return estrndup("localhost", lenof("localhost"));
}
-static php_url *php_http_url_from_env(php_url *url TSRMLS_DC)
+#define url(buf) ((php_http_url_t *) (buf).data)
+
+static php_http_url_t *php_http_url_from_env(TSRMLS_D)
{
zval *https, *zhost, *zport;
long port;
+ php_http_buffer_t buf;
- if (!url) {
- url = ecalloc(1, sizeof(*url));
- }
-
- /* port */
- zport = php_http_env_get_server_var(ZEND_STRL("SERVER_PORT"), 1 TSRMLS_CC);
- if (zport && IS_LONG == is_numeric_string(Z_STRVAL_P(zport), Z_STRLEN_P(zport), &port, NULL, 0)) {
- url->port = port;
- }
+ php_http_buffer_init_ex(&buf, MAX(PHP_HTTP_BUFFER_DEFAULT_SIZE, sizeof(php_http_url_t)<<2), PHP_HTTP_BUFFER_INIT_PREALLOC);
+ php_http_buffer_account(&buf, sizeof(php_http_url_t));
+ memset(buf.data, 0, buf.used);
/* scheme */
+ url(buf)->scheme = &buf.data[buf.used];
https = php_http_env_get_server_var(ZEND_STRL("HTTPS"), 1 TSRMLS_CC);
if (https && !strcasecmp(Z_STRVAL_P(https), "ON")) {
- url->scheme = estrndup("https", lenof("https"));
+ php_http_buffer_append(&buf, "https", sizeof("https"));
} else {
- url->scheme = estrndup("http", lenof("http"));
+ php_http_buffer_append(&buf, "http", sizeof("http"));
}
/* host */
+ url(buf)->host = &buf.data[buf.used];
if ((((zhost = php_http_env_get_server_var(ZEND_STRL("HTTP_HOST"), 1 TSRMLS_CC)) ||
(zhost = php_http_env_get_server_var(ZEND_STRL("SERVER_NAME"), 1 TSRMLS_CC)) ||
(zhost = php_http_env_get_server_var(ZEND_STRL("SERVER_ADDR"), 1 TSRMLS_CC)))) && Z_STRLEN_P(zhost)) {
size_t stop_at = strspn(Z_STRVAL_P(zhost), "0123456789abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ-.");
- url->host = estrndup(Z_STRVAL_P(zhost), stop_at);
+ php_http_buffer_append(&buf, Z_STRVAL_P(zhost), stop_at);
+ php_http_buffer_append(&buf, "", 1);
} else {
- url->host = localhostname();
+ char *host_str = localhostname();
+
+ php_http_buffer_append(&buf, host_str, strlen(host_str) + 1);
+ efree(host_str);
+ }
+
+ /* port */
+ zport = php_http_env_get_server_var(ZEND_STRL("SERVER_PORT"), 1 TSRMLS_CC);
+ if (zport && IS_LONG == is_numeric_string(Z_STRVAL_P(zport), Z_STRLEN_P(zport), &port, NULL, 0)) {
+ url(buf)->port = port;
}
/* path */
if (SG(request_info).request_uri && SG(request_info).request_uri[0]) {
const char *q = strchr(SG(request_info).request_uri, '?');
+ url(buf)->path = &buf.data[buf.used];
+
if (q) {
- url->path = estrndup(SG(request_info).request_uri, q - SG(request_info).request_uri);
+ php_http_buffer_append(&buf, SG(request_info).request_uri, q - SG(request_info).request_uri);
+ php_http_buffer_append(&buf, "", 1);
} else {
- url->path = estrdup(SG(request_info).request_uri);
+ php_http_buffer_append(&buf, SG(request_info).request_uri, strlen(SG(request_info).request_uri) + 1);
}
}
/* query */
if (SG(request_info).query_string && SG(request_info).query_string[0]) {
- url->query = estrdup(SG(request_info).query_string);
+ url(buf)->query = &buf.data[buf.used];
+ php_http_buffer_append(&buf, SG(request_info).query_string, strlen(SG(request_info).query_string) + 1);
}
- return url;
+ return url(buf);
}
-void php_http_url(int flags, const php_url *old_url, const php_url *new_url, php_url **url_ptr, char **url_str, size_t *url_len TSRMLS_DC)
+#define url_isset(u,n) \
+ ((u)&&(u)->n)
+#define url_append(buf, append) do { \
+ char *_ptr = (buf)->data; \
+ php_http_url_t *_url = (php_http_url_t *) _ptr, _mem = *_url; \
+ append; \
+ /* relocate */ \
+ if (_ptr != (buf)->data) { \
+ ptrdiff_t diff = (buf)->data - _ptr; \
+ _url = (php_http_url_t *) (buf)->data; \
+ if (_mem.scheme) _url->scheme += diff; \
+ if (_mem.user) _url->user += diff; \
+ if (_mem.pass) _url->pass += diff; \
+ if (_mem.host) _url->host += diff; \
+ if (_mem.path) _url->path += diff; \
+ if (_mem.query) _url->query += diff; \
+ if (_mem.fragment) _url->fragment += diff; \
+ } \
+} while (0)
+#define url_copy(n) do { \
+ if (url_isset(new_url, n)) { \
+ url(buf)->n = &buf.data[buf.used]; \
