X-Git-Url: https://git.m6w6.name/?p=m6w6%2Fext-http;a=blobdiff_plain;f=php_http_url.c;h=05850e8c42a74482c90a8725147ebf4d4d36f751;hp=3c3fa4d58d8e2c687e7f91216f73c1b81eacda22;hb=31636a8ec8578e7e4d45da180d55e1260bb527db;hpb=a760e170d36cc40ce992ebcdb13f0877ddfa3d73

diff --git a/php_http_url.c b/php_http_url.c
index 3c3fa4d..05850e8 100644
--- a/php_http_url.c
+++ b/php_http_url.c
@@ -314,6 +314,19 @@ STATUS php_http_url_encode_hash_ex(HashTable *hash, php_http_buffer_t *qstr, con
 	return SUCCESS;
 }
 
+struct parse_state {
+	php_http_url_t url;
+#ifdef ZTS
+	void ***ts;
+#endif
+	const char *ptr;
+	const char *end;
+	size_t maxlen;
+	off_t offset;
+	unsigned flags;
+	char buffer[1]; /* last member */
+};
+
 void php_http_url_free(php_http_url_t **url)
 {
 	if (*url) {
@@ -322,7 +335,7 @@ void php_http_url_free(php_http_url_t **url)
 	}
 }
 
-static size_t parse_mb_utf8(php_http_url_t *url, const char *ptr, const char *end, zend_bool idn)
+static size_t parse_mb_utf8(unsigned *wc, const char *ptr, const char *end)
 {
 	unsigned wchar;
 	size_t consumed = utf8towc(&wchar, (const unsigned char *) ptr, end - ptr);
@@ -330,15 +343,15 @@ static size_t parse_mb_utf8(php_http_url_t *url, const char *ptr, const char *en
 	if (!consumed || consumed == (size_t) -1) {
 		return 0;
 	}
-	if (!idn && !isualnum(wchar)) {
-		return 0;
-	}
 
+	if (wc) {
+		*wc = wchar;
+	}
 	return consumed;
 }
 
 #ifdef PHP_HTTP_HAVE_WCHAR
-static size_t parse_mb_loc(php_http_url_t *url, const char *ptr, const char *end, zend_bool idn)
+static size_t parse_mb_loc(unsigned *wc, const char *ptr, const char *end)
 {
 	wchar_t wchar;
 	size_t consumed = 0;
@@ -353,10 +366,10 @@ static size_t parse_mb_loc(php_http_url_t *url, const char *ptr, const char *end
 	if (!consumed || consumed == (size_t) -1) {
 		return 0;
 	}
-	if (!idn && !iswalnum(wchar)) {
-		return 0;
-	}
 
+	if (wc) {
+		*wc = wchar;
+	}
 	return consumed;
 }
 #endif
@@ -379,37 +392,69 @@ static const char * const parse_what[] = {
 	"fragment"
 };
 
-static size_t parse_mb(php_http_url_t *url, parse_mb_what_t what, const char *ptr, const char *end, const char *begin, zend_bool silent)
+static const char parse_xdigits[] = "0123456789ABCDEF";
+
+static size_t parse_mb(struct parse_state *state, parse_mb_what_t what, const char *ptr, const char *end, const char *begin, zend_bool silent)
 {
+	unsigned wchar;
 	size_t consumed = 0;
-	zend_bool idn = (what == PARSE_HOSTINFO) && (url->flags & PHP_HTTP_URL_PARSE_IDN);
 
-	if (url->flags & PHP_HTTP_URL_PARSE_MBUTF8) {
-		consumed = parse_mb_utf8(url, ptr, end, idn);
+	if (state->flags & PHP_HTTP_URL_PARSE_MBUTF8) {
+		consumed = parse_mb_utf8(&wchar, ptr, end);
 	}
 #ifdef PHP_HTTP_HAVE_WCHAR
-	else if (url->flags & PHP_HTTP_URL_PARSE_MBLOC) {
-		consumed = parse_mb_loc(url, ptr, end, idn);
+	else if (state->flags & PHP_HTTP_URL_PARSE_MBLOC) {
+		consumed = parse_mb_loc(&wchar, ptr, end);
 	}
 #endif
 
