibility while preparing for
future optimizations.

🤖 Generated with [Claude Code](https://claude.ai/code)

Co-Authored-By: Claude <[email protected]>

Author: jserv

src/defs.h

@@ -19,8 +19,8 @@
 #define MAX_PARAMS 8
 #define MAX_LOCALS 1600
 #define MAX_FIELDS 64
-#define MAX_TYPES 128
-#define MAX_IR_INSTR 60000
+#define MAX_TYPES 256
+#define MAX_IR_INSTR 80000
 #define MAX_BB_PRED 128
 #define MAX_BB_DOM_SUCC 64
 #define MAX_BB_RDOM_SUCC 256
@@ -180,6 +180,37 @@ typedef enum {
     T_cppd_pragma
 } token_t;
 
+/* Source location tracking for better error reporting */
+typedef struct {
+    int line;
+    int column;
+    char *filename;
+} source_location_t;
+
+/* Token structure with metadata for enhanced lexing */
+typedef struct token_info {
+    token_t type;
+    char value[MAX_TOKEN_LEN];
+    source_location_t location;
+    struct token_info *next; /* For freelist management */
+} token_info_t;
+
+/* Token freelist for memory reuse */
+typedef struct {
+    token_info_t *freelist;
+    int allocated_count;
+    int reused_count; /* Statistics for debugging */
+} token_pool_t;
+
+/* Token buffer for improved lookahead */
+#define TOKEN_BUFFER_SIZE 8
+typedef struct {
+    token_info_t *tokens[TOKEN_BUFFER_SIZE];
+    int head;
+    int tail;
+    int count;
+} token_buffer_t;
+
 /* builtin types */
 typedef enum {
     TYPE_void = 0,