+ url_append(&buf, php_http_buffer_append(&buf, new_url->n, strlen(new_url->n) + 1)); \
+ } else if (url_isset(old_url, n)) { \
+ url(buf)->n = &buf.data[buf.used]; \
+ url_append(&buf, php_http_buffer_append(&buf, old_url->n, strlen(old_url->n) + 1)); \
+ } \
+} while (0)
+
+php_http_url_t *php_http_url_mod(const php_http_url_t *old_url, const php_http_url_t *new_url, unsigned flags TSRMLS_DC)
{
- php_url *url, *tmp_url = NULL;
+ php_http_url_t *tmp_url = NULL;
+ php_http_buffer_t buf;
+
+ php_http_buffer_init_ex(&buf, MAX(PHP_HTTP_BUFFER_DEFAULT_SIZE, sizeof(php_http_url_t)<<2), PHP_HTTP_BUFFER_INIT_PREALLOC);
+ php_http_buffer_account(&buf, sizeof(php_http_url_t));
+ memset(buf.data, 0, buf.used);
/* set from env if requested */
if (flags & PHP_HTTP_URL_FROM_ENV) {
- php_url *env_url = php_http_url_from_env(NULL TSRMLS_CC);
-
- php_http_url(flags ^ PHP_HTTP_URL_FROM_ENV, env_url, old_url, &tmp_url, NULL, NULL TSRMLS_CC);
+ php_http_url_t *env_url = php_http_url_from_env(TSRMLS_C);
- php_url_free(env_url);
- old_url = tmp_url;
+ old_url = tmp_url = php_http_url_mod(env_url, old_url, flags ^ PHP_HTTP_URL_FROM_ENV TSRMLS_CC);
+ php_http_url_free(&env_url);
}
- url = ecalloc(1, sizeof(*url));
+ url_copy(scheme);
-#define __URLSET(u,n) \
- ((u)&&(u)->n)
-#define __URLCPY(n) \
- url->n = __URLSET(new_url,n) ? estrdup(new_url->n) : (__URLSET(old_url,n) ? estrdup(old_url->n) : NULL)
-
- if (!(flags & PHP_HTTP_URL_STRIP_PORT)) {
- url->port = __URLSET(new_url, port) ? new_url->port : ((old_url) ? old_url->port : 0);
- }
if (!(flags & PHP_HTTP_URL_STRIP_USER)) {
- __URLCPY(user);
+ url_copy(user);
}
+
if (!(flags & PHP_HTTP_URL_STRIP_PASS)) {
- __URLCPY(pass);
+ url_copy(pass);
}
- __URLCPY(scheme);
- __URLCPY(host);
+ url_copy(host);
+ if (!(flags & PHP_HTTP_URL_STRIP_PORT)) {
+ url(buf)->port = url_isset(new_url, port) ? new_url->port : ((old_url) ? old_url->port : 0);
+ }
+
if (!(flags & PHP_HTTP_URL_STRIP_PATH)) {
- if ((flags & PHP_HTTP_URL_JOIN_PATH) && __URLSET(old_url, path) && __URLSET(new_url, path) && *new_url->path != '/') {
+ if ((flags & PHP_HTTP_URL_JOIN_PATH) && url_isset(old_url, path) && url_isset(new_url, path) && *new_url->path != '/') {
size_t old_path_len = strlen(old_url->path), new_path_len = strlen(new_url->path);
+ char *path = ecalloc(1, old_path_len + new_path_len + 1 + 1);
- url->path = ecalloc(1, old_path_len + new_path_len + 1 + 1);
+ strcat(path, old_url->path);
+ if (path[old_path_len - 1] != '/') {
+ php_dirname(path, old_path_len);
+ strcat(path, "/");
+ }
+ strcat(path, new_url->path);
- strcat(url->path, old_url->path);
- if (url->path[old_path_len - 1] != '/') {
- php_dirname(url->path, old_path_len);
- strcat(url->path, "/");
+ url(buf)->path = &buf.data[buf.used];
+ if (path[0] != '/') {
+ url_append(&buf, php_http_buffer_append(&buf, "/", 1));
}
- strcat(url->path, new_url->path);
+ url_append(&buf, php_http_buffer_append(&buf, path, strlen(path) + 1));
+ efree(path);
} else {
- __URLCPY(path);
+ const char *path = NULL;
+
+ if (url_isset(new_url, path)) {
+ path = new_url->path;
+ } else if (url_isset(old_url, path)) {
+ path = old_url->path;
+ }
+
+ if (path) {
+ url(buf)->path = &buf.data[buf.used];
+
+ url_append(&buf, php_http_buffer_append(&buf, path, strlen(path) + 1));
+ }
+
+
}
}
+
if (!(flags & PHP_HTTP_URL_STRIP_QUERY)) {
- if ((flags & PHP_HTTP_URL_JOIN_QUERY) && __URLSET(new_url, query) && __URLSET(old_url, query)) {
+ if ((flags & PHP_HTTP_URL_JOIN_QUERY) && url_isset(new_url, query) && url_isset(old_url, query)) {
zval qarr, qstr;
INIT_PZVAL(&qstr);
ZVAL_NULL(&qstr);
php_http_querystring_update(&qarr, NULL, &qstr TSRMLS_CC);
- url->query = Z_STRVAL(qstr);
+
+ url(buf)->query = &buf.data[buf.used];
+ url_append(&buf, php_http_buffer_append(&buf, Z_STRVAL(qstr), Z_STRLEN(qstr) + 1));
+
+ zval_dtor(&qstr);
zval_dtor(&qarr);
} else {
- __URLCPY(query);
+ url_copy(query);
}
}
+
if (!(flags & PHP_HTTP_URL_STRIP_FRAGMENT)) {