-	if (consumed) {
-		PHP_HTTP_DUFF(consumed, url->buffer[url->offset++] = *ptr++);
-	} else if (!silent) {
-		TSRMLS_FETCH_FROM_CTX(url->ts);
+	while (consumed) {
+		if (!(state->flags & PHP_HTTP_URL_PARSE_TOPCT) || what == PARSE_HOSTINFO || what == PARSE_SCHEME) {
+			if (what == PARSE_HOSTINFO && (state->flags & PHP_HTTP_URL_PARSE_TOIDN)) {
+				/* idna */
+			} else if (state->flags & PHP_HTTP_URL_PARSE_MBUTF8) {
+				if (!isualnum(wchar)) {
+					break;
+				}
+#ifdef PHP_HTTP_HAVE_WCHAR
+			} else if (state->flags & PHP_HTTP_URL_PARSE_MBLOC) {
+				if (!iswalnum(wchar)) {
+					break;
+				}
+#endif
+			}
+			PHP_HTTP_DUFF(consumed, state->buffer[state->offset++] = *ptr++);
+		} else {
+			int i = 0;
+
+			PHP_HTTP_DUFF(consumed,
+					state->buffer[state->offset++] = '%';
+					state->buffer[state->offset++] = parse_xdigits[((unsigned char) ptr[i]) >> 4];
+					state->buffer[state->offset++] = parse_xdigits[((unsigned char) ptr[i]) & 0xf];
+					++i;
+			);
+		}
+
+		return consumed;
+	}
+
+	if (!silent) {
+		TSRMLS_FETCH_FROM_CTX(state->ts);
 		php_error_docref(NULL TSRMLS_CC, E_WARNING,
 				"Failed to parse %s; unexpected byte 0x%02x at pos %u in '%s'",
 				parse_what[what], (unsigned char) *ptr, (unsigned) (ptr - begin), begin);
 	}
 
-	return consumed;
+	return 0;
 }
 
-static STATUS parse_userinfo(php_http_url_t *url, const char *ptr)
+static STATUS parse_userinfo(struct parse_state *state, const char *ptr)
 {
 	size_t mb;
-	const char *password = NULL, *end = url->ptr, *tmp = ptr;
-	TSRMLS_FETCH_FROM_CTX(url->ts);
+	const char *password = NULL, *end = state->ptr, *tmp = ptr;
+	TSRMLS_FETCH_FROM_CTX(state->ts);
+
+	state->url.user = &state->buffer[state->offset];
 
 	do {
 		switch (*ptr) {
@@ -421,7 +466,8 @@ static STATUS parse_userinfo(php_http_url_t *url, const char *ptr)
 				return FAILURE;
 			}
 			password = ptr + 1;
-			url->buffer[url->offset++] = *ptr;
+			state->buffer[state->offset++] = 0;
+			state->url.pass = &state->buffer[state->offset];
 			break;
 
 		case '%':
@@ -431,9 +477,9 @@ static STATUS parse_userinfo(php_http_url_t *url, const char *ptr)
 						(unsigned) (ptr - tmp), tmp);
 				return FAILURE;
 			}
-			url->buffer[url->offset++] = *ptr++;
-			url->buffer[url->offset++] = *ptr++;
-			url->buffer[url->offset++] = *ptr;
+			state->buffer[state->offset++] = *ptr++;
+			state->buffer[state->offset++] = *ptr++;
+			state->buffer[state->offset++] = *ptr;
 			break;
 
 		case '!': case '$': case '&': case '\'': case '(': case ')': case '*':
@@ -450,35 +496,28 @@ static STATUS parse_userinfo(php_http_url_t *url, const char *ptr)
 		case '0': case '1': case '2': case '3': case '4': case '5': case '6':
 		case '7': case '8': case '9':
 			/* allowed */
-			url->buffer[url->offset++] = *ptr;
+			state->buffer[state->offset++] = *ptr;
 			break;
 
 		default:
-			if (!(mb = parse_mb(url, PARSE_USERINFO, ptr, end, tmp, 0))) {
+			if (!(mb = parse_mb(state, PARSE_USERINFO, ptr, end, tmp, 0))) {
 				return FAILURE;
 			}
 			ptr += mb - 1;
 		}
 	} while(++ptr != end);
 
