Add basic AST functionality

Currently this is a subset of the grammar, enough to get reasonable work
going.

doc/parser_grammar.txt

@@ -0,0 +1,36 @@
+/* string literals are lexer identifier tokens with that particular value */
+<program>   ::= <statement>*
+<statement> ::= ( <label> | <directive> | <instruction> ) <newline>
+
+<label> ::= <identifier> <colon>
+
+<directive> ::= <dot> <section>
+
+<section>   ::= "section" <identifier>
+
+<instruction> ::= <identifier> <operands>
+
+<operands> ::= <operand> ( <comma> <operands> )*
+
+<operand>  ::= <register> | <immediate> | <memory>
+
+<register> ::= <register_base> | <register_extra>
+<register_base> ::= "rax" | "rbx" | "rcx" | "rdx" | "rsi" | "rdi" | "rbp" | "rsp"
+<register_extra> ::= "r8" | "r9" | "r10" | "r11" | "r12" | "r13" | "r14" | "r15" 
+
+<immediate> ::= <number> | <label_reference>
+<number> ::= <octal> | <binary> | <decimal> | <hexadecimal>
+
+<label_reference> ::= <identifier>
+
+<memory> ::= <lbracket> <memory_expression> <rbracket>
+
+<memory_expression> ::= <label_reference> | <register_expression>
+
+<register_expression> ::= <register> <register_index>? <register_offset>?
+
+<register_index> ::= <plus> <register> <asterisk> <number>
+
+<register_offset> ::= <plus_or_minus> <number>
+
+<plus_or_minus> ::= <plus> | <minus>

src/ast.c

@@ -0,0 +1,84 @@
+#include "ast.h"
+#include "error.h"
+#include <string.h>
+
+error_t *err_node_children_cap = &(error_t){
+    .message = "Failed to increase ast node children, max capacity reached"};
+
+error_t *ast_node_alloc(ast_node_t **output) {
+    *output = nullptr;
+
+    ast_node_t *node = calloc(1, sizeof(ast_node_t));
+    if (node == nullptr)
+        return err_allocation_failed;
+
+    *output = node;
+    return nullptr;
+}
+
+void ast_node_free_value(ast_node_t *node) {
+    // TODO: decide how value ownership will work and clean it up here
+}
+
+void ast_node_free(ast_node_t *node) {
+    if (node == nullptr)
+        return;
+    if (node->children) {
+        for (size_t i = 0; i < node->len; ++i)
+            ast_node_free(node->children[i]);
+    }
+
+    ast_node_free_value(node);
+
+    memset(node, 0, sizeof(ast_node_t));
+    free(node);
+}
+
+/**
+ * @pre node->children must be nullptr
+ */
+error_t *ast_node_alloc_children(ast_node_t *node) {
+    node->children = calloc(node_default_children_cap, sizeof(ast_node_t *));
+    if (node->children == nullptr)
+        return err_allocation_failed;
+
+    node->cap = node_default_children_cap;
+    return nullptr;
+}
+
+error_t *ast_node_grow_cap(ast_node_t *node) {
+    if (node->cap >= node_max_children_cap) {
+        return err_node_children_cap;
+    }
+
+    size_t new_cap = node->cap * 2;
+    if (new_cap > node_max_children_cap) {
+        new_cap = node_max_children_cap;
+    }
+
+    ast_node_t **new_children =
+        realloc(node->children, new_cap * sizeof(ast_node_t *));
+    if (new_children == nullptr) {
+        return err_allocation_failed;
+    }
+
+    node->children = new_children;
+    node->cap = new_cap;
+
+    return nullptr;
+}
+
+error_t *ast_node_add_child(ast_node_t *node, ast_node_t *child) {
+    error_t *err = nullptr;
+    if (node->children == nullptr)
+        err = ast_node_alloc_children(node);
+    else if (node->len >= node->cap)
+        err = ast_node_grow_cap(node);
+    if (err)
+        return err;
+
+    node->children[node->len] = child;
+    node->len += 1;
+
+    return nullptr;
+}