- __URLCPY(fragment);
+ url_copy(fragment);
}
/* done with copy & combine & strip */
if (flags & PHP_HTTP_URL_FROM_ENV) {
/* free old_url we tainted above */
- php_url_free(tmp_url);
- }
-
- /* set some sane defaults */
-
- if (!url->scheme) {
- url->scheme = estrndup("http", lenof("http"));
+ php_http_url_free(&tmp_url);
}
- if (!url->host) {
- url->host = estrndup("localhost", lenof("localhost"));
- }
-
- if (!url->path) {
- url->path = estrndup("/", 1);
- } else if (url->path[0] != '/') {
- size_t plen = strlen(url->path);
- char *path = emalloc(plen + 1 + 1);
-
- path[0] = '/';
- memcpy(&path[1], url->path, plen + 1);
- STR_SET(url->path, path);
- }
/* replace directory references if path is not a single slash */
if ((flags & PHP_HTTP_URL_SANITIZE_PATH)
- && url->path[0] && (url->path[0] != '/' || url->path[1])) {
- char *ptr, *end = url->path + strlen(url->path) + 1;
+ && url(buf)->path[0] && url(buf)->path[1]) {
+ char *ptr, *end = url(buf)->path + strlen(url(buf)->path) + 1;
- for (ptr = strchr(url->path, '/'); ptr; ptr = strchr(ptr, '/')) {
+ for (ptr = strchr(url(buf)->path, '/'); ptr; ptr = strchr(ptr, '/')) {
switch (ptr[1]) {
case '/':
memmove(&ptr[1], &ptr[2], end - &ptr[2]);
case '.':
if (ptr[3] == '/') {
char *pos = &ptr[4];
- while (ptr != url->path) {
+ while (ptr != url(buf)->path) {
if (*--ptr == '/') {
break;
}
}
}
/* unset default ports */
- if (url->port) {
- if ( ((url->port == 80) && !strcmp(url->scheme, "http"))
- || ((url->port ==443) && !strcmp(url->scheme, "https"))
+ if (url(buf)->port) {
+ if ( ((url(buf)->port == 80) && url(buf)->scheme && !strcmp(url(buf)->scheme, "http"))
+ || ((url(buf)->port ==443) && url(buf)->scheme && !strcmp(url(buf)->scheme, "https"))
) {
- url->port = 0;
+ url(buf)->port = 0;
}
}
+ return url(buf);
+}
+
+char *php_http_url_to_string(const php_http_url_t *url, char **url_str, size_t *url_len, zend_bool persistent)
+{
+ php_http_buffer_t buf;
+
+ php_http_buffer_init_ex(&buf, PHP_HTTP_BUFFER_DEFAULT_SIZE, persistent ?
+ PHP_HTTP_BUFFER_INIT_PERSISTENT : 0);
+
+ if (url->scheme && *url->scheme) {
+ php_http_buffer_appendl(&buf, url->scheme);
+ php_http_buffer_appends(&buf, "://");
+ } else if ((url->user && *url->user) || (url->host && *url->host)) {
+ php_http_buffer_appends(&buf, "//");
+ }
+
+ if (url->user && *url->user) {
+ php_http_buffer_appendl(&buf, url->user);
+ if (url->pass && *url->pass) {
+ php_http_buffer_appends(&buf, ":");
+ php_http_buffer_appendl(&buf, url->pass);
+ }
+ php_http_buffer_appends(&buf, "@");
+ }
+
+ if (url->host && *url->host) {
+ php_http_buffer_appendl(&buf, url->host);
+ if (url->port) {
+ php_http_buffer_appendf(&buf, ":%hu", url->port);
+ }
+ }
+
+ if (url->path && *url->path) {
+ if (*url->path != '/') {
+ php_http_buffer_appends(&buf, "/");
+ }
+ php_http_buffer_appendl(&buf, url->path);
+ } else if (buf.used) {
+ php_http_buffer_appends(&buf, "/");
+ }
+
+ if (url->query && *url->query) {
+ php_http_buffer_appends(&buf, "?");
+ php_http_buffer_appendl(&buf, url->query);
+ }
+
+ if (url->fragment && *url->fragment) {
+ php_http_buffer_appends(&buf, "#");
+ php_http_buffer_appendl(&buf, url->fragment);
+ }
+
+ php_http_buffer_shrink(&buf);
+ php_http_buffer_fix(&buf);
+
+ if (url_len) {
+ *url_len = buf.used;
+ }
+
if (url_str) {
- php_http_url_to_string(url, url_str, url_len TSRMLS_CC);
+ *url_str = buf.data;
}
-
- if (url_ptr) {
- *url_ptr = url;
+
+ return buf.data;
+}
+
+char *php_http_url_authority_to_string(const php_http_url_t *url, char **url_str, size_t *url_len)
+{
+ php_http_buffer_t buf;
+
+ php_http_buffer_init(&buf);
+
+ if (url->user && *url->user) {
+ php_http_buffer_appendl(&buf, url->user);
+ if (url->pass && *url->pass) {
+ php_http_buffer_appends(&buf, ":");
+ php_http_buffer_appendl(&buf, url->pass);
+ }
+ php_http_buffer_appends(&buf, "@");
+ }
+
+ if (url->host && *url->host) {
+ php_http_buffer_appendl(&buf, url->host);
+ if (url->port) {
