+struct parse_state {
+ php_http_url_t url;
+ 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) {
+ efree(*url);
+ *url = NULL;
+ }
+}
+
+php_http_url_t *php_http_url_copy(const php_http_url_t *url, zend_bool persistent)
+{
+ php_http_url_t *cpy;
+ const char *end = NULL, *url_ptr = (const char *) url;
+ char *cpy_ptr;
+
+ end = MAX(url->scheme, end);
+ end = MAX(url->pass, end);
+ end = MAX(url->user, end);
+ end = MAX(url->host, end);
+ end = MAX(url->path, end);
+ end = MAX(url->query, end);
+ end = MAX(url->fragment, end);
+
+ if (end) {
+ end += strlen(end) + 1;
+ cpy_ptr = pecalloc(1, end - url_ptr, persistent);
+ cpy = (php_http_url_t *) cpy_ptr;
+
+ memcpy(cpy_ptr + sizeof(*cpy), url_ptr + sizeof(*url), end - url_ptr - sizeof(*url));
+
+ cpy->scheme = url->scheme ? cpy_ptr + (url->scheme - url_ptr) : NULL;
+ cpy->pass = url->pass ? cpy_ptr + (url->pass - url_ptr) : NULL;
+ cpy->user = url->user ? cpy_ptr + (url->user - url_ptr) : NULL;
+ cpy->host = url->host ? cpy_ptr + (url->host - url_ptr) : NULL;
+ cpy->path = url->path ? cpy_ptr + (url->path - url_ptr) : NULL;
+ cpy->query = url->query ? cpy_ptr + (url->query - url_ptr) : NULL;
+ cpy->fragment = url->fragment ? cpy_ptr + (url->fragment - url_ptr) : NULL;
+ } else {
+ cpy = ecalloc(1, sizeof(*url));
+ }
+
+ cpy->port = url->port;
+
+ return cpy;
+}
+
+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);
+
+ if (!consumed || consumed == (size_t) -1) {
+ return 0;
+ }
+
+ if (wc) {
+ *wc = wchar;
+ }
+ return consumed;
+}
+
+#ifdef PHP_HTTP_HAVE_WCHAR
+static size_t parse_mb_loc(unsigned *wc, const char *ptr, const char *end)
+{
+ wchar_t wchar;
+ size_t consumed = 0;
+#if defined(HAVE_MBRTOWC)
+ mbstate_t ps;
+
+ memset(&ps, 0, sizeof(ps));
+ consumed = mbrtowc(&wchar, ptr, end - ptr, &ps);
+#elif defined(HAVE_MBTOWC)
+ consumed = mbtowc(&wchar, ptr, end - ptr);
+#endif
+
+ if (!consumed || consumed == (size_t) -1) {
+ return 0;
+ }
+
+ if (wc) {
+ *wc = wchar;
+ }
+ return consumed;
+}
+#endif
+
+typedef enum parse_mb_what {
+ PARSE_SCHEME,
+ PARSE_USERINFO,
+ PARSE_HOSTINFO,
+ PARSE_PATH,
+ PARSE_QUERY,
+ PARSE_FRAGMENT
+} parse_mb_what_t;
+
+static const char * const parse_what[] = {
+ "scheme",
+ "userinfo",
+ "hostinfo",
+ "path",
+ "query",
+ "fragment"
+};
+
+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;
+
+ if (state->flags & PHP_HTTP_URL_PARSE_MBUTF8) {
+ consumed = parse_mb_utf8(&wchar, ptr, end);
+ }
+#ifdef PHP_HTTP_HAVE_WCHAR
+ else if (state->flags & PHP_HTTP_URL_PARSE_MBLOC) {
+ consumed = parse_mb_loc(&wchar, ptr, end);
+ }
+#endif
+
+ 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) {
+ if (consumed) {
+ php_error_docref(NULL, E_WARNING,
+ "Failed to parse %s; unexpected multibyte sequence 0x%x at pos %u in '%s'",
+ parse_what[what], wchar, (unsigned) (ptr - begin), begin);
+ } else {
+ php_error_docref(NULL, 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 0;
+}
+
