libhl/source/regex.c

/* regex.c
 * Copyright 2023 Anon Anonson, Ognjen 'xolatile' Milan Robovic, Emil Williams
 * SPDX Identifier: GPL-3.0-only / NO WARRANTY / NO GUARANTEE */

#include "regex.h"

#include <assert.h>
#include <string.h>

// ------------------
// ### Char tests ###
// ------------------
static bool is_quantifier(const char c) {
	for (const char * s = "+*?"; *s != '\00'; s++) {
		if (*s == c) {
			return true;
		}
	}
	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;
}



// ----------------------
// ### Internal Types ###
// ----------------------
typedef struct {
	int in;
	char input;
	int to;
} delta_t;

typedef struct {
	int in;
	int to;
} offshoot_t;



// ----------------------------------
// ### Regex creation/destruction ###
// ----------------------------------
static int escape_1_to_1(const char c, char * whitelist) {
	switch(c) {
		case 't': {
			strcat(whitelist, "\t");
		} return 1;
		case 'n': {
			strcat(whitelist, "\n");
		} return 1;
		case 'r': {
			strcat(whitelist, "\r");
		} return 1;
		case 'b': {
			strcat(whitelist, "\b");
		} return 1;
		case '[': {
			strcat(whitelist, "[");
		} return 1;
		case ']': {
			strcat(whitelist, "]");
		} return 1;
		case '.': {
			strcat(whitelist, ".");
		} return 1;
		case '?': {
			strcat(whitelist, "?");
		} return 1;
		case '+': {
			strcat(whitelist, "+");
		} return 1;
		case '*': {
			strcat(whitelist, "*");
		} return 1;
		case '\\': {
			strcat(whitelist, "\\");
		} return 1;
	}

	return 0;
}

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";
			strcpy(whitelist, identifier_chars);
			return sizeof(identifier_chars)-1;
		};
		case 'I': {
			const char identifier_chars[] = "@_\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";
			strcpy(whitelist, identifier_chars);
			return sizeof(identifier_chars)-1;
		};
		case 'k': {
			const char keyword_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";
			strcpy(whitelist, keyword_chars);
			return sizeof(keyword_chars)-1;
		};
		case 'K': {
			const char keyword_chars[] = "@_\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";
			strcpy(whitelist, keyword_chars);
			return sizeof(keyword_chars)-1;
		};
		case 'f': {
			const char filename_chars[] = "@0123456789/.-_+,#$%~=";
			strcpy(whitelist, filename_chars);
			return sizeof(filename_chars)-1;
		};
		case 'F': {
			const char filename_chars[] = "@/.-_+,#$%~=";
			strcpy(whitelist, filename_chars);
			return sizeof(filename_chars)-1;
		};
		case 'p': {
			const char printable_chars[] = "@\241\242\243\244\245\246\247\250\251\252\253\254\255\256\257\260\261\262\263\264\265\266\267\270\271\272\273\274\275\276\277\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";
			strcpy(whitelist, printable_chars);
			return sizeof(printable_chars)-1;
		};
		case 'P': {
			const char printable_chars[] = "@\241\242\243\244\245\246\247\250\251\252\253\254\255\256\257\260\261\262\263\264\265\266\267\270\271\272\273\274\275\276\277\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";
			strcpy(whitelist, printable_chars);
			return sizeof(printable_chars)-1;
		};
		case 's': {
			const char whitespace_chars[] = " \t\v\n";
			strcpy(whitelist, whitespace_chars);
			return sizeof(whitespace_chars)-1;
		};
		case 'd': {
			const char digit_chars[] = "0123456789";
			strcpy(whitelist, digit_chars);
			return sizeof(digit_chars)-1;
		};
		case 'x': {
			const char hex_chars[] = "0123456789abcdefABCDEF";
			strcpy(whitelist, hex_chars);
			return sizeof(hex_chars)-1;
		};
		case 'o': {
			const char oct_chars[] = "01234567";
			strcpy(whitelist, oct_chars);
			return sizeof(oct_chars)-1;
		};
		case 'w': {
			const char word_chars[] = "0123456789abcdefghijklmnopqrstuwxyzABCDEFGHIJKLMNOPQRSTUWXYZ_";
			strcpy(whitelist, word_chars);
			return sizeof(word_chars)-1;
		};
		case 'h': {
			const char very_word_chars[] = "abcdefghijklmnopqrstuwxyzABCDEFGHIJKLMNOPQRSTUWXYZ_";
			strcpy(whitelist, very_word_chars);
			return sizeof(very_word_chars)-1;
		};
		case 'a': {
			const char alpha_chars[] = "abcdefghijklmnopqrstuwxyzABCDEFGHIJKLMNOPQRSTUWXYZ";
			strcpy(whitelist, alpha_chars);
			return sizeof(alpha_chars)-1;
		};
		case 'l': {
			const char lower_alpha_chars[] = "abcdefghijklmnopqrstuwxyz";
			strcpy(whitelist, lower_alpha_chars);
			return sizeof(lower_alpha_chars)-1;
		};
		case 'u': {
			const char upper_alpha_chars[] = "ABCDEFGHIJKLMNOPQRSTUWXYZ";
			strcpy(whitelist, upper_alpha_chars);
			return sizeof(upper_alpha_chars)-1;
		};
	}

