Starting with the lexer, i'm starting to comprehend better what pratt parsing is and how to apply it. For now just declaring basic structure and functions

include/lexer.h

@@ -0,0 +1,78 @@
+#ifndef LEXER_H
+#define LEXER_H
+
+#include <stddef.h>
+#include <stdbool.h>
+#include <stdint.h>
+
+// For identifing
+typedef enum {
+    NODE_INTEGER,
+    NODE_BINARY_OP,
+} ASTNodeType;
+
+// For classify operators
+typedef enum {
+    OP_ADD,
+    OP_SUB,
+    OP_MUL,
+    OP_DIV
+} Operator;
+
+typedef enum {
+    ARRAY_OK = 0,
+    ARRAY_NULL,
+    ARRAY_EMPTY,
+    ARRAY_OUT_OF_BOUNDS,
+    ARRAY_NULL_ARG,
+    ARRAY_ALLOC,
+} ASTNodeArrayErr;
+
+typedef enum {
+    LEXER_OK = 0,
+    LEXER_INT_OVERFLOW,
+    LEXER_FAILED_NUMBER_CONVERSION,
+    LEXER_NOT_RECOGNIZED_SYMBOL,
+    LEXER_EMPTY_INPUT,
+    LEXER_NULL_ARG,
+    LEXER_WRONG_SYNTAX,
+    LEXER_BUF_OVERFLOW,
+} LexerErr;
+
+// Can be thought as tokens, they will be used by the parser.
+typedef struct ASTNode {
+    ASTNodeType type;
+    union {
+        int64_t integer;
+        struct {
+            struct ASTNode *left;
+            struct ASTNode *right;
+            Operator op;
+        } binary;
+    } data;
+} ASTNode;
+
+// I prefer ot have a dynamic array for storing the "tokens"
+typedef struct {
+    size_t len;
+    size_t cap;
+    ASTNode *data;
+} ASTNodeArray;
+
+// Basic array functionality
+ASTNodeArray ASTNodeArray_init(size_t size);
+void ASTNodeArray_free(ASTNodeArray *arr);
+ASTNodeArrayErr ASTNodeArray_push(ASTNodeArray *arr, ASTNode node);
+ASTNodeArrayErr ASTNodeArray_get(const ASTNodeArray *arr, size_t index, ASTNode *out);
+// Out in pop can be NULL so it doesn't return anything
+ASTNodeArrayErr ASTNodeArray_pop(ASTNodeArray *arr, size_t index, ASTNode *out);
+size_t ASTNodeArray_len(ASTNodeArray *arr);
+
+// Lexer funtions as well as few functionality
+LexerErr tokenize(const char* input, ASTNodeArray *out);
+LexerErr tokenize_number(const char* input, size_t *offset, ASTNode *out);
+LexerErr string_to_integer(const char buf[], int64_t *number);
+bool isoperator(int c);
+Operator char_to_operator(int c);
+
+#endif // !LEXER_H

include/parser.h

@@ -0,0 +1,14 @@
+#include "lexer.h"
+
+typedef struct {
+    ASTNode *head;
+} AST;
+
+typedef enum {
+    PARSER_NUD, // Null Denotation
+    PARSER_LED, // Left Denotation
+} ParserState;
+
+size_t node_lbp(Operator op);
+size_t node_rbp(Operator op);
+AST parse(ASTNodeArray arr);