+static ZEND_RESULT_CODE parse_userinfo(struct parse_state *state, const char *ptr)
+{
+ size_t mb;
+ const char *password = NULL, *end = state->ptr, *tmp = ptr;
+
+ state->url.user = &state->buffer[state->offset];
+
+ do {
+ switch (*ptr) {
+ case ':':
+ if (password) {
+ php_error_docref(NULL, E_WARNING,
+ "Failed to parse password; duplicate ':' at pos %u in '%s'",
+ (unsigned) (ptr - tmp), tmp);
+ return FAILURE;
+ }
+ password = ptr + 1;
+ state->buffer[state->offset++] = 0;
+ state->url.pass = &state->buffer[state->offset];
+ break;
+
+ case '%':
+ if (ptr[1] != '%' && (end - ptr <= 2 || !isxdigit(*(ptr+1)) || !isxdigit(*(ptr+2)))) {
+ php_error_docref(NULL, E_WARNING,
+ "Failed to parse userinfo; invalid percent encoding at pos %u in '%s'",
+ (unsigned) (ptr - tmp), tmp);
+ return FAILURE;
+ }
+ state->buffer[state->offset++] = *ptr++;
+ state->buffer[state->offset++] = *ptr++;
+ state->buffer[state->offset++] = *ptr;
+ break;
+
+ case '!': case '$': case '&': case '\'': case '(': case ')': case '*':
+ case '+': case ',': case ';': case '=': /* sub-delims */
+ case '-': case '.': case '_': case '~': /* unreserved */
+ case 'A': case 'B': case 'C': case 'D': case 'E': case 'F': case 'G':
+ case 'H': case 'I': case 'J': case 'K': case 'L': case 'M': case 'N':
+ case 'O': case 'P': case 'Q': case 'R': case 'S': case 'T': case 'U':
+ case 'V': case 'W': case 'X': case 'Y': case 'Z':
+ case 'a': case 'b': case 'c': case 'd': case 'e': case 'f': case 'g':
+ case 'h': case 'i': case 'j': case 'k': case 'l': case 'm': case 'n':
+ case 'o': case 'p': case 'q': case 'r': case 's': case 't': case 'u':
+ case 'v': case 'w': case 'x': case 'y': case 'z':
+ case '0': case '1': case '2': case '3': case '4': case '5': case '6':
+ case '7': case '8': case '9':
+ /* allowed */
+ state->buffer[state->offset++] = *ptr;
+ break;
+
+ default:
+ if (!(mb = parse_mb(state, PARSE_USERINFO, ptr, end, tmp, 0))) {
+ return FAILURE;
+ }
+ ptr += mb - 1;
+ }
+ } while(++ptr != end);
+
+
+ state->buffer[state->offset++] = 0;
+
+ return SUCCESS;
+}
+
+#if defined(PHP_WIN32) || defined(HAVE_UIDNA_IDNTOASCII)
+typedef size_t (*parse_mb_func)(unsigned *wc, const char *ptr, const char *end);
+static ZEND_RESULT_CODE to_utf16(parse_mb_func fn, const char *u8, uint16_t **u16, size_t *len)
+{
+ size_t offset = 0, u8_len = strlen(u8);
+
+ *u16 = ecalloc(4 * sizeof(uint16_t), u8_len + 1);
+ *len = 0;
+
+ while (offset < u8_len) {
+ unsigned wc;
+ uint16_t buf[2], *ptr = buf;
+ size_t consumed = fn(&wc, &u8[offset], &u8[u8_len]);
+
+ if (!consumed) {
+ efree(*u16);
+ php_error_docref(NULL, E_WARNING, "Failed to parse UTF-8 at pos %zu of '%s'", offset, u8);
+ return FAILURE;
+ } else {
+ offset += consumed;
+ }
+
+ switch (wctoutf16(buf, wc)) {
+ case 2:
+ (*u16)[(*len)++] = *ptr++;
+ /* no break */
+ case 1:
+ (*u16)[(*len)++] = *ptr++;
+ break;
+ case 0:
+ default:
+ efree(*u16);
+ php_error_docref(NULL, E_WARNING, "Failed to convert UTF-32 'U+%X' to UTF-16", wc);
+ return FAILURE;
