bugs go on and on, and they dont end

Makefile

@@ -1,4 +1,8 @@
 CXXFLAGS := -fuse-ld=mold -ggdb -Wall -Wextra -Wpedantic
+ifeq (${DEBUG}, 1)
+	CXXFLAGS += -DDEBUG
+endif
+
 OUT := regtest
 
 main:

documentation/README.md

@@ -98,5 +98,13 @@ SOW must match:
 Not only that, this combination is key,
 either it has to be the start of the string
 or there has to be at least something which is not a symbol char.
-With out the last condition "eexample" would match "\\\<exaplme\\\>"
+With out the last condition "eexample" would match "\\\<exaple\\\>"
 as the iteration of `regex_match()` reaches "example".
+
+From a more practical perspective:
+``` C
+\<myword\>
+// Must match
+"myword"
+" myword"
+```

documentation/TODO.md

@@ -2,3 +2,4 @@
 [ ] UTF-8 support
 [ ] arbitrary memory support (this probably covers UTF-8 support)
 [ ] documentation thats not shit
+[ ] HOOK\_ALL / OFFSHOOT width parameter inconsistency

source/jeger.c

@@ -8,8 +8,13 @@
 #include <string.h>
 #include <limits.h>
 #include <stdlib.h>
+#if DEBUG
+# include <stdio.h>
+#endif
 
-#define JEGER_INIT_STATE    2
+#define JEGER_SOS_STATE   0
+#define JEGER_NSOS_STATE  1
+#define JEGER_INIT_STATE  2
 
 // ------------------
 // ### Char tests ###
@@ -40,6 +45,16 @@ bool is_magic(const char c) {
 		;
 }
 
+// -------------------
+// ### Match tests ###
+// -------------------
+static inline
+bool is_sentinel(const match_t * const match) {
+	return (match->position == -1)
+	    && (match->width    == -1)
+	    ;
+}
+
 // -----------------
 // ### Char sets ###
 // -----------------
@@ -95,11 +110,12 @@ typedef struct {
 } offshoot_t;
 
 enum {
-	DO_CATCH              = 0x00000001 << 0,
-	IS_NEGATIVE           = 0x00000001 << 1,
-	IS_AT_THE_BEGINNING   = 0x00000001 << 2,
-	FORCE_START_OF_STRING = 0x00000001 << 3,
-	INCREMENT_STATE       = 0x00000001 << 4,
+	DO_CATCH                  = 0x00000001 << 0,
+	IS_NEGATIVE               = 0x00000001 << 1,
+	IS_AT_THE_BEGINNING       = 0x00000001 << 2,
+	FORCE_START_OF_STRING     = 0x00000001 << 3,
+	DO_FORBID_START_OF_STRING = 0x00000001 << 4,
+	INCREMENT_STATE           = 0x00000001 << 5,
 };
 
 typedef struct {
@@ -346,7 +362,7 @@ int escape_1_to_N(const char                    c,
 
 static inline
 int escape_to_negative(const char                    c,
-	                         compiler_state * const cs) {
+                             compiler_state * const cs) {
 	switch (c) {
 		case 'D': {
 			const char digit_chars[] = JEGER_CHAR_SET_digits;
@@ -539,21 +555,38 @@ regex_t * regex_compile(const char * const pattern) {
 				s += 1;
 			} break;
 			case '<': {
-				unsigned true_inc = 1;
-				if ((cs.flags & DO_CATCH)
-				||  (cs.flags & IS_NEGATIVE)) {
-					OFFSHOOT(0,  +1, 1, 1, &cs, regex);
-					OFFSHOOT(+1, +2, 1, 1, &cs, regex);
-					++true_inc;
-				} else {
+				// XXX: make this legible
+				if (cs.flags & IS_AT_THE_BEGINNING
+				&& !(cs.flags & DO_CATCH)
+				&& !(cs.flags & IS_NEGATIVE)
+				&& whitelist[0] == '\0') {
+					// ---
 					cs.flags |= INCREMENT_STATE;
+					cs.flags |= DO_FORBID_START_OF_STRING;
+					strcat(whitelist, JEGER_CHAR_symbol_chars);
+					// ---
+					ABSOLUTE_OFFSHOOT( JEGER_SOS_STATE, JEGER_INIT_STATE+1, 0, 0, regex);
+					ABSOLUTE_OFFSHOOT(JEGER_INIT_STATE, JEGER_INIT_STATE+2, 1, 0, regex);
+					HOOK_ALL(0, whitelist, HALT_AND_CATCH_FIRE, &cs, regex);
+					// ---
+					++cs.state;
+					cs.width = 0;
+					HOOK_ALL(0, whitelist, +1, &cs, regex);
+					cs.width = 1;
+					OFFSHOOT(0, +1, 1, 0, &cs, regex);
+					// ---
+				} else {
+					HOOK_ALL(0, whitelist, +1, &cs, regex);
+					if ((cs.flags & DO_CATCH)
+					||  (cs.flags & IS_NEGATIVE)) {
+						OFFSHOOT(+1, +2, 1, 1, &cs, regex);
+					} else {
+						cs.flags |= INCREMENT_STATE;
+					}
+					OFFSHOOT(0, +1, 1, 0, &cs, regex);
 				}
 				cs.flags |= IS_NEGATIVE;
-				if (cs.flags & IS_AT_THE_BEGINNING) {
-					ABSOLUTE_OFFSHOOT(0, JEGER_INIT_STATE + true_inc, 0, 0, regex);
-				}
 				strcat(blacklist, JEGER_CHAR_symbol_chars);
-				//OFFSHOOT(0 + (true_inc-1), +true_inc, 1, 0, &cs, regex);
 				s += 1;
 			} break;
 			case '>': {
@@ -622,11 +655,13 @@ regex_t * regex_compile(const char * const pattern) {
 	}
 