-	if (password) {
-		url->user = &url->buffer[url->offset - (end - password) - (password - tmp)];
-		url->buffer[url->offset - (end - password) - 1] = 0;
-		url->pass = &url->buffer[url->offset - (end - password)];
-		url->buffer[url->offset++] = 0;
-	} else {
-		url->user = &url->buffer[url->offset - (end - tmp)];
-		url->buffer[url->offset++] = 0;
-	}
+
+	state->buffer[state->offset++] = 0;
 
 	return SUCCESS;
 }
 
-static STATUS parse_hostinfo(php_http_url_t *url, const char *ptr)
+static STATUS parse_hostinfo(struct parse_state *state, const char *ptr)
 {
 	size_t mb, len;
-	const char *end = url->ptr, *tmp = ptr, *port = NULL;
-	TSRMLS_FETCH_FROM_CTX(url->ts);
+	const char *end = state->ptr, *tmp = ptr, *port = NULL;
+	TSRMLS_FETCH_FROM_CTX(state->ts);
 
 
 #ifdef HAVE_INET_PTON
@@ -492,12 +531,12 @@ static STATUS parse_hostinfo(php_http_url_t *url, const char *ptr)
 
 			efree(addr);
 			if (rv == 1) {
-				url->buffer[url->offset] = '[';
-				url->host = &url->buffer[url->offset];
-				inet_ntop(AF_INET6, buf, url->host + 1, url->maxlen - url->offset);
-				url->offset += strlen(url->host);
-				url->buffer[url->offset++] = ']';
-				url->buffer[url->offset++] = 0;
+				state->buffer[state->offset] = '[';
+				state->url.host = &state->buffer[state->offset];
+				inet_ntop(AF_INET6, buf, state->url.host + 1, state->maxlen - state->offset);
+				state->offset += strlen(state->url.host);
+				state->buffer[state->offset++] = ']';
+				state->buffer[state->offset++] = 0;
 				ptr = tmp + 1;
 			} else if (rv == -1) {
 				error = strerror(errno);
@@ -519,7 +558,7 @@ static STATUS parse_hostinfo(php_http_url_t *url, const char *ptr)
 		case ':':
 			if (port) {
 				php_error_docref(NULL TSRMLS_CC, E_WARNING,
-						"Failed to parse port; duplicate ':' at pos %u in '%s'",
+						"Failed to parse port; unexpected ':' at pos %u in '%s'",
 						(unsigned) (ptr - tmp), tmp);
 				return FAILURE;
 			}
@@ -533,9 +572,9 @@ static STATUS parse_hostinfo(php_http_url_t *url, const char *ptr)
 						(unsigned) (ptr - tmp), tmp);
 				return FAILURE;
 			}
-			url->buffer[url->offset++] = *ptr++;
-			url->buffer[url->offset++] = *ptr++;
-			url->buffer[url->offset++] = *ptr;
+			state->buffer[state->offset++] = *ptr++;
+			state->buffer[state->offset++] = *ptr++;
+			state->buffer[state->offset++] = *ptr;
 			break;
 
 		case '!': case '$': case '&': case '\'': case '(': case ')': case '*':
@@ -560,10 +599,10 @@ static STATUS parse_hostinfo(php_http_url_t *url, const char *ptr)
 		case '7': case '8': case '9':
 			/* allowed */
 			if (port) {
-				url->port *= 10;
-				url->port += *ptr - '0';
+				state->url.port *= 10;
+				state->url.port += *ptr - '0';
 			} else {
-				url->buffer[url->offset++] = *ptr;
+				state->buffer[state->offset++] = *ptr;
 			}
 			break;
 
