|
@@ -530,33 +530,33 @@ typedef const char *SOCK_OPT_TYPE;
|
|
|
|
|
|
#if defined(_WIN32)
|
|
|
/* The Microsoft Visual Studio SDK defines MAX_PATH, but not PATH_MAX.
|
|
|
-* Linux defines PATH_MAX but not MAX_PATH. MinGW defines both.
|
|
|
-* Linux uses UTF-8 filenames, and (typically) defines PATH_MAX as 4096
|
|
|
-* (see https://eklitzke.org/path-max-is-tricky).
|
|
|
-* Windows cannot handle UTF-8 directly but uses UCS-2 (UTF-16 without checking
|
|
|
-* surrogate pairs in the file API).
|
|
|
-* CivetWeb internally uses UTF-8 on every system, thus it needs to convert
|
|
|
-* all filen ames for Windows.
|
|
|
-* Windows has an internal limit of 259 UCS-2 characters (plus termination).
|
|
|
-*/
|
|
|
+ * Linux defines PATH_MAX but not MAX_PATH. MinGW defines both.
|
|
|
+ * Linux uses UTF-8 filenames, and (typically) defines PATH_MAX as 4096
|
|
|
+ * (see https://eklitzke.org/path-max-is-tricky).
|
|
|
+ * Windows cannot handle UTF-8 directly but uses UCS-2 (UTF-16 without checking
|
|
|
+ * surrogate pairs in the file API).
|
|
|
+ * CivetWeb internally uses UTF-8 on every system, thus it needs to convert
|
|
|
+ * all filen ames for Windows.
|
|
|
+ * Windows has an internal limit of 259 UCS-2 characters (plus termination).
|
|
|
+ */
|
|
|
/* PATH_MAX is used in CivetWeb as char[] buffer size for UTF-8 name
|
|
|
-* representations. Every character in the BMP (Basic Multilingual Plane)
|
|
|
-* encoded as UTF-8 requires 1, 2 or 3 bytes. A BMP character encoded as
|
|
|
-* UTF-16 always requires one 16-bit word. A character outside the BMP
|
|
|
-* requires 4 bytes in UTF-8 and two 16-bit words in UTF-16 (a surrogate
|
|
|
-* pair).
|
|
|
-* Thus, a 3*259+1 bytes buffer for UTF-8 encoded text will allways be
|
|
|
-* sufficent to hold a 295+1 word UTF-16 encoded text.
|
|
|
-* We do not need 4*259+1 bytes for the UTF-8 text buffer, since none BMP
|
|
|
-* characters would require two words in UTF-16 and hit the UCS-2 length
|
|
|
-* limit of the Windows file API.
|
|
|
-*/
|
|
|
+ * representations. Every character in the BMP (Basic Multilingual Plane)
|
|
|
+ * encoded as UTF-8 requires 1, 2 or 3 bytes. A BMP character encoded as
|
|
|
+ * UTF-16 always requires one 16-bit word. A character outside the BMP
|
|
|
+ * requires 4 bytes in UTF-8 and two 16-bit words in UTF-16 (a surrogate
|
|
|
+ * pair).
|
|
|
+ * Thus, a 3*259+1 bytes buffer for UTF-8 encoded text will allways be
|
|
|
+ * sufficent to hold a 295+1 word UTF-16 encoded text.
|
|
|
+ * We do not need 4*259+1 bytes for the UTF-8 text buffer, since none BMP
|
|
|
+ * characters would require two words in UTF-16 and hit the UCS-2 length
|
|
|
+ * limit of the Windows file API.
|
|
|
+ */
|
|
|
/* See also https://github.com/civetweb/civetweb/issues/937 */
|
|
|
/* W_PATH_MAX is a char buffer size for 259 BMP characters in UTF-8 plus
|
|
|
-* null termination, rounded up to the next 4 bytes boundary */
|
|
|
+ * null termination, rounded up to the next 4 bytes boundary */
|
|
|
#define PATH_MAX (3 * 260)
|
|
|
/* W_PATH_MAX is the wchar_t buffer size required for 259 BMP characters
|
|
|
-* plus termination. (Note: wchar_t is 16 bit on Windows) */
|
|
|
+ * plus termination. (Note: wchar_t is 16 bit on Windows) */
|
|
|
#define W_PATH_MAX (260)
|
|
|
#endif
|
|
|
|
|
@@ -21212,12 +21212,13 @@ mg_get_context_info(const struct mg_context *ctx, char *buffer, int buflen)
|
|
|
}
|
|
|
|
|
|
|
|
|
-void mg_disable_connection_keep_alive(struct mg_connection *conn)
|
|
|
+void
|
|
|
+mg_disable_connection_keep_alive(struct mg_connection *conn)
|
|
|
{
|
|
|
- /* https://github.com/civetweb/civetweb/issues/727 */
|
|
|
- if (conn != NULL) {
|
|
|
- conn->must_close = 1;
|
|
|
- }
|
|
|
+ /* https://github.com/civetweb/civetweb/issues/727 */
|
|
|
+ if (conn != NULL) {
|
|
|
+ conn->must_close = 1;
|
|
|
+ }
|
|
|
}
|
|
|
|
|
|
|