	return 0;
}

static int compile_range(const char * const     range,
                               char *       whitelist,
							   bool *     is_negative) {
	assert(range[0] == '[' && "Not a range.");

	int r = 0;
	const char * s;
	if (range[1] == '^') {
		*is_negative = true;
		s = range + 2;
	} else {
		s = range + 1;
	}
	for (; *s != ']'; s++) {
		assert(*s != '\00' && "Unclosed range.");
		char c = *s;
		if (escape_1_to_1(c, whitelist)
		||  escape_1_to_N(c, whitelist)) {
			;
		} else if (*(s+1) == '-') {
			char end = *(s+2);
			assert(c < end && "Endless range.");
			for (char cc = c; cc < end+1; cc++) {
				strncat(whitelist,   &cc, 1);
				strncat(whitelist, "\00", 1);
			}
			s += 2;
		} else {
			++r;
			strncat(whitelist,    &c, 1);
			strncat(whitelist, "\00", 1);
		}
	}

	return ((s - range) + 1);
}

static bool catch_(const regex_t * const regex,
                         int     * const state) {

	for (int i = 0; i < regex->catch_table.element_size; i++){
		const offshoot_t * const offshoot = (vector_get(&regex->catch_table, i));
		if (offshoot->in == *state) {
			*state = offshoot->to;
			return true;
		}
	}
	return false;
}

#define HALT_AND_CATCH_FIRE -1

#define HOOK_ALL(from, str, to) do {                   \
	int hook_to = (is_negative) ? -1 : state + to;     \
	for (char * s = str; *s != '\00'; s++) {           \
		vector_push(&regex->delta_table,               \
			&(delta_t){state + from, *s, hook_to}      \
		);                                             \
	}                                                  \
	if (do_catch) {                                    \
		vector_push(&regex->catch_table,               \
			&(offshoot_t){state + from, hook_to}       \
		);                                             \
	}                                                  \
} while (0)

#define EAT(n) do { \
	s += n;         \
} while (0)

regex_t * regex_compile(const char * const pattern) {
	regex_t * regex = (regex_t *)malloc(sizeof(regex_t));
	regex->str = strdup(pattern);
	vector_init(&regex->delta_table, sizeof(delta_t), 32);
	vector_init(&regex->catch_table, sizeof(offshoot_t), 16);

	int state = 0;

	char whitelist[64];
	bool do_catch;
	bool is_negative;
	for (const char * s = pattern; *s != '\00';) {
		// Get token
		assert(!is_quantifier(*pattern) && "Pattern starts with quantifier.");
		whitelist[0] = '\00';
		do_catch     = false;

		switch (*s) {
			case '.': {
				do_catch = true;
			} break;
			case '\\': {
				EAT(1);
				if(escape_1_to_1(*s, whitelist)
				|| escape_1_to_N(*s, whitelist)){
					;
				} else {
					assert(!"Unknown escape.");
				}
			} break;
			case '[': {
				EAT(compile_range(s, whitelist, &is_negative)-1);
			} break;
			default: {
				whitelist[0] = *s;
				whitelist[1] = '\00';
			} break;
		}

		EAT(1);

		// Get quantifier
		switch (*s) {
			case '?': {
				HOOK_ALL(0, whitelist, +1);
				EAT(1);
			} break;
			case '*': {
				HOOK_ALL(0, whitelist,  0);
				EAT(1);
			} break;
			case '+': {
				HOOK_ALL(0, whitelist, +1);
				state += 1;
				HOOK_ALL(0, whitelist,  0);
				EAT(1);
			} break;
			default: { // Literal
				HOOK_ALL(0, whitelist, +1);
				state += 1;
			} break;
		}
	}

	regex->accepting_state = state;

	return regex;
}

int regex_free(regex_t * const regex) {
	free(regex->str);
	vector_free(&regex->delta_table);
	vector_free(&regex->catch_table);
	free(regex);
	return 0;
}



// -----------------
// ### Searching ###
// -----------------
static bool regex_assert(const regex_t * const  regex,
                         const char    * const string,
				               int              state) {

	for (const char * s = string; *s != '\00'; s++) {
		// delta
		for (int i = 0; i < regex->delta_table.element_count; i++) {
			const delta_t * const delta = (delta_t *)(vector_get(&regex->delta_table, i));
			if ((delta->in == state) 
			&&  (delta->input == *s)) {
				if(regex_assert(regex, s+1, delta->to)){
					return true;
				}
			}
		}

		if (catch_(regex, &state)) {
			continue;
		}

		return false;
	}

	return (state == regex->accepting_state);
}

bool regex_search(      regex_t *        regex,
                  const char    * const string) {

	if (regex == NULL) {
		return false;
	}
	if (string == NULL) {
		return true;
	}

	return regex_assert(regex, string, 0);
}