+ }
+ }
+
+ return SUCCESS;
+}
+#endif
+
+#ifndef MAXHOSTNAMELEN
+# define MAXHOSTNAMELEN 256
+#endif
+
+#if PHP_HTTP_HAVE_IDN2
+static ZEND_RESULT_CODE parse_idn2(struct parse_state *state, size_t prev_len)
+{
+ char *idn = NULL;
+ int rv = -1;
+
+ if (state->flags & PHP_HTTP_URL_PARSE_MBUTF8) {
+ rv = idn2_lookup_u8((const unsigned char *) state->url.host, (unsigned char **) &idn, IDN2_NFC_INPUT);
+ }
+# ifdef PHP_HTTP_HAVE_WCHAR
+ else if (state->flags & PHP_HTTP_URL_PARSE_MBLOC) {
+ rv = idn2_lookup_ul(state->url.host, &idn, 0);
+ }
+# endif
+ if (rv != IDN2_OK) {
+ php_error_docref(NULL, E_WARNING, "Failed to parse IDN; %s", idn2_strerror(rv));
+ return FAILURE;
+ } else {
+ size_t idnlen = strlen(idn);
+ memcpy(state->url.host, idn, idnlen + 1);
+ free(idn);
+ state->offset += idnlen - prev_len;
+ return SUCCESS;
+ }
+}
+#elif PHP_HTTP_HAVE_IDN
+static ZEND_RESULT_CODE parse_idn(struct parse_state *state, size_t prev_len)
+{
+ char *idn = NULL;
+ int rv = -1;
+
+ 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 (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) {
+ php_error_docref(NULL, E_WARNING, "Failed to parse IDN; %s", idna_strerror(rv));
+ return FAILURE;
+ } else {
+ size_t idnlen = strlen(idn);
+ memcpy(state->url.host, idn, idnlen + 1);
+ free(idn);
+ state->offset += idnlen - prev_len;
+ return SUCCESS;
+ }
+}
+#endif
+
+#ifdef HAVE_UIDNA_IDNTOASCII
+# if HAVE_UNICODE_UIDNA_H
+# include <unicode/uidna.h>
+# else
+typedef uint16_t UChar;
+typedef enum { U_ZERO_ERROR = 0 } UErrorCode;
+int32_t uidna_IDNToASCII(const UChar *src, int32_t srcLength, UChar *dest, int32_t destCapacity, int32_t options, void *parseError, UErrorCode *status);
+# endif
+static ZEND_RESULT_CODE parse_uidn(struct parse_state *state)
+{
+ char *host_ptr;
+ uint16_t *uhost_str, ahost_str[MAXHOSTNAMELEN], *ahost_ptr;
+ size_t uhost_len, ahost_len;
+ UErrorCode error = U_ZERO_ERROR;
+
+ if (state->flags & PHP_HTTP_URL_PARSE_MBUTF8) {
+ if (SUCCESS != to_utf16(parse_mb_utf8, state->url.host, &uhost_str, &uhost_len)) {
+ return FAILURE;
+ }
+#ifdef PHP_HTTP_HAVE_WCHAR
+ } else if (state->flags & PHP_HTTP_URL_PARSE_MBLOC) {
+ if (SUCCESS != to_utf16(parse_mb_loc, state->url.host, &uhost_str, &uhost_len)) {
+ return FAILURE;
+ }
+#endif
+ } else {
+ php_error_docref(NULL, E_WARNING, "Failed to parse IDN; codepage not specified");
+ return FAILURE;
+ }
+
+ ahost_len = uidna_IDNToASCII(uhost_str, uhost_len, ahost_str, MAXHOSTNAMELEN, 3, NULL, &error);
+ efree(uhost_str);
+
+ if (error != U_ZERO_ERROR) {
+ php_error_docref(NULL, E_WARNING, "Failed to parse IDN; ICU error %d", error);
+ return FAILURE;
+ }
+
+ host_ptr = state->url.host;
+ ahost_ptr = ahost_str;
+ PHP_HTTP_DUFF(ahost_len, *host_ptr++ = *ahost_ptr++);
+
+ *host_ptr = '\0';
+ state->offset += host_ptr - state->url.host;
+
+ return SUCCESS;
+}
+#endif
+
+#if 0 && defined(PHP_WIN32)
+static ZEND_RESULT_CODE parse_widn(struct parse_state *state)
+{
+ char *host_ptr;
