From 94b151462a4ae16826ae2a157efbd287885dc563 Mon Sep 17 00:00:00 2001
From: anon <anon@anon.anon>
Date: Thu, 24 Aug 2023 04:11:46 +0200
Subject: [PATCH] auto esc magic char tokens

---
 source/hl.h    | 31 +++++++++++++++++--------------
 source/regex.c | 20 ++++++++++++++++----
 source/regex.h |  2 ++
 3 files changed, 35 insertions(+), 18 deletions(-)

diff --git a/source/hl.h b/source/hl.h
index 22ae2a7..7528277 100644
--- a/source/hl.h
+++ b/source/hl.h
@@ -34,7 +34,7 @@ typedef struct {
 typedef struct {
 	hl_group_t   * hl;
 	token_type_t   t;
-	char         * syntax;
+	regex_t      * syntax;
 } token_t;
 
 // GLOBALS
@@ -56,7 +56,7 @@ void new_display_mode(display_t * mode) {
 
 int free_token(token_t * token) {
 	free(token->hl);
-	free(token->syntax);
+	regex_free(token->syntax);
 
 	return 0;
 }
@@ -67,15 +67,23 @@ int append_token(token_t * token) {
 	return 0;
 }
 
-token_t * new_symbol_token(const char         * const word,
-                                 hl_group_t   * const    g) {
-	char * new_word = strdup(word);
+token_t * new_symbol_token(const char         * const c,
+                                 hl_group_t   * const g) {
+	char * new_word;
+	if (is_magic(*c)) {
+		new_word = (char *)malloc(sizeof(char)*3);
+		new_word[0] = '\\';
+		new_word[1] = *c;
+		new_word[2] = '\00';
+	} else {
+		new_word = strdup(c);
+	}
 
 	token_t * mt = (token_t*)malloc(sizeof(token_t));
 
 	mt->hl     = g;
 	mt->t      = KEYSYMBOL;
-	mt->syntax = new_word;
+	mt->syntax = regex_compile(new_word);
 
 	append_token(mt);
 
@@ -126,7 +134,7 @@ token_t * new_keyword_token(const char         * const word,
 
 	mt->hl     = g;
 	mt->t      = KEYWORD;
-	mt->syntax = new_word;
+	mt->syntax = regex_compile(new_word);
 
 	append_token(mt);
 
@@ -173,15 +181,10 @@ token_t * new_token(const char         * const word,
 int token_fits(const token_t * const       token,
                const char    * const       to,
                const int                   string_offset,
-			             int * match_offset) {
-	const char * const pattern = token->syntax;
-
-	if (! pattern) {
-		return true;
-	}
+			         int     *             match_offset) {
 
 	//return regex_match(pattern, to, string_offset, match_offset);
-	return regex_search(NULL, pattern + string_offset);
+	return regex_search(token->syntax, to + string_offset);
 }
 
 void render_string(const char * const string,
diff --git a/source/regex.c b/source/regex.c
index cef41ad..a662b8f 100644
--- a/source/regex.c
+++ b/source/regex.c
@@ -33,7 +33,7 @@ typedef struct {
 	s += n;         \
 } while (0)
 
-bool is_quantifier(const char c){
+static bool is_quantifier(const char c) {
 	for (const char * s = "+*?"; *s != '\00'; s++) {
 		if (*s == c) {
 			return true;
@@ -42,8 +42,20 @@ bool is_quantifier(const char c){
 	return false;
 }
 
+bool is_magic(const char c) {
+	if (is_quantifier(c)) {
+		return true;
+	}
+	for (const char * s = "\\[]"; *s != '\00'; s++) {
+		if (*s == c) {
+			return true;
+		}
+	}
+	return false;
+}
 
-int escape_1_to_1(const char c, char * whitelist) {
+
+static int escape_1_to_1(const char c, char * whitelist) {
 	switch(c) {
 		case 't': {
 			strcat(whitelist, "\t");
@@ -83,7 +95,7 @@ int escape_1_to_1(const char c, char * whitelist) {
 	return 0;
 }
 
-int escape_1_to_N(const char c, char * whitelist) {
+static int escape_1_to_N(const char c, char * whitelist) {
 	switch(c) {
 		case 'i': {
 			const char identifier_chars[] = "@0123456789_\300\301\302\303\304\305\306\307\310\311\312\313\314\315\316\317\320\321\322\323\324\325\326\327\330\331\332\333\334\335\336\337";
@@ -175,7 +187,7 @@ int escape_1_to_N(const char c, char * whitelist) {
 	return 0;
 }
 
-int compile_range(const char * const     range,
+static int compile_range(const char * const     range,
                         char *       whitelist) {
 	assert(range[0] == '[' && "Not a range.");
 
diff --git a/source/regex.h b/source/regex.h
index c6c714f..7b1ef10 100644
--- a/source/regex.h
+++ b/source/regex.h
@@ -17,4 +17,6 @@ extern regex_t * regex_compile(const char * const pattern);
 extern bool regex_search(regex_t * regex, const char * const string);
 extern int       regex_free(regex_t * const regex);
 
+extern bool is_magic(const char c);
+
 #endif