+ php_http_buffer_appendf(&buf, ":%hu", url->port);
+ }
+ }
+
+ php_http_buffer_shrink(&buf);
+ php_http_buffer_fix(&buf);
+
+ if (url_len) {
+ *url_len = buf.used;
+ }
+
+ if (url_str) {
+ *url_str = buf.data;
+ }
+
+ return buf.data;
+}
+
+php_http_url_t *php_http_url_from_zval(zval *value, unsigned flags TSRMLS_DC)
+{
+ zval *zcpy;
+ php_http_url_t *purl;
+
+ switch (Z_TYPE_P(value)) {
+ case IS_ARRAY:
+ case IS_OBJECT:
+ purl = php_http_url_from_struct(HASH_OF(value));
+ break;
+
+ default:
+ zcpy = php_http_ztyp(IS_STRING, value);
+ purl = php_http_url_parse(Z_STRVAL_P(zcpy), Z_STRLEN_P(zcpy), flags TSRMLS_CC);
+ zval_ptr_dtor(&zcpy);
+ }
+
+ return purl;
+}
+
+php_http_url_t *php_http_url_from_struct(HashTable *ht)
+{
+ zval **e;
+ php_http_buffer_t buf;
+
+ php_http_buffer_init_ex(&buf, MAX(PHP_HTTP_BUFFER_DEFAULT_SIZE, sizeof(php_http_url_t)<<2), PHP_HTTP_BUFFER_INIT_PREALLOC);
+ php_http_buffer_account(&buf, sizeof(php_http_url_t));
+ memset(buf.data, 0, buf.used);
+
+ if (SUCCESS == zend_hash_find(ht, "scheme", sizeof("scheme"), (void *) &e)) {
+ zval *cpy = php_http_ztyp(IS_STRING, *e);
+ url(buf)->scheme = &buf.data[buf.used];
+ url_append(&buf, php_http_buffer_append(&buf, Z_STRVAL_P(cpy), Z_STRLEN_P(cpy) + 1));
+ zval_ptr_dtor(&cpy);
+ }
+ if (SUCCESS == zend_hash_find(ht, "user", sizeof("user"), (void *) &e)) {
+ zval *cpy = php_http_ztyp(IS_STRING, *e);
+ url(buf)->user = &buf.data[buf.used];
+ url_append(&buf, php_http_buffer_append(&buf, Z_STRVAL_P(cpy), Z_STRLEN_P(cpy) + 1));
+ zval_ptr_dtor(&cpy);
+ }
+ if (SUCCESS == zend_hash_find(ht, "pass", sizeof("pass"), (void *) &e)) {
+ zval *cpy = php_http_ztyp(IS_STRING, *e);
+ url(buf)->pass = &buf.data[buf.used];
+ url_append(&buf, php_http_buffer_append(&buf, Z_STRVAL_P(cpy), Z_STRLEN_P(cpy) + 1));
+ zval_ptr_dtor(&cpy);
+ }
+ if (SUCCESS == zend_hash_find(ht, "host", sizeof("host"), (void *) &e)) {
+ zval *cpy = php_http_ztyp(IS_STRING, *e);
+ url(buf)->host = &buf.data[buf.used];
+ url_append(&buf, php_http_buffer_append(&buf, Z_STRVAL_P(cpy), Z_STRLEN_P(cpy) + 1));
+ zval_ptr_dtor(&cpy);
+ }
+ if (SUCCESS == zend_hash_find(ht, "port", sizeof("port"), (void *) &e)) {
+ zval *cpy = php_http_ztyp(IS_LONG, *e);
+ url(buf)->port = (unsigned short) Z_LVAL_P(cpy);
+ zval_ptr_dtor(&cpy);
+ }
+ if (SUCCESS == zend_hash_find(ht, "path", sizeof("path"), (void *) &e)) {
+ zval *cpy = php_http_ztyp(IS_STRING, *e);
+ url(buf)->path = &buf.data[buf.used];
+ url_append(&buf, php_http_buffer_append(&buf, Z_STRVAL_P(cpy), Z_STRLEN_P(cpy) + 1));
+ zval_ptr_dtor(&cpy);
+ }
+ if (SUCCESS == zend_hash_find(ht, "query", sizeof("query"), (void *) &e)) {
+ zval *cpy = php_http_ztyp(IS_STRING, *e);
+ url(buf)->query = &buf.data[buf.used];
+ url_append(&buf, php_http_buffer_append(&buf, Z_STRVAL_P(cpy), Z_STRLEN_P(cpy) + 1));
+ zval_ptr_dtor(&cpy);
+ }
+ if (SUCCESS == zend_hash_find(ht, "fragment", sizeof("fragment"), (void *) &e)) {
+ zval *cpy = php_http_ztyp(IS_STRING, *e);
+ url(buf)->fragment = &buf.data[buf.used];
+ url_append(&buf, php_http_buffer_append(&buf, Z_STRVAL_P(cpy), Z_STRLEN_P(cpy) + 1));
+ zval_ptr_dtor(&cpy);
+ }
+
+ return url(buf);
+}
+
+HashTable *php_http_url_to_struct(const php_http_url_t *url, zval *strct TSRMLS_DC)
+{
+ zval arr;
+
+ if (strct) {
+ switch (Z_TYPE_P(strct)) {
+ default:
+ zval_dtor(strct);
+ array_init(strct);
+ /* no break */
+ case IS_ARRAY:
+ case IS_OBJECT:
+ INIT_PZVAL_ARRAY((&arr), HASH_OF(strct));
+ break;
+ }
} else {
- php_url_free(url);
+ INIT_PZVAL(&arr);
+ array_init(&arr);
+ }
+
+ if (url) {
+ if (url->scheme) {
+ add_assoc_string(&arr, "scheme", url->scheme, 1);
+ }
+ if (url->user) {
+ add_assoc_string(&arr, "user", url->user, 1);
+ }
+ if (url->pass) {
+ add_assoc_string(&arr, "pass", url->pass, 1);
+ }
+ if (url->host) {
+ add_assoc_string(&arr, "host", url->host, 1);
+ }
+ if (url->port) {
+ add_assoc_long(&arr, "port", (long) url->port);
+ }