+ uint16_t *uhost_str, ahost_str[MAXHOSTNAMELEN], *ahost_ptr;
+ size_t uhost_len;
+
+ if (state->flags & PHP_HTTP_URL_PARSE_MBUTF8) {
+ if (SUCCESS != to_utf16(parse_mb_utf8, state->url.host, &uhost_str, &uhost_len)) {
+ php_error_docref(NULL, E_WARNING, "Failed to parse IDN");
+ return FAILURE;
+ }
+#ifdef PHP_HTTP_HAVE_WCHAR
+ } else if (state->flags & PHP_HTTP_URL_PARSE_MBLOC) {
+ if (SUCCESS != to_utf16(parse_mb_loc, state->url.host, &uhost_str, &uhost_len)) {
+ php_error_docref(NULL, E_WARNING, "Failed to parse IDN");
+ return FAILURE;
+ }
+#endif
+ } else {
+ php_error_docref(NULL, E_WARNING, "Failed to parse IDN");
+ return FAILURE;
+ }
+
+ if (!IdnToAscii(IDN_ALLOW_UNASSIGNED|IDN_USE_STD3_ASCII_RULES, uhost_str, uhost_len, ahost_str, MAXHOSTNAMELEN)) {
+ efree(uhost_str);
+ php_error_docref(NULL, E_WARNING, "Failed to parse IDN");
+ return FAILURE;
+ }
+
+ efree(uhost_str);
+ host_ptr = state->url.host;
+ ahost_ptr = ahost_str;
+ PHP_HTTP_DUFF(wcslen(ahost_str), *host_ptr++ = *ahost_ptr++);
+ efree(ahost_str);
+
+ *host_ptr = '\0';
+ state->offset += host_ptr - state->url.host;
+
+ return SUCCESS;
+}
+#endif
+
+static ZEND_RESULT_CODE parse_hostinfo(struct parse_state *state, const char *ptr)
+{
+ size_t mb, len;
+ const char *end = state->ptr, *tmp = ptr, *port = NULL;
+
+#ifdef HAVE_INET_PTON
+ if (*ptr == '[') {
+ char *error = NULL, *tmp = memchr(ptr, ']', end - ptr);
+
+ if (tmp) {
+ size_t addrlen = tmp - ptr + 1;
+ char buf[16], *addr = estrndup(ptr + 1, addrlen - 2);
+ int rv = inet_pton(AF_INET6, addr, buf);
+
+ efree(addr);
+ if (rv == 1) {
+ 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);
+ } else {
+ error = "unexpected '['";
+ }
+ } else {
+ error = "expected ']'";
+ }
+
+ if (error) {
+ php_error_docref(NULL, E_WARNING, "Failed to parse hostinfo; %s", error);
+ return FAILURE;
+ }
+ }
+#endif
+ if (ptr != end) do {
+ switch (*ptr) {
+ case ':':
+ if (port) {
+ php_error_docref(NULL, E_WARNING,
+ "Failed to parse port; unexpected ':' at pos %u in '%s'",
+ (unsigned) (ptr - tmp), tmp);
+ return FAILURE;
+ }
+ port = ptr + 1;
+ break;
+
+ case '%':
+ if (ptr[1] != '%' && (end - ptr <= 2 || !isxdigit(*(ptr+1)) || !isxdigit(*(ptr+2)))) {
+ php_error_docref(NULL, E_WARNING,
+ "Failed to parse hostinfo; invalid percent encoding at pos %u in '%s'",
+ (unsigned) (ptr - tmp), tmp);
+ return FAILURE;
+ }
+ state->buffer[state->offset++] = *ptr++;
+ state->buffer[state->offset++] = *ptr++;
+ state->buffer[state->offset++] = *ptr;
+ break;
+
+ case '!': case '$': case '&': case '\'': case '(': case ')': case '*':
+ case '+': case ',': case ';': case '=': /* sub-delims */
+ case '-': case '.': case '_': case '~': /* unreserved */
+ case 'A': case 'B': case 'C': case 'D': case 'E': case 'F': case 'G':
+ case 'H': case 'I': case 'J': case 'K': case 'L': case 'M': case 'N':
+ case 'O': case 'P': case 'Q': case 'R': case 'S': case 'T': case 'U':
+ case 'V': case 'W': case 'X': case 'Y': case 'Z':
