sic

commit d24e1b588e62bef0a64aafbf34f720a0bf6e3867
parent d8295a48cf792eafc3d1fb12c80b2946bfa66427
Author: citbl <citbl@citbl.org>
Date:   Wed, 13 May 2026 22:25:26 +1000

refac and ident lexing

Diffstat:
M
src/array.h
 | 
1
+
A
src/check.c
 | 
10
++++++++++
A
src/check.h
 | 
5
+++++
M
src/common.h
 | 
68
++------------------------------------------------------------------
M
src/lexer.c
 | 
23
++++++++++++++++++++---
M
src/str.c
 | 
1
-
A
src/token.h
 | 
63
+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
M
src/utils.c
 | 
7
-------

8 files changed, 101 insertions(+), 77 deletions(-)
diff --git a/src/array.h b/src/array.h
@@ -1,5 +1,6 @@
 #pragma once
 
+#include <stdlib.h>
 #include <stdio.h>
 
 #define ARRAY_PUSH(dest, len, cap, item)                                                \
diff --git a/src/check.c b/src/check.c
@@ -0,0 +1,10 @@
+#include "check.h"
+#include <stdio.h>
+#include <stdlib.h>
+
+void check(bool condition, const char* message)
+{
+    if (!condition) return;
+    fprintf(stderr, "%s", message);
+    exit(1);
+}
diff --git a/src/check.h b/src/check.h
@@ -0,0 +1,5 @@
+#pragma once
+
+#include <stdbool.h>
+
+void check(bool condition, const char* message);
diff --git a/src/common.h b/src/common.h
@@ -5,74 +5,10 @@
 #include <stdio.h>
 #include <stdlib.h>
 
-#include "str.h"
-
-typedef enum Token_Type {
-    NOTYETSET = 9,
-    IDENT,
-    KEYWORD,
-    SYMBOL,
-    PLUS,
-    MIN,
-    STAR,
-    SLASH,
-    BACKSLASH,
-    PERCENT,
-    LBRACE,
-    RBRACE,
-    LPAREN,
-    RPAREN,
-    EQ_EQ,
-    BANG_EQ,
-    EQ_GT,
-    LT_EQ,
-    LT_DASH,
-    DASH_GT,
-    EQ,
-    BANG,
-    LIT_STRING,
-    LIT_DECIMAL,
-    LIT_INT,
-    LIT_BOOL,
-    LIT_VOID,
-} Token_Type;
-
-typedef enum Keyword {
-    IF = 139,
-    ELSE,
-    WHILE,
-    OPT,
-    LAZY,
-    MATCH,
-    DEF,
-    FOR,
-    EACH,
-    SOME,
-    NONE,
-    OK,
-    ERR,
-    IN,
-    IS,
-    CAST
-} Keyword;
+#include "token.h"
 
 /////////////////////////////////////////////////
 
-typedef struct String {
-    char* value;
-    size_t cap;
-    size_t len;
-} String;
-
-typedef struct Token {
-    Token_Type type;
-    Str lexeme;
-    const char* path;
-    const char* filename;
-    size_t line;
-    size_t col;
-} Token;
-
 typedef struct Lexer_State {
     size_t pos;
     size_t line;
@@ -92,4 +28,4 @@ typedef struct Lexer {
 
 /////////////////////////////////////////////////
 
-void check(bool condition, const char* message);
+// void check(bool condition, const char* message);
diff --git a/src/lexer.c b/src/lexer.c
@@ -1,9 +1,10 @@
-#include <stdio.h>
+
 #include <string.h>
 #include <ctype.h>
 
 #include "lexer.h"
 #include "array.h"
+#include "str.h"
 
 static void add_token(Lexer* lex, Token t)
 {
@@ -104,6 +105,21 @@ static void lex_number(Lexer* lex, Token* tok)
     add_token(lex, *tok);
 }
 
+static void lex_ident(Lexer* lex, Token* tok)
+{
+    char c = lex->code[lex->state.pos];
+    Str* str = &tok->lexeme;
+    str_append(str, c);
+    tok->type = IDENT;
+    while (lex->state.pos < lex->code_len) {
+        advance(lex);
+        c = peek(lex);
+        if (!isalnum((unsigned char)c)) break;
+        str_append(str, c);
+    }
+    add_token(lex, *tok);
+}
+
 static Token new_token(Lexer* lex)
 {
     return (Token){
@@ -135,9 +151,10 @@ Lexer* lexer_lex(Lexer* lex)
             run_until_char(lex, '\n');
             continue;
         }
-
+        if (isalpha((unsigned char)c)) {
+            lex_ident(lex, &t);
+        }
         if (isdigit((unsigned char)c)) {
-            t.type = LIT_INT;
             lex_number(lex, &t);
         }
 
diff --git a/src/str.c b/src/str.c
@@ -1,4 +1,3 @@
-#include <stdlib.h>
 #include "str.h"
 #include "array.h"
 
diff --git a/src/token.h b/src/token.h
@@ -0,0 +1,63 @@
+#pragma once
+#include "str.h"
+
+typedef enum Token_Type {
+    NOTYETSET = 9,
+    IDENT,
+    KEYWORD,
+    SYMBOL,
+    PLUS,
+    MIN,
+    STAR,
+    SLASH,
+    BACKSLASH,
+    PERCENT,
+    LBRACE,
+    RBRACE,
+    LPAREN,
+    RPAREN,
+    EQ_EQ,
+    BANG_EQ,
+    EQ_GT,
+    LT_EQ,
+    LT_DASH,
+    DASH_GT,
+    EQ,
+    BANG,
+    LIT_STRING,
+    LIT_DECIMAL,
+    LIT_INT,
+    LIT_BOOL,
+    LIT_VOID,
+} Token_Type;
+
+typedef enum Keyword {
+    IF = 139,
+    ELSE,
+    WHILE,
+    OPT,
+    LAZY,
+    MATCH,
+    DEF,
+    FOR,
+    EACH,
+    SOME,
+    NONE,
+    OK,
+    ERR,
+    IN,
+    IS,
+    CAST
+} Keyword;
+
+/////////////////////////////////////////////////
+
+typedef struct Token {
+    Token_Type type;
+    Str lexeme;
+    const char* path;
+    const char* filename;
+    size_t line;
+    size_t col;
+} Token;
+
diff --git a/src/utils.c b/src/utils.c
@@ -56,10 +56,3 @@ void separate_file_from_path(const char* fullpath, char** out_path, char** out_f
     *out_filename = strdup(filename);
     free(path);
 }
-
-void check(bool condition, const char* message)
-{
-    if (!condition) return;
-    fprintf(stderr, "%s", message);
-    exit(1);
-}