@@ -573,30 +612,30 @@ static STATUS parse_hostinfo(php_http_url_t *url, const char *ptr)
 						"Failed to parse port; unexpected byte 0x%02x at pos %u in '%s'",
 						(unsigned char) *ptr, (unsigned) (ptr - tmp), tmp);
 				return FAILURE;
-			} else if (!(mb = parse_mb(url, PARSE_HOSTINFO, ptr, end, tmp, 0))) {
+			} else if (!(mb = parse_mb(state, PARSE_HOSTINFO, ptr, end, tmp, 0))) {
 				return FAILURE;
 			}
 			ptr += mb - 1;
 		}
 	} while (++ptr != end);
 
-	if (!url->host) {
+	if (!state->url.host) {
 		len = (port ? port - tmp - 1 : end - tmp);
-		url->host = &url->buffer[url->offset - len];
-		url->buffer[url->offset++] = 0;
+		state->url.host = &state->buffer[state->offset - len];
+		state->buffer[state->offset++] = 0;
 	}
 
 #ifdef PHP_HTTP_HAVE_IDN
-	if (url->flags & PHP_HTTP_URL_PARSE_IDN) {
+	if (state->flags & PHP_HTTP_URL_PARSE_TOIDN) {
 		char *idn = NULL;
 		int rv = -1;
 
-		if (url->flags & PHP_HTTP_URL_PARSE_MBUTF8) {
-			rv = idna_to_ascii_8z(url->host, &idn, IDNA_ALLOW_UNASSIGNED|IDNA_USE_STD3_ASCII_RULES);
+		if (state->flags & PHP_HTTP_URL_PARSE_MBUTF8) {
+			rv = idna_to_ascii_8z(state->url.host, &idn, IDNA_ALLOW_UNASSIGNED|IDNA_USE_STD3_ASCII_RULES);
 		}
 #	ifdef PHP_HTTP_HAVE_WCHAR
-		else if (url->flags & PHP_HTTP_URL_PARSE_MBLOC) {
-			rv = idna_to_ascii_lz(url->host, &idn, IDNA_ALLOW_UNASSIGNED|IDNA_USE_STD3_ASCII_RULES);
+		else if (state->flags & PHP_HTTP_URL_PARSE_MBLOC) {
+			rv = idna_to_ascii_lz(state->url.host, &idn, IDNA_ALLOW_UNASSIGNED|IDNA_USE_STD3_ASCII_RULES);
 		}
 #	endif
 		if (rv != IDNA_SUCCESS) {
@@ -604,9 +643,9 @@ static STATUS parse_hostinfo(php_http_url_t *url, const char *ptr)
 			return FAILURE;
 		} else {
 			size_t idnlen = strlen(idn);
-			memcpy(url->host, idn, idnlen + 1);
+			memcpy(state->url.host, idn, idnlen + 1);
 			free(idn);
-			url->offset += idnlen - len;
+			state->offset += idnlen - len;
 		}
 	}
 #endif
@@ -614,18 +653,25 @@ static STATUS parse_hostinfo(php_http_url_t *url, const char *ptr)
 	return SUCCESS;
 }
 
-static const char *parse_authority(php_http_url_t *url)
+static const char *parse_authority(struct parse_state *state)
 {
-	const char *tmp = url->ptr;
+	const char *tmp = state->ptr, *host = NULL;
 
 	do {
-		switch (*url->ptr) {
+		switch (*state->ptr) {
 		case '@':
 			/* userinfo delimiter */
-			if (tmp != url->ptr && SUCCESS != parse_userinfo(url, tmp)) {
+			if (host) {
+				TSRMLS_FETCH_FROM_CTX(state->ts);
+				php_error_docref(NULL TSRMLS_CC, E_WARNING,
+						"Failed to parse userinfo; unexpected '@'");
+				return NULL;
+			}
+			host = state->ptr + 1;
+			if (tmp != state->ptr && SUCCESS != parse_userinfo(state, tmp)) {
 				return NULL;
 			}
-			tmp = url->ptr + 1;
+			tmp = state->ptr + 1;
 			break;
 