+ case 'a': case 'b': case 'c': case 'd': case 'e': case 'f': case 'g':
+ case 'h': case 'i': case 'j': case 'k': case 'l': case 'm': case 'n':
+ case 'o': case 'p': case 'q': case 'r': case 's': case 't': case 'u':
+ case 'v': case 'w': case 'x': case 'y': case 'z':
+ if (port) {
+ php_error_docref(NULL, E_WARNING,
+ "Failed to parse port; unexpected char '%c' at pos %u in '%s'",
+ (unsigned char) *ptr, (unsigned) (ptr - tmp), tmp);
+ return FAILURE;
+ }
+ /* no break */
+ case '0': case '1': case '2': case '3': case '4': case '5': case '6':
+ case '7': case '8': case '9':
+ /* allowed */
+ if (port) {
+ state->url.port *= 10;
+ state->url.port += *ptr - '0';
+ } else {
+ state->buffer[state->offset++] = *ptr;
+ }
+ break;
+
+ default:
+ if (ptr == end) {
+ break;
+ } else if (port) {
+ php_error_docref(NULL, E_WARNING,
+ "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(state, PARSE_HOSTINFO, ptr, end, tmp, 0))) {
+ return FAILURE;
+ }
+ ptr += mb - 1;
+ }
+ } while (++ptr != end);
+
+ if (!state->url.host) {
+ len = (port ? port - tmp - 1 : end - tmp);
+ state->url.host = &state->buffer[state->offset - len];
+ state->buffer[state->offset++] = 0;
+ }
+
+ if (state->flags & PHP_HTTP_URL_PARSE_TOIDN) {
+#if PHP_HTTP_HAVE_IDN2
+ return parse_idn2(state, len);
+#elif PHP_HTTP_HAVE_IDN
+ return parse_idn(state, len);
+#endif
+#ifdef HAVE_UIDNA_IDNTOASCII
+ return parse_uidn(state);
+#endif
+#if 0 && defined(PHP_WIN32)
+ return parse_widn(state);
+#endif
+ }
+
+ return SUCCESS;
+}
+
+static const char *parse_authority(struct parse_state *state)
+{
+ const char *tmp = state->ptr, *host = NULL;
+
+ do {
+ switch (*state->ptr) {
+ case '@':
+ /* userinfo delimiter */
+ if (host) {
+ php_error_docref(NULL, 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 = state->ptr + 1;
+ break;
+
+ case '/':
+ case '?':
+ case '#':
+ case '\0':
+ EOD:
+ /* host delimiter */
+ if (tmp != state->ptr && SUCCESS != parse_hostinfo(state, tmp)) {
+ return NULL;
+ }
+ return state->ptr;
+ }
+ } while (++state->ptr <= state->end);
+
+ --state->ptr;
+ goto EOD;
+}
+
+static const char *parse_path(struct parse_state *state)
+{
+ size_t mb;
+ const char *tmp;
+
+ /* is there actually a path to parse? */
+ if (!*state->ptr) {
+ return state->ptr;
+ }
+ tmp = state->ptr;
+ state->url.path = &state->buffer[state->offset];
+
+ do {
+ switch (*state->ptr) {
+ case '#':
+ case '?':
+ goto done;
+
+ case '%':
+ if (state->ptr[1] != '%' && (state->end - state->ptr <= 2 || !isxdigit(*(state->ptr+1)) || !isxdigit(*(state->ptr+2)))) {
+ php_error_docref(NULL, E_WARNING,
+ "Failed to parse path; invalid percent encoding at pos %u in '%s'",
+ (unsigned) (state->ptr - tmp), tmp);
+ return NULL;
+ }
+ state->buffer[state->offset++] = *state->ptr++;
+ state->buffer[state->offset++] = *state->ptr++;
+ state->buffer[state->offset++] = *state->ptr;
+ break;
+
+ case '/': /* yeah, well */
+ case '!': case '$': case '&': case '\'': case '(': case ')': case '*':