src/lexer.c

@@ -0,0 +1,250 @@
+#include "lexer.h"
+#include <ctype.h>
+#include <math.h>
+#include <stdint.h>
+#include <stdlib.h>
+#include <strings.h>
+#include <limits.h>
+
+#define NODE_ARRAY_DEFAULT_SIZE 64
+// Helps  state machine for the lexer :)
+typedef enum {
+    WAIT_FOR_NUMBER,
+    WAIT_FOR_OPERATOR,
+} LexerState;
+
+ASTNodeArray ASTNodeArray_init(size_t size) {
+    ASTNodeArray new;
+    new.len = 0;      // if 0 then use default
+    new.cap = size == 0 ? NODE_ARRAY_DEFAULT_SIZE : size;
+    new.data = malloc(new.cap * sizeof(ASTNode));
+    return new;
+}
+
+void ASTNodeArray_free(ASTNodeArray *arr) {
+    free(arr->data);
+    arr->cap = 0;
+    arr->len = 0;
+}
+
+ASTNodeArrayErr ASTNodeArray_get(const ASTNodeArray *arr, size_t index, ASTNode *out) {
+    if (arr == NULL) {
+        return ARRAY_NULL;
+    }
+
+    if (out == NULL) {
+        return ARRAY_NULL_ARG;
+    }
+
+    if (arr->len == 0) {
+        return ARRAY_EMPTY;
+    }
+
+    if (index >= arr->len) {
+        return ARRAY_OUT_OF_BOUNDS;
+    }
+
+    *out = arr->data[index];
+
+    return ARRAY_OK;
+}
+
+ASTNodeArrayErr ASTNodeArray_push(ASTNodeArray *arr, ASTNode node) {
+    if (arr == NULL) {
+        return ARRAY_NULL;
+    }
+
+    if (arr->len >= arr->cap) {
+        size_t new_cap = arr->cap * 2;
+        ASTNode *tmp = realloc(arr->data, new_cap * sizeof(ASTNode));
+        if (tmp == NULL) {
+            return ARRAY_ALLOC;
+        }
+        arr->data = tmp;
+        arr->cap = new_cap;
+    }
+
+    arr->data[arr->len] = node;
+    arr->len = arr->len + 1;
+
+    return ARRAY_OK;
+}
+
+ASTNodeArrayErr ASTNodeArray_pop(ASTNodeArray *arr, size_t index, ASTNode *out) {
+    if (arr == NULL) {
+        return ARRAY_NULL;
+    }
+
+    if (arr->len == 0) {
+        return ARRAY_EMPTY;
+    }
+
+    if (index >= arr->len) {
+        return ARRAY_OUT_OF_BOUNDS;
+    }
+
+    if (arr->cap / 4 > arr->len) {
+        size_t new_cap = arr->cap / 2;
+        ASTNode *tmp = realloc(arr->data, new_cap * sizeof(ASTNode));
+        if (tmp == NULL) {
+            return ARRAY_ALLOC;
+        }
+        arr->data = tmp;
+        arr->cap = new_cap;
+    }
+    
+    if (out != NULL) {
+        ASTNode node_to_delete = arr->data[index];
+        *out = node_to_delete;
+    }
+
+    for (size_t i = index; i < arr->len - 1; i++) {
+        arr->data[index] = arr->data[index + 1];
+    }
+
+    return ARRAY_OK;
+}
+
+size_t ASTNodeArray_len(ASTNodeArray *arr) {
+    if (arr == NULL) {
+        return 0;
+    }
+    return arr->len;
+}
+
+LexerErr tokenize(const char *input, ASTNodeArray *out) {
+    size_t offset = 0;
+    LexerState state = WAIT_FOR_NUMBER;
+    ASTNodeArray arr = ASTNodeArray_init(0); // 0 defaults to 64
+
+    while (input[offset] != '\n' && input[offset] != '\0') {
+        int current = input[offset];
+
+        if (isdigit(current)) {
+            if (state != WAIT_FOR_NUMBER) {
+                ASTNodeArray_free(&arr);
+                return LEXER_WRONG_SYNTAX;
+            }
+            ASTNode new_node;