+ if (url->path) {
+ add_assoc_string(&arr, "path", url->path, 1);
+ }
+ if (url->query) {
+ add_assoc_string(&arr, "query", url->query, 1);
+ }
+ if (url->fragment) {
+ add_assoc_string(&arr, "fragment", url->fragment, 1);
+ }
}
+
+ return Z_ARRVAL(arr);
}
-STATUS php_http_url_encode_hash(HashTable *hash, const char *pre_encoded_str, size_t pre_encoded_len, char **encoded_str, size_t *encoded_len TSRMLS_DC)
+ZEND_RESULT_CODE php_http_url_encode_hash(HashTable *hash, const char *pre_encoded_str, size_t pre_encoded_len, char **encoded_str, size_t *encoded_len TSRMLS_DC)
{
const char *arg_sep_str;
size_t arg_sep_len;
return SUCCESS;
}
-STATUS php_http_url_encode_hash_ex(HashTable *hash, php_http_buffer_t *qstr, const char *arg_sep_str, size_t arg_sep_len, const char *val_sep_str, size_t val_sep_len, const char *pre_encoded_str, size_t pre_encoded_len TSRMLS_DC)
+ZEND_RESULT_CODE php_http_url_encode_hash_ex(HashTable *hash, php_http_buffer_t *qstr, const char *arg_sep_str, size_t arg_sep_len, const char *val_sep_str, size_t val_sep_len, const char *pre_encoded_str, size_t pre_encoded_len TSRMLS_DC)
{
if (pre_encoded_len && pre_encoded_str) {
php_http_buffer_append(qstr, pre_encoded_str, pre_encoded_len);
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) {
}
}
-static size_t parse_mb_utf8(php_http_url_t *url, const char *ptr, const char *end, zend_bool idn)
+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 (!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;
#if defined(HAVE_MBRTOWC)
- mbstate_t ps = {0};
+ 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);
if (!consumed || consumed == (size_t) -1) {
return 0;
}
- if (!idn && !iswalnum(wchar)) {
- return 0;
- }
+ if (wc) {
+ *wc = wchar;
+ }
return consumed;
}
#endif
static const char parse_xdigits[] = "0123456789ABCDEF";
-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 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) {
- if (!(url->flags & PHP_HTTP_URL_PARSE_PCTENC) || what == PARSE_HOSTINFO || what == PARSE_SCHEME) {
- PHP_HTTP_DUFF(consumed, url->buffer[url->offset++] = *ptr++);
+ 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,
- url->buffer[url->offset++] = '%';
- url->buffer[url->offset++] = parse_xdigits[((unsigned char) ptr[i]) >> 4];
- url->buffer[url->offset++] = parse_xdigits[((unsigned char) ptr[i]) & 0xf];
+ 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;
);
}
- } else if (!silent) {
- TSRMLS_FETCH_FROM_CTX(url->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 consumed;
+ if (!silent) {
+ TSRMLS_FETCH_FROM_CTX(state->ts);
+ if (consumed) {
+ php_error_docref(NULL TSRMLS_CC, 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 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 0;
}
-static STATUS parse_userinfo(php_http_url_t *url, const char *ptr)
+static ZEND_RESULT_CODE 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);
- url->user = &url->buffer[url->offset];
+ state->url.user = &state->buffer[state->offset];
do {
switch (*ptr) {
return FAILURE;
}
password = ptr + 1;
- url->buffer[url->offset++] = 0;
- url->pass = &url->buffer[url->offset];
+ state->buffer[state->offset++] = 0;
+ state->url.pass = &state->buffer[state->offset];
break;
case '%':
(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 '*':
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);
- url->buffer[url->offset++] = 0;
+ state->buffer[state->offset++] = 0;
return SUCCESS;
}
-static STATUS parse_hostinfo(php_http_url_t *url, const char *ptr)
+#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 mb, len;
- const char *end = url->ptr, *tmp = ptr, *port = NULL;
- TSRMLS_FETCH_FROM_CTX(url->ts);
+ 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 TSRMLS_CC, 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 TSRMLS_CC, E_WARNING, "Failed to convert UTF-32 'U+%X' to UTF-16", wc);
+ return FAILURE;
+ }
+ }
+
+ return SUCCESS;
+}
+#endif
+
+#ifndef MAXHOSTNAMELEN