 	// Init state hookups
-	ABSOLUTE_OFFSHOOT(0, JEGER_INIT_STATE, 0, 0, regex);
+	if (!(cs.flags & DO_FORBID_START_OF_STRING)) {
+		ABSOLUTE_OFFSHOOT(JEGER_SOS_STATE, JEGER_INIT_STATE, 0, 0, regex);
+	}
 	if (cs.flags & FORCE_START_OF_STRING) {
-		ABSOLUTE_OFFSHOOT(1, HALT_AND_CATCH_FIRE, 0, 0, regex);
+		ABSOLUTE_OFFSHOOT(JEGER_NSOS_STATE, HALT_AND_CATCH_FIRE, 0, 0, regex);
 	} else {
-		ABSOLUTE_OFFSHOOT(1,    JEGER_INIT_STATE, 0, 0, regex);
+		ABSOLUTE_OFFSHOOT(JEGER_NSOS_STATE,    JEGER_INIT_STATE, 0, 0, regex);
 	}
 
 	regex->accepting_state = cs.state;
@@ -702,9 +737,6 @@ bool regex_assert(const regex_t * const         regex,
 				was_found = true;
 				const int r = regex_assert(regex, s + delta->pattern_width, delta->to, match);
 				if(r){
-					if (match->position == -1) {
-						match->position = (s - string);
-					}
 					match->width += delta->match_width;
 					return r;
 				}
@@ -718,6 +750,9 @@ bool regex_assert(const regex_t * const         regex,
 			if (my_catch && (!my_catch->pattern_width || !last_stand)) {
 				state = my_catch->to;
 				s += my_catch->pattern_width;
+				if (match->position < 1) {
+					match->position = my_catch->match_width;
+				}
 				match->width += my_catch->match_width;
 				goto LOOP;
 			}
@@ -756,7 +791,11 @@ match_t * regex_match(const regex_t * const              regex,
 			};
 
 			if (regex_assert(regex, s, initial_state, match)) {
+				if(match->position == -1){
 				match->position = (s - string);
+				}else{
+				match->position += (s - string);
+				}
 
 				vector_push(&matches, match);
 
@@ -790,7 +829,7 @@ bool regex_search(const regex_t * const  regex,
                   const char    * const string) {
 
 	match_t * m = regex_match(regex, string, true);
-	const bool r = (m->position != -1);
+	const bool r = !is_sentinel(m);
 	free(m);
 
 	return r;

source/main.cpp

@@ -107,6 +107,14 @@ signed main() {
 	TEST(R"del(.\<print\>.)del",  "printf", false);
 	TEST(R"del(.\<print\>.)del", "fprintf", false);
 
+	puts("");
+
+	TEST(R"del(\<while\>)del",       "while", true);
+	TEST(R"del(\<while\>)del",     " while ", true);
+	TEST(R"del(\<while\>)del",     "9while ", true);
+	TEST(R"del(\<while\>)del", "for while {", true);
+	TEST(R"del(\<while\>)del",  "for while{", true);
+
 	if (test_counter == passed_tests) {
 		fputs("\033[32m", stdout);
 	} else {
@@ -143,6 +151,15 @@ signed main() {
 	TEST2( R"del(a+a)del", "   aaa", match_t{ 3, strlen("aaa")});
 	TEST2(R"del(a+\+)del",   "aaa+", match_t{ 0, strlen("aaa+")});
 
+	puts("");
+	puts("");
+
+	TEST2(R"del(\<while\>)del",      "while", match_t{0, strlen("while")});
+	TEST2(R"del(\<while\>)del",     " while", match_t{1, strlen("while")});
+	TEST2(R"del(\<while\>)del",  "for while", match_t{4, strlen("while")});
+	TEST2(R"del(\<while\>)del",  "for9while", match_t{4, strlen("while")});
+	TEST2(R"del(\<while\>)del", "for9while ", match_t{4, strlen("while")});
+
 	if(test_counter2 == passed_tests2) {
 		fputs("\033[32m", stdout);
 	} else {