+ case '+': case ',': case ';': case '=': /* sub-delims */
+ case '-': case '.': case '_': case '~': /* unreserved */
+ case ':': case '@': /* pchar */
+ case 'A': case 'B': case 'C': case 'D': case 'E': case 'F': case 'G':
+ case 'H': case 'I': case 'J': case 'K': case 'L': case 'M': case 'N':
+ case 'O': case 'P': case 'Q': case 'R': case 'S': case 'T': case 'U':
+ case 'V': case 'W': case 'X': case 'Y': case 'Z':
+ case 'a': case 'b': case 'c': case 'd': case 'e': case 'f': case 'g':
+ case 'h': case 'i': case 'j': case 'k': case 'l': case 'm': case 'n':
+ case 'o': case 'p': case 'q': case 'r': case 's': case 't': case 'u':
+ case 'v': case 'w': case 'x': case 'y': case 'z':
+ case '0': case '1': case '2': case '3': case '4': case '5': case '6':
+ case '7': case '8': case '9':
+ /* allowed */
+ state->buffer[state->offset++] = *state->ptr;
+ break;
+
+ default:
+ if (!(mb = parse_mb(state, PARSE_PATH, state->ptr, state->end, tmp, 0))) {
+ return NULL;
+ }
+ state->ptr += mb - 1;
+ }
+ } while (++state->ptr < state->end);
+
+ done:
+ /* did we have any path component ? */
+ if (tmp != state->ptr) {
+ state->buffer[state->offset++] = 0;
+ } else {
+ state->url.path = NULL;
+ }
+ return state->ptr;
+}
+
+static const char *parse_query(struct parse_state *state)
+{
+ size_t mb;
+ const char *tmp = state->ptr + !!*state->ptr;
+
+ /* is there actually a query to parse? */
+ if (*state->ptr != '?') {
+ return state->ptr;
+ }
+
+ /* skip initial '?' */
+ tmp = ++state->ptr;
+ state->url.query = &state->buffer[state->offset];
+
+ while (state->ptr < state->end) {
+ switch (*state->ptr) {
+ case '#':
+ goto done;
+
+ case '%':
+ if (state->ptr[1] != '%' && (state->end - state->ptr <= 2 || !isxdigit(*(state->ptr+1)) || !isxdigit(*(state->ptr+2)))) {
+ php_error_docref(NULL, E_WARNING,
+ "Failed to parse query; invalid percent encoding at pos %u in '%s'",
+ (unsigned) (state->ptr - tmp), tmp);
+ return NULL;
+ }
+ state->buffer[state->offset++] = *state->ptr++;
+ state->buffer[state->offset++] = *state->ptr++;
+ state->buffer[state->offset++] = *state->ptr;
+ break;
+
+ /* RFC1738 unsafe */
+ case '{': case '}':
+ case '<': case '>':
+ case '[': case ']':
+ case '|': case '\\': case '^': case '`': case '"': case ' ':
+ if (state->flags & PHP_HTTP_URL_PARSE_TOPCT) {
+ state->buffer[state->offset++] = '%';
+ state->buffer[state->offset++] = parse_xdigits[((unsigned char) *state->ptr) >> 4];
+ state->buffer[state->offset++] = parse_xdigits[((unsigned char) *state->ptr) & 0xf];
+ break;
+ }
+ /* no break */
+
+ case '?': case '/': /* yeah, well */
+ case '!': case '$': case '&': case '\'': case '(': case ')': case '*':
+ case '+': case ',': case ';': case '=': /* sub-delims */
+ case '-': case '.': case '_': case '~': /* unreserved */
+ case ':': case '@': /* pchar */
+ case 'A': case 'B': case 'C': case 'D': case 'E': case 'F': case 'G':
+ case 'H': case 'I': case 'J': case 'K': case 'L': case 'M': case 'N':