+# define MAXHOSTNAMELEN 256
+#endif
+
+#ifdef PHP_HTTP_HAVE_IDN
+static ZEND_RESULT_CODE parse_idn(struct parse_state *state, size_t prev_len)
+{
+ char *idn = NULL;
+ int rv = -1;
+ TSRMLS_FETCH_FROM_CTX(state->ts);
+
+ 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 TSRMLS_CC, 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;
+ TSRMLS_FETCH_FROM_CTX(state->ts);
+
+ 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 TSRMLS_CC, 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 TSRMLS_CC, 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;
+ TSRMLS_FETCH_FROM_CTX(state->ts);
+
+ 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 TSRMLS_CC, 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 TSRMLS_CC, E_WARNING, "Failed to parse IDN");
+ return FAILURE;
+ }
+#endif
+ } else {
+ php_error_docref(NULL TSRMLS_CC, 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 TSRMLS_CC, 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;
+ TSRMLS_FETCH_FROM_CTX(state->ts);
#ifdef HAVE_INET_PTON
if (*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);
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;
}
(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 '*':
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;
default:
- if (port) {
+ if (ptr == end) {
+ break;
+ } else if (port) {
php_error_docref(NULL TSRMLS_CC, 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(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;
}
+ if (state->flags & PHP_HTTP_URL_PARSE_TOIDN) {
#ifdef PHP_HTTP_HAVE_IDN
- if (url->flags & PHP_HTTP_URL_PARSE_IDN) {
- 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);
- }
-# 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);
- }
-# endif
- if (rv != IDNA_SUCCESS) {
- php_error_docref(NULL TSRMLS_CC, E_WARNING, "Failed to parse IDN; %s", idna_strerror(rv));
- return FAILURE;
- } else {
- size_t idnlen = strlen(idn);
- memcpy(url->host, idn, idnlen + 1);
- free(idn);
- url->offset += idnlen - len;
- }
- }
+ 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(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;
}
- tmp = url->ptr + 1;
+ 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 != 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;
+ --state->ptr;
+ goto EOD;
}
-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;
- url->path = &url->buffer[url->offset];
+ 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->buffer[url->offset++] = 0;
- }
- return url->ptr;
+ goto done;
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 */
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;
+ 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(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;
- url->query = &url->buffer[url->offset];
+ tmp = ++state->ptr;
+ state->url.query = &state->buffer[state->offset];
do {
- switch (*url->ptr) {
+ switch (*state->ptr) {
case '#':
- case '\0':
- url->buffer[url->offset++] = 0;
- return url->ptr;
+ goto done;
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 '[':
+ 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 '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;
+ done:
+ state->buffer[state->offset++] = 0;
+ return state->ptr;
}
-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;
- url->fragment = &url->buffer[url->offset];
+ tmp = ++state->ptr;
+ state->url.fragment = &state->buffer[state->offset];
do {
- switch (*url->ptr) {
- case '\0':
- url->buffer[url->offset++] = 0;
- return url->ptr;
-
+ switch (*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 '/':
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;
+ state->buffer[state->offset++] = 0;
+ return state->ptr;
}
-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 */
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;
+ return state->ptr = 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);
+
+ state->end = str + len;
+ state->ptr = str;
+ state->flags = flags;