src/ast.h

@@ -0,0 +1,72 @@
+#ifndef INCLUDE_SRC_AST_H_
+#define INCLUDE_SRC_AST_H_
+
+#include "error.h"
+#include "lexer.h"
+#include <stddef.h>
+#include <stdint.h>
+
+typedef enum node_id {
+    NODE_PROGRAM,
+    NODE_DIRECTIVE,
+    NODE_LABEL,
+    NODE_INSTRUCTION
+} node_id_t;
+
+typedef struct ast_node ast_node_t;
+
+constexpr size_t node_default_children_cap = 8;
+/* 65K ought to be enough for anybody */
+constexpr size_t node_max_children_cap = 1 << 16;
+
+struct ast_node {
+    node_id_t id;
+    lexer_token_t *token;
+    size_t len;
+    size_t cap;
+    ast_node_t **children;
+
+    union {
+        struct {
+            uint64_t value;
+            int size;
+        } integer;
+        char *name;
+    } value;
+};
+
+/**
+ * @brief Allocates a new AST node
+ *
+ * Creates and initializes a new AST node with default (zero) values.
+ *
+ * @param[out] output Pointer to store the allocated node
+ * @return error_t* nullptr on success, allocation error on failure
+ */
+error_t *ast_node_alloc(ast_node_t **node);
+
+/**
+ * @brief Frees an AST node and all its children recursively
+ *
+ * Recursively frees all children of the node, then frees the node itself.
+ * If node is nullptr, the function returns without doing anything.
+ *
+ * @param node The node to free
+ */
+void ast_node_free(ast_node_t *node);
+
+/**
+ * @brief Adds a child node to a parent node
+ *
+ * Adds the specified child node to the parent's children array.
+ * If this is the first child, the function allocates the children array.
+ * If the children array is full, the function increases its capacity.
+ *
+ * @param node The parent node to add the child to
+ * @param child The child node to add
+ * @return error_t* nullptr on success, allocation error on failure,
+ *                  or err_node_children_cap if maximum capacity is reached
+ */
+error_t *ast_node_add_child(ast_node_t *node, ast_node_t *child);
+
+#endif // INCLUDE_SRC_AST_H_

src/error.c

@@ -10,6 +10,9 @@ error_t *const err_errorf_length = &(error_t){
     .message =
         "Formatting of another error failed to determine the error length"};
 
+error_t *err_allocation_failed =
+    &(error_t){.message = "Memory allocation failed"};
+
 error_t *errorf(const char *fmt, ...) {
     error_t *err = calloc(1, sizeof(error_t));
     if (err == nullptr)

src/error.h

@@ -18,4 +18,7 @@ static inline void error_free(error_t *err) {
     free(err);
 }
 
+/* Some global errors */
+extern error_t *err_allocation_failed;
+
 #endif // INCLUDE_SRC_ERROR_H_

src/lexer.c

@@ -20,9 +20,6 @@ error_t *err_eof =
 
 error_t *err_unknown_read = &(error_t){.message = "Unknown read error"};
 
-error_t *err_allocation_failed =
-    &(error_t){.message = "Memory allocation failed"};
-
 typedef bool (*char_predicate_t)(char);
 
 const char *lexer_token_id_to_cstr(lexer_token_id_t id) {

src/main.c

@@ -1,5 +1,6 @@
 #include "error.h"
 #include "lexer.h"
+#include "tokenlist.h"
 
 #include <limits.h>
 #include <stdio.h>
@@ -39,24 +40,33 @@ int main(int argc, char *argv[]) {
         print_fn = print_value;
     }
 
-    lexer_t lex = {0};
+    lexer_t *lex = &(lexer_t){};
-    lexer_token_t token;
+    error_t *err = lexer_open(lex, filename);
-    error_t *err = lexer_open(&lex, filename);
+    if (err)
-    if (err) {
+        goto cleanup_error;
+
+    tokenlist_t *list;
+    err = tokenlist_alloc(&list);
+    if (err)
+        goto cleanup_lexer;
+
+    err = tokenlist_fill(list, lex);
+    if (err)
+        goto cleanup_tokens;
+
+    for (auto entry = list->head; entry; entry = entry->next) {
+        print_fn(&entry->token);
+    }
+    tokenlist_free(list);
+    error_free(err);
+    return 0;
+
+cleanup_tokens:
+    tokenlist_free(list);
+cleanup_lexer:
+    lexer_close(lex);
+cleanup_error:
     puts(err->message);
     error_free(err);
     return 1;
 }
-
-    bool keep_going = true;
-    while (keep_going && (err = lexer_next(&lex, &token)) == nullptr) {
-        keep_going = print_fn(&token);
-        free(token.value);
-    }
-
-    if (err && err != err_eof) {
-        puts(err->message);
-    }
-    error_free(err);
-    return 0;
-}