+            LexerErr result = tokenize_number(input, &offset, &new_node);
+
+            if (result != LEXER_OK) {
+                ASTNodeArray_free(&arr);
+                return result;
+            }
+
+            ASTNodeArray_push(&arr, new_node);
+            state = WAIT_FOR_OPERATOR;
+        } else if (isoperator(current)) {
+            if (state != WAIT_FOR_OPERATOR) {
+                return LEXER_WRONG_SYNTAX;
+            }
+            ASTNode new_node = {
+                .type = NODE_BINARY_OP,
+                .data.binary.op = char_to_operator(current),
+                .data.binary.right = NULL,
+                .data.binary.left = NULL,
+            };
+
+            ASTNodeArray_push(&arr, new_node);
+            state = WAIT_FOR_NUMBER;
+        } else if (isspace(current)) {
+            // Nothing...
+        } else {
+            ASTNodeArray_free(&arr);
+            return LEXER_NOT_RECOGNIZED_SYMBOL;
+        }
+
+        offset++;
+    }
+
+    if (arr.len < 1) {
+        return LEXER_EMPTY_INPUT;
+    }
+
+    *out = arr;
+    return LEXER_OK;
+}
+
+// CURRENTLY, it only supports ints, not clear how floating
+// point is implemented but i'll figure it out
+LexerErr tokenize_number(const char *input, size_t *offset, ASTNode *out) {
+    char buf[128] = { '\0' };
+    size_t buf_pos = 0;
+    bool is_integer = true; // Will later be used to differentiate fractions
+
+    size_t current = *offset;
+    while (isdigit(input[current])) {
+        buf[buf_pos] = input[current];
+        
+        if (buf_pos >= sizeof(buf)) {
+            return LEXER_BUF_OVERFLOW;
+        }
+        current++;
+        buf_pos++;
+    }
+
+    ASTNode new_node;
+    if (is_integer) {
+        new_node.type = NODE_INTEGER;
+        LexerErr status = string_to_integer(buf, &new_node.data.integer);
+        if (status == LEXER_OK) {
+            *out = new_node;
+        }
+        *offset = current;
+        return status;
+    }
+
+    return LEXER_FAILED_NUMBER_CONVERSION;
+}
+
+LexerErr string_to_integer(const char *buf, int64_t *number) {
+    int c = 0;
+    int64_t count = 0;
+    while (buf[c] != '\0') {
+        
+        int digit = buf[c] - '0';
+
+        if (count > (INT64_MAX - digit) / 10) {
+            return LEXER_INT_OVERFLOW;
+        }
+        count = count * 10;
+        count += digit;
+        
+        c++;
+    }
+
+    *number = count;
+    return LEXER_OK;
+}
+
+bool isoperator(int c) {
+    switch (c) {
+        case '+':
+        case '-':
+        case '/':
+        case '*':
+            return true;
+        default:
+            return false;
+    }
+}
+
+Operator char_to_operator(int c) {
+    switch (c) {
+        case '+':
+            return OP_ADD;
+            break;
+        case '-':
+            return OP_SUB;
+            break;
+        case '*':
+            return OP_MUL;
+            break;
+        case '/':
+            return OP_DIV;
+            break;
+        default: // I mean shouldn't be used, we assume
+            return -1;
+    }
+}