+ state->maxlen = maxlen;
+ TSRMLS_SET_CTX(state->ts);
- url->end = str + len;
- url->ptr = str;
- url->flags = flags;
- url->maxlen = maxlen;
- TSRMLS_SET_CTX(url->ts);
+ 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(state)) {
+ efree(state);
+ return NULL;
+ }
- 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_query(state)) {
+ php_error_docref(NULL TSRMLS_CC, E_WARNING, "Failed to parse URL query: '%s'", state->ptr);
+ efree(state);
return NULL;
}
- if (!parse_hier(url)) {
- 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;
}
- 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);
+ return (php_http_url_t *) state;
+}
+
+php_http_url_t *php_http_url_parse_authority(const char *str, size_t len, unsigned flags TSRMLS_DC)
+{
+ 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;
+ TSRMLS_SET_CTX(state->ts);
+
+ if (!(state->ptr = parse_authority(state))) {
+ 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 (state->ptr != state->end) {
+ php_error_docref(NULL TSRMLS_CC, E_WARNING,
+ "Failed to parse URL authority, unexpected character at pos %u in '%s'",
+ (unsigned) (state->ptr - str), str);
+ efree(state);
return NULL;
}
- return url;
+ return (php_http_url_t *) state;
}
ZEND_BEGIN_ARG_INFO_EX(ai_HttpUrl___construct, 0, 0, 0)
zend_replace_error_handling(EH_THROW, php_http_exception_bad_url_class_entry, &zeh TSRMLS_CC);
{
- php_url *res_purl, *new_purl = NULL, *old_purl = NULL;
+ php_http_url_t *res_purl, *new_purl = NULL, *old_purl = NULL;
if (new_url) {
- switch (Z_TYPE_P(new_url)) {
- case IS_OBJECT:
- case IS_ARRAY:
- new_purl = php_http_url_from_struct(NULL, HASH_OF(new_url) TSRMLS_CC);
- break;
- default: {
- zval *cpy = php_http_ztyp(IS_STRING, new_url);
-
- new_purl = php_url_parse(Z_STRVAL_P(cpy));
- zval_ptr_dtor(&cpy);
- break;
- }
- }
+ new_purl = php_http_url_from_zval(new_url, flags TSRMLS_CC);
if (!new_purl) {
zend_restore_error_handling(&zeh TSRMLS_CC);
return;
}
}
if (old_url) {
- switch (Z_TYPE_P(old_url)) {
- case IS_OBJECT:
- case IS_ARRAY:
- old_purl = php_http_url_from_struct(NULL, HASH_OF(old_url) TSRMLS_CC);
- break;
- default: {
- zval *cpy = php_http_ztyp(IS_STRING, old_url);
-
- old_purl = php_url_parse(Z_STRVAL_P(cpy));
- zval_ptr_dtor(&cpy);
- break;
- }
- }
+ old_purl = php_http_url_from_zval(old_url, flags TSRMLS_CC);
if (!old_purl) {
if (new_purl) {
- php_url_free(new_purl);
+ php_http_url_free(&new_purl);
}
zend_restore_error_handling(&zeh TSRMLS_CC);
return;
}
}
- php_http_url(flags, old_purl, new_purl, &res_purl, NULL, NULL TSRMLS_CC);
+ res_purl = php_http_url_mod(old_purl, new_purl, flags TSRMLS_CC);
php_http_url_to_struct(res_purl, getThis() TSRMLS_CC);
- php_url_free(res_purl);
+ php_http_url_free(&res_purl);
if (old_purl) {
- php_url_free(old_purl);
+ php_http_url_free(&old_purl);
}
if (new_purl) {
- php_url_free(new_purl);
+ php_http_url_free(&new_purl);
}
}
zend_restore_error_handling(&zeh TSRMLS_CC);
zend_replace_error_handling(EH_THROW, php_http_exception_bad_url_class_entry, &zeh TSRMLS_CC);
{
- php_url *new_purl = NULL, *old_purl = NULL;
+ php_http_url_t *new_purl = NULL, *old_purl = NULL;
if (new_url) {
- switch (Z_TYPE_P(new_url)) {
- case IS_OBJECT:
- case IS_ARRAY:
- new_purl = php_http_url_from_struct(NULL, HASH_OF(new_url) TSRMLS_CC);
- break;
- default: {
- zval *cpy = php_http_ztyp(IS_STRING, new_url);
-
- new_purl = php_url_parse(Z_STRVAL_P(new_url));
- zval_ptr_dtor(&cpy);
- break;
- }
- }
+ new_purl = php_http_url_from_zval(new_url, flags TSRMLS_CC);
if (!new_purl) {
zend_restore_error_handling(&zeh TSRMLS_CC);
return;
}
}
- if ((old_purl = php_http_url_from_struct(NULL, HASH_OF(getThis()) TSRMLS_CC))) {
- php_url *res_purl;
+ if ((old_purl = php_http_url_from_struct(HASH_OF(getThis())))) {
+ php_http_url_t *res_purl;
ZVAL_OBJVAL(return_value, zend_objects_clone_obj(getThis() TSRMLS_CC), 0);
- php_http_url(flags, old_purl, new_purl, &res_purl, NULL, NULL TSRMLS_CC);
