hlc

commit 75b1868517b633d8dec332fefc4ae697539687ba
parent 0e31cba13af2a9fbb5111e36dfbc5748d6a4add2
Author: thing1 <l.standen@posteo.com>
Date:   Sun,  5 Oct 2025 22:58:38 +0100

Started lexer

Diffstat:
M
Makefile
 | 
14
+++++++++++++-
M
lex.c
 | 
135
+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++------------------
A
lex.h
 | 
43
+++++++++++++++++++++++++++++++++++++++++++
A
main.c
 | 
15
+++++++++++++++
A
test.hlc
 | 
1
+

5 files changed, 177 insertions(+), 31 deletions(-)
diff --git a/Makefile b/Makefile
@@ -1,7 +1,19 @@
-all: spec
+CC=c89
+CFLAGS=-ggdb
+
+SRC = lex.c main.c
+OBJ = ${SRC:.c=.o}
+
+all: spec hlc
 
 spec: spec.md
 	smu spec.md > spec.html
 
+.c.o:
+	${CC} -c ${CFLAGS} $<
+
+hlc: ${OBJ}
+	${CC} -o $@ ${OBJ} ${CFLAGS}
+
 clean:
 	rm -rf *.html *.o hlc
diff --git a/lex.c b/lex.c
@@ -1,33 +1,108 @@
-#include <stdlib.h>
 #include <stdio.h>
+#include <stdlib.h>
 #include <string.h>
+#include <ctype.h>
+
+#include "lex.h"
+
+#define strnul(s) (s + strlen(s))
+
+lex_val lex_error = {UNKNOWN, NULL};
+lex_val lv = { 0 };
+
+FILE *input;
+lex_val *(*nextfn)(void);
+
+int
+getchr(FILE *in) {
+	int c;
+	while (isblank((c = getc(in)))) continue;
+	return c;
+}
+
+char *
+check_bytes(char *str) {
+	char *s;
+	for (s = str; *s; s++) if (getchr(input) != *s) return NULL;
+	return str;
+}
+
+int
+peekc(FILE *in) {
+	int c = getchr(in);
+	ungetc(c, in);
+	return c;
+}
+
+lex_val *
+lex_type(void) {
+	switch (peekc(input)) {
+	case 'b': 
+		if (!(lv.data = check_bytes("byte"))) return &lex_error;
+		lv.type = BYTE;
+		break;
+	case 's':
+		if (!(lv.data = check_bytes("short"))) return &lex_error;
+		lv.type = SHORT;
+		break;
+	case 'l':
+		if (!(lv.data = check_bytes("long"))) return &lex_error;
+		lv.type = LONG;
+		break;
+	case '*':
+		lv.type = PTR;
+		lv.data = check_bytes("*");
+		break;
+	default: return &lex_error;
+	}
+	if (peekc(input) == '*') nextfn = &lex_type;
+	else nextfn = &lex_name;
+	return &lv;
+}
+
+lex_val *
+lex_name(void) {
+	static char name[32] = { 0 };
+	char c, len = 0;
+	if (!isalpha((c = getchr(input)))) return &lex_error;
+	do {
+		memcpy(strnul(name), &c, 1);
+		if ((len++ + 1) == 32) return &lex_error;
+	} while(isalnum((c = getc(input))));
+	lv.data = name;
+	lv.type = NAME;
+
+	/*
+	switch (peekc(input)) {
+	case '=': nextfn = &lex_assign; break; CHECK FOR == and =
+	case ',': nextfn = &lex_comma; break;
+	case ';': nextfn = &lex_semi; break;
+
+	case '<':
+	case '>':
+	case '!':
+	case '+': 
+	case '-':
+	case '/': 
+	case '*': CHECK FOR MUL AND PTR
+	case '&':
+		  nextfn = &lex_sym; break;
+		  break;
+	default: return &lex_error;
+	}
+	*/
+
+	return &lv;
+}
+
+void
+init_lexer(FILE *in) {
+	input = in;
+	nextfn = &lex_type;
+}
 
-enum lex_type {
-	STRING,
-	NAME,
-	NUM,
-	SEMI,
-	DEREF,
-	UNARRAY,
-	QUOTE,
-	OBRACE,
-	CBRACE,
-	ASSIGN,
-	EQU,
-	NEQ,
-	LT,
-	GT,
-	LTE,
-	GTE,
-	FUNC,
-	COMMA,
-	ADD,
-	SUB,
-	DIV,
-	MUL,
-	BYTE,
-	SHORT,
-	LONG,
-	OCBRACE,
-	CCBRACE,
-};
+lex_val *
+get_next() {
+	memset(&lv, 0, sizeof(typeof(lv)));
+	return (nextfn) ? nextfn() : NULL;
+}
diff --git a/lex.h b/lex.h
@@ -0,0 +1,43 @@
+
+enum lex_type {
+	UNKNOWN,
+
+	STRING,
+	NAME,
+	NUM,
+	SEMI,
+	PTR,
+	UNARRAY,
+	QUOTE,
+	OBRACE,
+	CBRACE,
+	ASSIGN,
+	EQU,
+	NEQ,
+	LT,
+	GT,
+	LTE,
+	GTE,
+	FUNC,
+	COMMA,
+	ADD,
+	SUB,
+	DIV,
+	MUL,
+	BYTE,
+	SHORT,
+	LONG,
+	OCBRACE,
+	CCBRACE,
+};
+
+typedef struct lex_val {
+	enum lex_type type;	
+	char *data;
+} lex_val;
+
+lex_val *lex_name(void);
+lex_val *lex_type(void);
+
+lex_val *get_next(void);
+void init_lexer(FILE *f);
diff --git a/main.c b/main.c
@@ -0,0 +1,15 @@
+#include <stdio.h>
+
+#include "lex.h"
+
+int
+main() {
+	FILE *f = fopen("test.hlc", "r");
+	init_lexer(f);
+
+	lex_val *val;
+	while (val = get_next()) 
+		continue;
+
+	fclose(f);
+}
diff --git a/test.hlc b/test.hlc
@@ -0,0 +1 @@
+byte *name