 		case '/':
@@ -633,49 +679,52 @@ static const char *parse_authority(php_http_url_t *url)
 		case '#':
 		case '\0':
 			/* host delimiter */
-			if (tmp != url->ptr && SUCCESS != parse_hostinfo(url, tmp)) {
+			if (tmp != state->ptr && SUCCESS != parse_hostinfo(state, tmp)) {
 				return NULL;
 			}
-			return url->ptr;
+			return state->ptr;
 		}
-	} while (++url->ptr <= url->end);
+	} while (++state->ptr <= state->end);
 
 	return NULL;
 }
 
-static const char *parse_path(php_http_url_t *url)
+static const char *parse_path(struct parse_state *state)
 {
 	size_t mb;
 	const char *tmp;
-	TSRMLS_FETCH_FROM_CTX(url->ts);
+	TSRMLS_FETCH_FROM_CTX(state->ts);
 
 	/* is there actually a path to parse? */
-	if (!*url->ptr) {
-		return url->ptr;
+	if (!*state->ptr) {
+		return state->ptr;
 	}
-	tmp = url->ptr;
+	tmp = state->ptr;
+	state->url.path = &state->buffer[state->offset];
 
 	do {
-		switch (*url->ptr) {
+		switch (*state->ptr) {
+		case '#':
 		case '?':
 		case '\0':
 			/* did we have any path component ? */
-			if (tmp != url->ptr) {
-				url->path = &url->buffer[url->offset - (url->ptr - tmp)];
-				url->buffer[url->offset++] = 0;
+			if (tmp != state->ptr) {
+				state->buffer[state->offset++] = 0;
+			} else {
+				state->url.path = NULL;
 			}
-			return url->ptr;
+			return state->ptr;
 
 		case '%':
-			if (url->ptr[1] != '%' && (url->end - url->ptr <= 2 || !isxdigit(*(url->ptr+1)) || !isxdigit(*(url->ptr+2)))) {
+			if (state->ptr[1] != '%' && (state->end - state->ptr <= 2 || !isxdigit(*(state->ptr+1)) || !isxdigit(*(state->ptr+2)))) {
 				php_error_docref(NULL TSRMLS_CC, E_WARNING,
 						"Failed to parse path; invalid percent encoding at pos %u in '%s'",
-						(unsigned) (url->ptr - tmp), tmp);
+						(unsigned) (state->ptr - tmp), tmp);
 				return NULL;
 			}
-			url->buffer[url->offset++] = *url->ptr++;
-			url->buffer[url->offset++] = *url->ptr++;
-			url->buffer[url->offset++] = *url->ptr;
+			state->buffer[state->offset++] = *state->ptr++;
+			state->buffer[state->offset++] = *state->ptr++;
+			state->buffer[state->offset++] = *state->ptr;
 			break;
 
 		case '/': /* yeah, well */
@@ -694,52 +743,52 @@ static const char *parse_path(php_http_url_t *url)
 		case '0': case '1': case '2': case '3': case '4': case '5': case '6':
 		case '7': case '8': case '9':
 			/* allowed */
-			url->buffer[url->offset++] = *url->ptr;
+			state->buffer[state->offset++] = *state->ptr;
 			break;
 
 		default:
-			if (!(mb = parse_mb(url, PARSE_PATH, url->ptr, url->end, tmp, 0))) {
+			if (!(mb = parse_mb(state, PARSE_PATH, state->ptr, state->end, tmp, 0))) {
 				return NULL;
 			}
-			url->ptr += mb - 1;
+			state->ptr += mb - 1;
 		}
-	} while (++url->ptr <= url->end);
+	} while (++state->ptr <= state->end);
 
 	return NULL;
 }
 
-static const char *parse_query(php_http_url_t *url)
+static const char *parse_query(struct parse_state *state)
 {
 	size_t mb;
-	const char *tmp = url->ptr + !!*url->ptr;
-	TSRMLS_FETCH_FROM_CTX(url->ts);
+	const char *tmp = state->ptr + !!*state->ptr;
+	TSRMLS_FETCH_FROM_CTX(state->ts);
 