+ case 'O': case 'P': case 'Q': case 'R': case 'S': case 'T': case 'U':
+ case 'V': case 'W': case 'X': case 'Y': case 'Z':
+ case 'a': case 'b': case 'c': case 'd': case 'e': case 'f': case 'g':
+ case 'h': case 'i': case 'j': case 'k': case 'l': case 'm': case 'n':
+ case 'o': case 'p': case 'q': case 'r': case 's': case 't': case 'u':
+ case 'v': case 'w': case 'x': case 'y': case 'z':
+ case '0': case '1': case '2': case '3': case '4': case '5': case '6':
+ case '7': case '8': case '9':
+ /* allowed */
+ state->buffer[state->offset++] = *state->ptr;
+ break;
+
+ default:
+ if (!(mb = parse_mb(state, PARSE_QUERY, state->ptr, state->end, tmp, 0))) {
+ return NULL;
+ }
+ state->ptr += mb - 1;
+ }
+
+ ++state->ptr;
+ }
+
+ done:
+ state->buffer[state->offset++] = 0;
+ return state->ptr;
+}
+
+static const char *parse_fragment(struct parse_state *state)
+{
+ size_t mb;
+ const char *tmp;
+
+ /* is there actually a fragment to parse? */
+ if (*state->ptr != '#') {
+ return state->ptr;
+ }
+
+ /* skip initial '#' */
+ tmp = ++state->ptr;
+ state->url.fragment = &state->buffer[state->offset];
+
+ do {
+ switch (*state->ptr) {
+ case '%':
+ if (state->ptr[1] != '%' && (state->end - state->ptr <= 2 || !isxdigit(*(state->ptr+1)) || !isxdigit(*(state->ptr+2)))) {
+ php_error_docref(NULL, E_WARNING,
+ "Failed to parse fragment; invalid percent encoding at pos %u in '%s'",
+ (unsigned) (state->ptr - tmp), tmp);
+ return NULL;
+ }
+ state->buffer[state->offset++] = *state->ptr++;
+ state->buffer[state->offset++] = *state->ptr++;
+ state->buffer[state->offset++] = *state->ptr;
+ break;
+
+ /* RFC1738 unsafe */
+ case '{': case '}':
+ case '<': case '>':
+ case '[': case ']':
+ case '|': case '\\': case '^': case '`': case '"': case ' ':
+ if (state->flags & PHP_HTTP_URL_PARSE_TOPCT) {
+ state->buffer[state->offset++] = '%';
+ state->buffer[state->offset++] = parse_xdigits[((unsigned char) *state->ptr) >> 4];
+ state->buffer[state->offset++] = parse_xdigits[((unsigned char) *state->ptr) & 0xf];
+ break;
+ }
+ /* no break */
+
+ case '?': case '/':
+ case '!': case '$': case '&': case '\'': case '(': case ')': case '*':
+ case '+': case ',': case ';': case '=': /* sub-delims */
+ case '-': case '.': case '_': case '~': /* unreserved */
+ case ':': case '@': /* pchar */
+ case 'A': case 'B': case 'C': case 'D': case 'E': case 'F': case 'G':
+ case 'H': case 'I': case 'J': case 'K': case 'L': case 'M': case 'N':
+ case 'O': case 'P': case 'Q': case 'R': case 'S': case 'T': case 'U':
+ case 'V': case 'W': case 'X': case 'Y': case 'Z':
+ case 'a': case 'b': case 'c': case 'd': case 'e': case 'f': case 'g':
+ case 'h': case 'i': case 'j': case 'k': case 'l': case 'm': case 'n':
+ case 'o': case 'p': case 'q': case 'r': case 's': case 't': case 'u':
+ case 'v': case 'w': case 'x': case 'y': case 'z':
+ case '0': case '1': case '2': case '3': case '4': case '5': case '6':
+ case '7': case '8': case '9':
+ /* allowed */