src/tokenlist.c

@@ -0,0 +1,83 @@
+#include "tokenlist.h"
+#include "error.h"
+#include "lexer.h"
+#include <stdlib.h>
+
+error_t *tokenlist_alloc(tokenlist_t **output) {
+    *output = nullptr;
+
+    tokenlist_t *list = calloc(1, sizeof(tokenlist_t));
+    if (list == nullptr)
+        return err_allocation_failed;
+
+    list->head = nullptr;
+    list->tail = nullptr;
+
+    *output = list;
+    return nullptr;
+}
+
+error_t *tokenlist_entry_alloc(tokenlist_entry_t **output) {
+    *output = nullptr;
+
+    tokenlist_entry_t *entry = calloc(1, sizeof(tokenlist_entry_t));
+    if (entry == nullptr)
+        return err_allocation_failed;
+
+    entry->next = nullptr;
+    entry->prev = nullptr;
+
+    *output = entry;
+    return nullptr;
+}
+
+void tokenlist_append(tokenlist_t *list, tokenlist_entry_t *entry) {
+    if (list->head == nullptr) {
+        list->head = entry;
+        list->tail = entry;
+        entry->next = nullptr;
+        entry->prev = nullptr;
+    } else {
+        entry->prev = list->tail;
+        entry->next = nullptr;
+        list->tail->next = entry;
+        list->tail = entry;
+    }
+}
+
+void tokenlist_entry_free(tokenlist_entry_t *entry) {
+    lexer_token_cleanup(&entry->token);
+    free(entry);
+}
+
+void tokenlist_free(tokenlist_t *list) {
+    if (list == nullptr)
+        return;
+
+    tokenlist_entry_t *current = list->head;
+    while (current) {
+        tokenlist_entry_t *next = current->next;
+        tokenlist_entry_free(current);
+        current = next;
+    }
+
+    free(list);
+}
+
+error_t *tokenlist_fill(tokenlist_t *list, lexer_t *lex) {
+    error_t *err = nullptr;
+    lexer_token_t token = {};
+    while ((err = lexer_next(lex, &token)) == nullptr) {
+        tokenlist_entry_t *entry;
+        err = tokenlist_entry_alloc(&entry);
+        if (err) {
+            lexer_token_cleanup(&token);
+            return err;
+        }
+        entry->token = token;
+        tokenlist_append(list, entry);
+    }
+    if (err != err_eof)
+        return err;
+    return nullptr;
+}

src/tokenlist.h

@@ -0,0 +1,30 @@
+#ifndef INCLUDE_SRC_TOKENLIST_H_
+#define INCLUDE_SRC_TOKENLIST_H_
+#include "lexer.h"
+
+typedef struct tokenlist_entry tokenlist_entry_t;
+
+struct tokenlist_entry {
+    lexer_token_t token;
+    tokenlist_entry_t *next;
+    tokenlist_entry_t *prev;
+};
+
+typedef struct tokenlist {
+    tokenlist_entry_t *head;
+    tokenlist_entry_t *tail;
+} tokenlist_t;
+
+/**
+ * @brief Allocate a new doubly linked list of lexer tokens
+ */
+error_t *tokenlist_alloc(tokenlist_t **list);
+
+/**
+ * Consume all tokens from the lexer and add them to the list
+ */
+error_t *tokenlist_fill(tokenlist_t *list, lexer_t *lex);
+
+void tokenlist_free(tokenlist_t *list);
+
+#endif // INCLUDE_SRC_TOKENLIST_H_