-	/* is there actually a query to parse ? */
-	if (!*url->ptr || *url->ptr != '?') {
-		return url->ptr;
+	/* is there actually a query to parse? */
+	if (*state->ptr != '?') {
+		return state->ptr;
 	}
 
 	/* skip initial '?' */
-	tmp = url->ptr + 1;
+	tmp = ++state->ptr;
+	state->url.query = &state->buffer[state->offset];
 
 	do {
-		switch (*url->ptr) {
+		switch (*state->ptr) {
 		case '#':
 		case '\0':
-			url->query = &url->buffer[url->offset - (url->ptr - tmp)];
-			url->buffer[url->offset++] = 0;
-			return url->ptr;
+			state->buffer[state->offset++] = 0;
+			return state->ptr;
 
 		case '%':
-			if (url->ptr[1] != '%' && (url->end - url->ptr <= 2 || !isxdigit(*(url->ptr+1)) || !isxdigit(*(url->ptr+2)))) {
+			if (state->ptr[1] != '%' && (state->end - state->ptr <= 2 || !isxdigit(*(state->ptr+1)) || !isxdigit(*(state->ptr+2)))) {
 				php_error_docref(NULL TSRMLS_CC, E_WARNING,
 						"Failed to parse query; invalid percent encoding at pos %u in '%s'",
-						(unsigned) (url->ptr - tmp), tmp);
+						(unsigned) (state->ptr - tmp), tmp);
 				return NULL;
 			}
-			url->buffer[url->offset++] = *url->ptr++;
-			url->buffer[url->offset++] = *url->ptr++;
-			url->buffer[url->offset++] = *url->ptr;
+			state->buffer[state->offset++] = *state->ptr++;
+			state->buffer[state->offset++] = *state->ptr++;
+			state->buffer[state->offset++] = *state->ptr;
 			break;
 
 		case '?': case '/': /* yeah, well */
@@ -758,54 +807,54 @@ static const char *parse_query(php_http_url_t *url)
 		case '0': case '1': case '2': case '3': case '4': case '5': case '6':
 		case '7': case '8': case '9':
 			/* allowed */
-			url->buffer[url->offset++] = *url->ptr;
+			state->buffer[state->offset++] = *state->ptr;
 			break;
 
 		default:
-			if (!(mb = parse_mb(url, PARSE_QUERY, url->ptr, url->end, tmp, 0))) {
+			if (!(mb = parse_mb(state, PARSE_QUERY, state->ptr, state->end, tmp, 0))) {
 				return NULL;
 			}
-			url->ptr += mb - 1;
+			state->ptr += mb - 1;
 		}
-	} while (++url->ptr <= url->end);
+	} while (++state->ptr <= state->end);
 
 	return NULL;
 }
 
-static const char *parse_fragment(php_http_url_t *url)
+static const char *parse_fragment(struct parse_state *state)
 {
 	size_t mb;
 	const char *tmp;
-	TSRMLS_FETCH_FROM_CTX(url->ts);
+	TSRMLS_FETCH_FROM_CTX(state->ts);
 
-	/* is there actually a fragment to parse */
-	if (!*url->ptr || *url->ptr != '#') {
-		return url->ptr;
+	/* is there actually a fragment to parse? */
+	if (*state->ptr != '#') {
+		return state->ptr;
 	}
 
 	/* skip initial '#' */
-	tmp = url->ptr + 1;
+	tmp = ++state->ptr;
+	state->url.fragment = &state->buffer[state->offset];
 
 	do {
-		switch (*url->ptr) {
+		switch (*state->ptr) {
 		case '\0':
-			url->fragment = &url->buffer[url->offset - (url->ptr - tmp)];
-			url->buffer[url->offset++] = 0;
-			return url->ptr;
+			state->buffer[state->offset++] = 0;
+			return state->ptr;
 