src/parser.c

@@ -0,0 +1,27 @@
+#include "parser.h"
+#include "lexer.h"
+#include <stdlib.h>
+
+size_t node_lbp(Operator op) {
+    switch (op) {
+        case OP_ADD:
+        case OP_SUB:
+            return 10;
+            break;
+        case OP_DIV:
+        case OP_MUL:
+            return 20;
+    }
+}
+
+size_t node_rbp(Operator op) {
+    switch (op) {
+        case OP_ADD:
+        case OP_SUB:
+            return 10;
+            break;
+        case OP_DIV:
+        case OP_MUL:
+            return 20;
+    }
+}

test/CMakeLists.txt

@@ -1,10 +1,17 @@
 find_package(cmocka REQUIRED)
 
-add_executable(test_parser test_parser.c)
+add_executable(test_nodeArray test_ASTNodeArray.c)
+add_executable(test_lexer test_lexer.c)
 
-target_link_libraries(test_parser
+target_link_libraries(test_nodeArray
     calculator_lib
     cmocka::cmocka
 )
 
-add_test(NAME parser_tests COMMAND test_parser)
+target_link_libraries(test_lexer
+    calculator_lib
+    cmocka::cmocka
+)
+
+add_test(NAME nodeArray_tests COMMAND test_nodeArray)
+add_test(NAME lexer_tests COMMAND test_lexer)

test/test_ASTNodeArray.c

@@ -0,0 +1,86 @@
+#include "lexer.h"
+#include <stdarg.h>
+#include <stdbool.h>
+#include <stddef.h>
+#include <stdint.h>
+#include <setjmp.h>
+#include <cmocka.h>
+
+static void test_array_push(void **state) {
+    (void) state;
+
+    // We use 2 to force resize and checking anything wrong with malloc
+    ASTNodeArray arr = ASTNodeArray_init(2);
+    ASTNode node1 = {
+        .type = NODE_INTEGER,
+        .data = { .integer = 90 }
+    };
+
+    ASTNode node2 = {
+        .type = NODE_INTEGER,
+        .data = { .integer = 80 }
+    };
+
+    ASTNode node3 = {
+        .type = NODE_INTEGER,
+        .data = { .integer = 70 }
+    };
+
+    assert_int_equal(ASTNodeArray_push(&arr, node1), ARRAY_OK);
+    assert_int_equal(ASTNodeArray_len(&arr), 1);
+
+    assert_int_equal(ASTNodeArray_push(&arr, node2), ARRAY_OK);
+    assert_int_equal(ASTNodeArray_len(&arr), 2);
+
+    assert_int_equal(ASTNodeArray_push(&arr, node3), ARRAY_OK);
+    assert_int_equal(ASTNodeArray_len(&arr), 3);
+
+    ASTNodeArray_free(&arr);
+}
+
+static void test_array_pop(void **state) {
+    (void) state;
+
+    // Set to force desize
+    ASTNodeArray arr = ASTNodeArray_init(16);
+    ASTNode node1 = {
+        .type = NODE_INTEGER,
+        .data = { .integer = 90 }
+    };
+
+    ASTNode node2 = {
+        .type = NODE_INTEGER,
+        .data = { .integer = 80 }
+    };
+
+    ASTNode node3 = {
+        .type = NODE_INTEGER,
+        .data = { .integer = 70 }
+    };
+
+    assert_int_equal(ASTNodeArray_push(&arr, node1), ARRAY_OK);
+    assert_int_equal(ASTNodeArray_len(&arr), 1);
+
+    assert_int_equal(ASTNodeArray_push(&arr, node2), ARRAY_OK);
+    assert_int_equal(ASTNodeArray_len(&arr), 2);
+
+    assert_int_equal(ASTNodeArray_push(&arr, node3), ARRAY_OK);
+    assert_int_equal(ASTNodeArray_len(&arr), 3);
+
+    ASTNode node4;
+    assert_int_equal(ASTNodeArray_pop(&arr, 1, &node4), ARRAY_OK);
+    assert_int_equal(node4.type, NODE_INTEGER);
+    assert_int_equal(node4.data.integer, 80);
+
+    ASTNodeArray_free(&arr);
+}
+
+
+int main(void) {
+    const struct CMUnitTest tests[] = {
+        cmocka_unit_test(test_array_push),
+        cmocka_unit_test(test_array_pop),
+    };
+
+    return cmocka_run_group_tests(tests, NULL, NULL);
+}