+ state->buffer[state->offset++] = *state->ptr;
+ break;
+
+ default:
+ if (!(mb = parse_mb(state, PARSE_FRAGMENT, state->ptr, state->end, tmp, 0))) {
+ return NULL;
+ }
+ state->ptr += mb - 1;
+ }
+ } while (++state->ptr < state->end);
+
+ state->buffer[state->offset++] = 0;
+ return state->ptr;
+}
+
+static const char *parse_hier(struct parse_state *state)
+{
+ 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(state);
+}
+
+static const char *parse_scheme(struct parse_state *state)
+{
+ size_t mb;
+ const char *tmp = state->ptr;
+
+ do {
+ switch (*state->ptr) {
+ case ':':
+ /* scheme delimiter */
+ 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 (state->ptr == tmp) {
+ return tmp;
+ }
+ /* no break */
+ case 'A': case 'B': case 'C': case 'D': case 'E': case 'F': case 'G':
+ case 'H': case 'I': case 'J': case 'K': case 'L': case 'M': case 'N':
+ case 'O': case 'P': case 'Q': case 'R': case 'S': case 'T': case 'U':
+ case 'V': case 'W': case 'X': case 'Y': case 'Z':
+ case 'a': case 'b': case 'c': case 'd': case 'e': case 'f': case 'g':
+ case 'h': case 'i': case 'j': case 'k': case 'l': case 'm': case 'n':
+ 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 */
+ state->buffer[state->offset++] = *state->ptr;
+ break;
+
+ default:
+ if (!(mb = parse_mb(state, PARSE_SCHEME, state->ptr, state->end, tmp, 1))) {
+ /* soft fail; parse path next */
+ return tmp;
+ }
+ state->ptr += mb - 1;
+ }
+ } while (++state->ptr != state->end);
+
+ return state->ptr = tmp;
+}
+
+php_http_url_t *php_http_url_parse(const char *str, size_t len, unsigned flags)
+{
+ size_t maxlen = 3 * len;
+ struct parse_state *state = ecalloc(1, sizeof(*state) + maxlen);
+
+ state->end = str + len;
+ state->ptr = str;
+ state->flags = flags;
+ state->maxlen = maxlen;
+
+ if (!parse_scheme(state)) {
+ php_error_docref(NULL, E_WARNING, "Failed to parse URL scheme: '%s'", state->ptr);
+ efree(state);
+ return NULL;
+ }
+
+ if (!parse_hier(state)) {
+ efree(state);
+ return NULL;
+ }
+
+ if (!parse_query(state)) {
+ php_error_docref(NULL, E_WARNING, "Failed to parse URL query: '%s'", state->ptr);
+ efree(state);
+ return NULL;
+ }
+
+ if (!parse_fragment(state)) {
+ php_error_docref(NULL, E_WARNING, "Failed to parse URL fragment: '%s'", state->ptr);
+ efree(state);
+ return NULL;
+ }
+
+ return (php_http_url_t *) state;
+}
+
+php_http_url_t *php_http_url_parse_authority(const char *str, size_t len, unsigned flags)
+{
+ size_t maxlen = 3 * len;
+ struct parse_state *state = ecalloc(1, sizeof(*state) + maxlen);
+
+ state->end = str + len;
+ state->ptr = str;
+ state->flags = flags;
+ state->maxlen = maxlen;
+
+ if (!(state->ptr = parse_authority(state))) {
+ efree(state);
+ return NULL;
+ }
+
+ if (state->ptr != state->end) {
+ php_error_docref(NULL, E_WARNING,
+ "Failed to parse URL authority, unexpected character at pos %u in '%s'",
+ (unsigned) (state->ptr - str), str);
+ efree(state);
+ return NULL;
+ }
+
+ return (php_http_url_t *) state;
+}
+