 		case '%':
-			if (url->ptr[1] != '%' && (url->end - url->ptr <= 2 || !isxdigit(*(url->ptr+1)) || !isxdigit(*(url->ptr+2)))) {
+			if (state->ptr[1] != '%' && (state->end - state->ptr <= 2 || !isxdigit(*(state->ptr+1)) || !isxdigit(*(state->ptr+2)))) {
 				php_error_docref(NULL TSRMLS_CC, E_WARNING,
 						"Failed to parse fragment; invalid percent encoding at pos %u in '%s'",
-						(unsigned) (url->ptr - tmp), tmp);
+						(unsigned) (state->ptr - tmp), tmp);
 				return NULL;
 			}
-			url->buffer[url->offset++] = *url->ptr++;
-			url->buffer[url->offset++] = *url->ptr++;
-			url->buffer[url->offset++] = *url->ptr;
+			state->buffer[state->offset++] = *state->ptr++;
+			state->buffer[state->offset++] = *state->ptr++;
+			state->buffer[state->offset++] = *state->ptr;
 			break;
 
-		case '?': case '/': /* yeah, well */
+		case '?': case '/':
 		case '!': case '$': case '&': case '\'': case '(': case ')': case '*':
 		case '+': case ',': case ';': case '=': /* sub-delims */
 		case '-': case '.': case '_': case '~': /* unreserved */
@@ -821,52 +870,52 @@ static const char *parse_fragment(php_http_url_t *url)
 		case '0': case '1': case '2': case '3': case '4': case '5': case '6':
 		case '7': case '8': case '9':
 			/* allowed */
-			url->buffer[url->offset++] = *url->ptr;
+			state->buffer[state->offset++] = *state->ptr;
 			break;
 
 		default:
-			if (!(mb = parse_mb(url, PARSE_FRAGMENT, url->ptr, url->end, tmp, 0))) {
+			if (!(mb = parse_mb(state, PARSE_FRAGMENT, state->ptr, state->end, tmp, 0))) {
 				return NULL;
 			}
-			url->ptr += mb - 1;
+			state->ptr += mb - 1;
 		}
-	} while (++url->ptr <= url->end);
+	} while (++state->ptr <= state->end);
 
 	return NULL;
 }
 
-static const char *parse_hier(php_http_url_t *url)
+static const char *parse_hier(struct parse_state *state)
 {
-	if (*url->ptr == '/') {
-		if (url->end - url->ptr > 1) {
-			if (*(url->ptr + 1) == '/') {
-				url->ptr += 2;
-				if (!(url->ptr = parse_authority(url))) {
+	if (*state->ptr == '/') {
+		if (state->end - state->ptr > 1) {
+			if (*(state->ptr + 1) == '/') {
+				state->ptr += 2;
+				if (!(state->ptr = parse_authority(state))) {
 					return NULL;
 				}
 			}
 		}
 	}
-	return parse_path(url);
+	return parse_path(state);
 }
 
-static const char *parse_scheme(php_http_url_t *url)
+static const char *parse_scheme(struct parse_state *state)
 {
 	size_t mb;
-	const char *tmp = url->ptr;
+	const char *tmp = state->ptr;
 
 	do {
-		switch (*url->ptr) {
+		switch (*state->ptr) {
 		case ':':
 			/* scheme delimiter */
-			url->scheme = &url->buffer[0];
-			url->buffer[url->offset++] = 0;
-			return ++url->ptr;
+			state->url.scheme = &state->buffer[0];
+			state->buffer[state->offset++] = 0;
+			return ++state->ptr;
 
 		case '0': case '1': case '2': case '3': case '4': case '5': case '6':
 		case '7': case '8': case '9':
 		case '+': case '-': case '.':
-			if (url->ptr == tmp) {
+			if (state->ptr == tmp) {
 				return tmp;
 			}
 			/* no break */
@@ -879,59 +928,56 @@ static const char *parse_scheme(php_http_url_t *url)
 		case 'o': case 'p': case 'q': case 'r': case 's': case 't': case 'u':
 		case 'v': case 'w': case 'x': case 'y': case 'z':
 			/* scheme part */
-			url->buffer[url->offset++] = *url->ptr;
+			state->buffer[state->offset++] = *state->ptr;
 			break;
 
