From f6060fb14216676533a90a5f7db6021e550d7147 Mon Sep 17 00:00:00 2001
From: MCorange <mcorange@mcorangehq.xyz>
Date: Wed, 25 Jun 2025 22:09:56 +0300
Subject: [PATCH] Add KW's and OP's

---
 src/include/token.h |  4 +---
 src/main.c          | 29 +++++++++++++++++++++---
 src/token.c         |  4 ++++
 src/tokeniser.c     | 55 ++++++++++++++++++++++++++++++++++++++++-----
 test.mrph           |  2 +-
 5 files changed, 81 insertions(+), 13 deletions(-)

diff --git a/src/include/token.h b/src/include/token.h
index 59523a3..8e1c856 100644
--- a/src/include/token.h
+++ b/src/include/token.h
@@ -5,7 +5,6 @@
 #include <stddef.h>
 #include <sys/types.h>
 
-
 typedef enum token_type_e {
     TT_NONE = 0,
     TT_KW,
@@ -16,6 +15,7 @@ typedef enum token_type_e {
     TT_PUSH_CHAR,
     TT_PUSH_INT,
     TT_PUSH_FLOAT,
+    TT_IDENT
 } token_type_t;
 
 typedef enum kw_type_e {
@@ -117,8 +117,6 @@ typedef struct token_s {
     };
 } token_t;
 
-
-
 extern const char* OP_LIST[];
 extern const char* KW_LIST[];
 #endif // _H_MORPH_TOKEN
diff --git a/src/main.c b/src/main.c
index 7032284..7beaa55 100644
--- a/src/main.c
+++ b/src/main.c
@@ -1,13 +1,36 @@
+#include "dynarray.h"
+#include "token.h"
 #include <util.h>
 #include <tokeniser.h>
 #include <stdio.h>
 
-int main(int argc, const char **argv) {
+int main(int argc, const char** argv) {
     if (argc < 2) {
         printf("Usage: %s [source]\n", argv[0]);
         return 1;
     }
     char* str = read_to_string(argv[1]);
-    tokenise_string(argv[1], str);
-    return 0; 
+    token_t* tokens = tokenise_string((char*)argv[1], str);
+    int token_count = dynarray_length(tokens);
+
+    for (int i = 0; i < token_count; i++) {
+        token_t tok = tokens[i];
+        switch (tok.type) {
+        case (TT_NONE):
+            break;
+        case (TT_KW): {
+            printf("TOK: KW %s\n", KW_LIST[tok.kw_type]);
+        } break;
+        case (TT_OP): {
+            printf("TOK: OP %s\n", OP_LIST[tok.op_type]);
+        } break;
+        case (TT_IDENT): {
+            printf("TOK: IDENT \"%s\"\n", tok.str_v);
+        } break;
+        default:
+            break;
+        }
+    }
+
+    return 0;
 }
diff --git a/src/token.c b/src/token.c
index cb98f31..4f38069 100644
--- a/src/token.c
+++ b/src/token.c
@@ -1,5 +1,7 @@
 #include <token.h>
 
+// clang-format off
+
 const char* OP_LIST[] = {
     [OP_ADD]  = "add",
     [OP_SUB]  = "sub",
@@ -59,3 +61,5 @@ const char* KW_LIST[] = {
     [KW_CONST] = "const",
     [KW_MEMORY] = "memory"
 };
+
+// clang-format on
diff --git a/src/tokeniser.c b/src/tokeniser.c
index ee318c5..3131f61 100644
--- a/src/tokeniser.c
+++ b/src/tokeniser.c
@@ -1,3 +1,4 @@
+#include <stdio.h>
 #include <stdlib.h>
 #include <string.h>
 #include <token.h>
@@ -28,16 +29,58 @@ token_t* tokenise_string(char* file_path, char* str) {
             // TODO: manage memory better
             // INFO: If you make an ident longer than 4kb i will murder you
             char* buf = malloc(1024 * 4);
-            int buf_counter = 1;
-            *buf = str[i];
-            while (i < str_len && (str[i + 1] != ' ' || str[i + 1] != '\t' || str[i + 1] != '\n' || str[i + 1] != '\r')) {
-                i += 1;
+            memset(buf, 0, 1024 * 4);
+            int buf_counter = 0;
+            while (i < str_len) {
+                if (str[i] == ' ' || str[i] == '\t' || str[i] == '\r' || str[i] == '\n') {
+                    break;
+                }
                 buf[buf_counter++] = str[i];
+                i += 1;
             }
+            buf[buf_counter + 1] = '\0';
+            printf("'%s'\n", buf);
+            // PERF: I dont know if this makes it faster or slower, need 2 check
             buf = realloc(buf, strlen(buf) + 1);
-            
+
+            bool found = false;
+            for (int i = 1; i < KW_COUNT__; i++) {
+                if (strcmp(KW_LIST[i], buf) == 0) {
+                    token_t tok = {
+                        .type = TT_KW,
+                        .kw_type = (kw_type_t)i,
+                    };
+                    dynarray_push(tokens, tok);
+                    found = true;
+                }
+            }
+
+            if (found) {
+                break;
+            }
+            for (int i = 1; i < OP_COUNT__; i++) {
+                if (strcmp(OP_LIST[i], buf) == 0) {
+                    token_t tok = {
+                        .type = TT_OP,
+                        .op_type = (op_type_t)i,
+                    };
+                    dynarray_push(tokens, tok);
+                    found = true;
+                }
+            }
+
+            if (found) {
+                break;
+            }
+            token_t tok = {
+                .type = TT_IDENT,
+                .str_v = strdup(buf),
+            };
+            dynarray_push(tokens, tok);
+
+            free(buf);
         }
         }
     }
-    return NULL;
+    return tokens;
 }
diff --git a/test.mrph b/test.mrph
index f0b6a76..45242ec 100644
--- a/test.mrph
+++ b/test.mrph
@@ -1,5 +1,5 @@
 
 
-fn main with void returs int do
+fn main with void returns int do
     34 35 add __print__
 end