test/test_lexer.c

@@ -0,0 +1,112 @@
+#include "lexer.h"
+#include <stdarg.h>
+#include <stdbool.h>
+#include <stddef.h>
+#include <stdint.h>
+#include <setjmp.h>
+#include <cmocka.h>
+
+static void test_tokenize_normal_expresion(void **state) {
+    (void) state;
+
+    char expr[256] = "2 + 3 / 66 * 789";
+    ASTNodeArray tokens;
+    ASTNode node;
+    
+    assert_int_equal(tokenize(expr, &tokens), LEXER_OK);
+    assert_int_equal(tokens.len, 7);
+
+    ASTNodeArray_get(&tokens, 0, &node);
+    assert_int_equal(node.type, NODE_INTEGER);
+    assert_int_equal(node.data.integer, 2);
+
+    ASTNodeArray_get(&tokens, 1, &node);
+    assert_int_equal(node.type, NODE_BINARY_OP);
+    assert_int_equal(node.data.binary.op, OP_ADD);
+
+    ASTNodeArray_get(&tokens, 2, &node);
+    assert_int_equal(node.type, NODE_INTEGER);
+    assert_int_equal(node.data.integer, 3);
+
+    ASTNodeArray_get(&tokens, 3, &node);
+    assert_int_equal(node.type, NODE_BINARY_OP);
+    assert_int_equal(node.data.binary.op, OP_DIV);
+
+    ASTNodeArray_get(&tokens, 4, &node);
+    assert_int_equal(node.type, NODE_INTEGER);
+    assert_int_equal(node.data.integer, 66);
+
+    ASTNodeArray_get(&tokens, 5, &node);
+    assert_int_equal(node.type, NODE_BINARY_OP);
+    assert_int_equal(node.data.binary.op, OP_MUL);
+
+    ASTNodeArray_get(&tokens, 6, &node);
+    assert_int_equal(node.type, NODE_INTEGER);
+    assert_int_equal(node.data.integer, 789);
+}
+
+static void test_tokenize_unrecognized_symbol(void **state) {
+    (void) state;
+
+    char expr[256] = " 2 j 3 / 66 } 789";
+    ASTNodeArray tokens = {
+        .len = 0,
+        .cap = 0,
+    };
+
+    assert_int_equal(tokenize(expr, &tokens), LEXER_NOT_RECOGNIZED_SYMBOL);
+    assert_int_equal(tokens.len, 0);
+    assert_int_equal(tokens.cap, 0);
+}
+
+static void test_tokenize_wrong_sintax(void **state) {
+    (void) state;
+
+    char expr[256] = "2 3 / 66 789";
+    ASTNodeArray tokens = {
+        .len = 0,
+        .cap = 0,
+    };
+
+    assert_int_equal(tokenize(expr, &tokens), LEXER_WRONG_SYNTAX);
+    assert_int_equal(tokens.len, 0);
+    assert_int_equal(tokens.cap, 0);
+}
+
+static void test_string_to_number_normal(void **state) {
+    (void) state;
+
+    char num[16] = "2333t55";
+    size_t offset = 0;
+    ASTNode result;
+
+    assert_int_equal(tokenize_number(num, &offset, &result), LEXER_OK);
+
+    assert_int_equal(offset, 4); // equal to t position in string
+    assert_int_equal(result.type, NODE_INTEGER);
+    assert_int_equal(result.data.integer, 2333);
+}
+
+static void test_string_to_number_overflow(void **state) {
+    (void) state;
+
+    // Number is INT64_MAX but with a extra 8 at the end
+    char num[32] = "92233720368547758078yy7";
+    size_t offset = 0;
+    ASTNode result;
+    assert_int_equal(tokenize_number(num, &offset, &result), LEXER_INT_OVERFLOW);
+    // Technically it can trigger a buf overflow error but obvioulsy 
+    // it will trigger int overflow error first
+}
+
+int main(void) {
+    const struct CMUnitTest tests[] = {
+        cmocka_unit_test(test_string_to_number_normal),
+        cmocka_unit_test(test_string_to_number_overflow),
+        cmocka_unit_test(test_tokenize_normal_expresion),
+        cmocka_unit_test(test_tokenize_unrecognized_symbol),
+        cmocka_unit_test(test_tokenize_wrong_sintax),
+    };
+
+    return cmocka_run_group_tests(tests, NULL, NULL);
+}