+ res_purl = php_http_url_mod(old_purl, new_purl, flags TSRMLS_CC);
php_http_url_to_struct(res_purl, return_value TSRMLS_CC);
- php_url_free(res_purl);
- php_url_free(old_purl);
+ php_http_url_free(&res_purl);
+ php_http_url_free(&old_purl);
}
if (new_purl) {
- php_url_free(new_purl);
+ php_http_url_free(&new_purl);
}
}
zend_restore_error_handling(&zeh TSRMLS_CC);
PHP_METHOD(HttpUrl, toString)
{
if (SUCCESS == zend_parse_parameters_none()) {
- php_url *purl;
+ php_http_url_t *purl;
- if ((purl = php_http_url_from_struct(NULL, HASH_OF(getThis()) TSRMLS_CC))) {
+ if ((purl = php_http_url_from_struct(HASH_OF(getThis())))) {
char *str;
size_t len;
- php_http_url(0, purl, NULL, NULL, &str, &len TSRMLS_CC);
- php_url_free(purl);
+ php_http_url_to_string(purl, &str, &len, 0);
+ php_http_url_free(&purl);
RETURN_STRINGL(str, len, 0);
}
}
ZEND_END_ARG_INFO();
PHP_METHOD(HttpUrl, toArray)
{
- php_url *purl;
+ php_http_url_t *purl;
if (SUCCESS != zend_parse_parameters_none()) {
return;
}
/* strip any non-URL properties */
- purl = php_http_url_from_struct(NULL, HASH_OF(getThis()) TSRMLS_CC);
+ purl = php_http_url_from_struct(HASH_OF(getThis()));
php_http_url_to_struct(purl, return_value TSRMLS_CC);
- php_url_free(purl);
-}
-
-ZEND_BEGIN_ARG_INFO_EX(ai_HttpUrl_parse, 0, 0, 1)
- ZEND_ARG_INFO(0, url)
- ZEND_ARG_INFO(0, flags)
-ZEND_END_ARG_INFO();
-PHP_METHOD(HttpUrl, parse)
-{
- char *str;
- int len;
- long flags = 0;
- php_http_url_t *url;
- zend_error_handling zeh;
-
- php_http_expect(SUCCESS == zend_parse_parameters(ZEND_NUM_ARGS() TSRMLS_CC, "s|l", &str, &len, &flags), invalid_arg, return);
-
- zend_replace_error_handling(EH_THROW, php_http_exception_bad_url_class_entry, &zeh TSRMLS_CC);
- if ((url = php_http_url_parse(str, len, flags TSRMLS_CC))) {
- object_init_ex(return_value, php_http_url_class_entry);
- if (url->scheme) {
- zend_update_property_string(php_http_url_class_entry, return_value,
- ZEND_STRL("scheme"), url->scheme TSRMLS_CC);
- }
- if (url->user) {
- zend_update_property_string(php_http_url_class_entry, return_value,
- ZEND_STRL("user"), url->user TSRMLS_CC);
- }
- if (url->pass) {
- zend_update_property_string(php_http_url_class_entry, return_value,
- ZEND_STRL("pass"), url->pass TSRMLS_CC);
- }
- if (url->host) {
- zend_update_property_string(php_http_url_class_entry, return_value,
- ZEND_STRL("host"), url->host TSRMLS_CC);
- }
- if (url->port) {
- zend_update_property_long(php_http_url_class_entry, return_value,
- ZEND_STRL("port"), url->port TSRMLS_CC);
- }
- if (url->path) {
- zend_update_property_string(php_http_url_class_entry, return_value,
- ZEND_STRL("path"), url->path TSRMLS_CC);
- }
- if (url->query) {
- zend_update_property_string(php_http_url_class_entry, return_value,
- ZEND_STRL("query"), url->query TSRMLS_CC);
- }
- if (url->fragment) {
- zend_update_property_string(php_http_url_class_entry, return_value,
- ZEND_STRL("fragment"), url->fragment TSRMLS_CC);
- }
- php_http_url_free(&url);
- }
- zend_restore_error_handling(&zeh TSRMLS_CC);
+ php_http_url_free(&purl);
}
static zend_function_entry php_http_url_methods[] = {
PHP_ME(HttpUrl, toString, ai_HttpUrl_toString, ZEND_ACC_PUBLIC)
ZEND_MALIAS(HttpUrl, __toString, toString, ai_HttpUrl_toString, ZEND_ACC_PUBLIC)
PHP_ME(HttpUrl, toArray, ai_HttpUrl_toArray, ZEND_ACC_PUBLIC)
- PHP_ME(HttpUrl, parse, ai_HttpUrl_parse, ZEND_ACC_PUBLIC|ZEND_ACC_STATIC)
EMPTY_FUNCTION_ENTRY
};
zend_declare_class_constant_long(php_http_url_class_entry, ZEND_STRL("PARSE_MBLOC"), PHP_HTTP_URL_PARSE_MBLOC TSRMLS_CC);
#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);
+#if defined(PHP_HTTP_HAVE_IDN) || defined(HAVE_UIDNA_IDNTOASCII)
+ 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_PCTENC"), PHP_HTTP_URL_PARSE_PCTENC TSRMLS_CC);
+ zend_declare_class_constant_long(php_http_url_class_entry, ZEND_STRL("PARSE_TOPCT"), PHP_HTTP_URL_PARSE_TOPCT TSRMLS_CC);
return SUCCESS;
}
projects / m6w6 / ext-http / blobdiff