 		default:
-			if (!(mb = parse_mb(url, PARSE_SCHEME, url->ptr, url->end, tmp, 1))) {
+			if (!(mb = parse_mb(state, PARSE_SCHEME, state->ptr, state->end, tmp, 1))) {
 				/* soft fail; parse path next */
 				return tmp;
 			}
-			url->ptr += mb - 1;
+			state->ptr += mb - 1;
 		}
-	} while (++url->ptr != url->end);
+	} while (++state->ptr != state->end);
 
 	return tmp;
 }
 
-struct parser_state {
-};
-
 php_http_url_t *php_http_url_parse(const char *str, size_t len, unsigned flags TSRMLS_DC)
 {
 	size_t maxlen = 3 * len;
-	php_http_url_t *url = ecalloc(1, sizeof(*url) + maxlen);
+	struct parse_state *state = ecalloc(1, sizeof(*state) + maxlen);
 
-	url->end = str + len;
-	url->ptr = str;
-	url->flags = flags;
-	url->maxlen = maxlen;
-	TSRMLS_SET_CTX(url->ts);
+	state->end = str + len;
+	state->ptr = str;
+	state->flags = flags;
+	state->maxlen = maxlen;
+	TSRMLS_SET_CTX(state->ts);
 
-	if (!parse_scheme(url)) {
-		php_error_docref(NULL TSRMLS_CC, E_WARNING, "Failed to parse URL scheme: '%s'", url->ptr);
-		php_http_url_free(&url);
+	if (!parse_scheme(state)) {
+		php_error_docref(NULL TSRMLS_CC, E_WARNING, "Failed to parse URL scheme: '%s'", state->ptr);
+		efree(state);
 		return NULL;
 	}
 
-	if (!parse_hier(url)) {
-		php_http_url_free(&url);
+	if (!parse_hier(state)) {
+		efree(state);
 		return NULL;
 	}
 
-	if (!parse_query(url)) {
-		php_error_docref(NULL TSRMLS_CC, E_WARNING, "Failed to parse URL query: '%s'", url->ptr);
-		php_http_url_free(&url);
+	if (!parse_query(state)) {
+		php_error_docref(NULL TSRMLS_CC, E_WARNING, "Failed to parse URL query: '%s'", state->ptr);
+		efree(state);
 		return NULL;
 	}
 
-	if (!parse_fragment(url)) {
-		php_error_docref(NULL TSRMLS_CC, E_WARNING, "Failed to parse URL fragment: '%s'", url->ptr);
-		php_http_url_free(&url);
+	if (!parse_fragment(state)) {
+		php_error_docref(NULL TSRMLS_CC, E_WARNING, "Failed to parse URL fragment: '%s'", state->ptr);
+		efree(state);
 		return NULL;
 	}
 
-	return url;
+	return (php_http_url_t *) state;
 }
 
 ZEND_BEGIN_ARG_INFO_EX(ai_HttpUrl___construct, 0, 0, 0)
@@ -1197,8 +1243,9 @@ PHP_MINIT_FUNCTION(http_url)
 #endif
 	zend_declare_class_constant_long(php_http_url_class_entry, ZEND_STRL("PARSE_MBUTF8"), PHP_HTTP_URL_PARSE_MBUTF8 TSRMLS_CC);
 #ifdef PHP_HTTP_HAVE_IDN
-	zend_declare_class_constant_long(php_http_url_class_entry, ZEND_STRL("PARSE_IDN"), PHP_HTTP_URL_PARSE_IDN TSRMLS_CC);
+	zend_declare_class_constant_long(php_http_url_class_entry, ZEND_STRL("PARSE_TOIDN"), PHP_HTTP_URL_PARSE_TOIDN TSRMLS_CC);
 #endif
+	zend_declare_class_constant_long(php_http_url_class_entry, ZEND_STRL("PARSE_TOPCT"), PHP_HTTP_URL_PARSE_TOPCT TSRMLS_CC);
 
 	return SUCCESS;
 }