t7x/deps/curl/lib/curl_multibyte.h
2024-06-17 19:04:36 -04:00

92 lines
3.3 KiB
C

#ifndef HEADER_CURL_MULTIBYTE_H
#define HEADER_CURL_MULTIBYTE_H
/***************************************************************************
* _ _ ____ _
* Project ___| | | | _ \| |
* / __| | | | |_) | |
* | (__| |_| | _ <| |___
* \___|\___/|_| \_\_____|
*
* Copyright (C) Daniel Stenberg, <daniel@haxx.se>, et al.
*
* This software is licensed as described in the file COPYING, which
* you should have received as part of this distribution. The terms
* are also available at https://curl.se/docs/copyright.html.
*
* You may opt to use, copy, modify, merge, publish, distribute and/or sell
* copies of the Software, and permit persons to whom the Software is
* furnished to do so, under the terms of the COPYING file.
*
* This software is distributed on an "AS IS" basis, WITHOUT WARRANTY OF ANY
* KIND, either express or implied.
*
* SPDX-License-Identifier: curl
*
***************************************************************************/
#include "curl_setup.h"
#if defined(WIN32)
/*
* MultiByte conversions using Windows kernel32 library.
*/
wchar_t *curlx_convert_UTF8_to_wchar(const char *str_utf8);
char *curlx_convert_wchar_to_UTF8(const wchar_t *str_w);
#endif /* WIN32 */
/*
* Macros curlx_convert_UTF8_to_tchar(), curlx_convert_tchar_to_UTF8()
* and curlx_unicodefree() main purpose is to minimize the number of
* preprocessor conditional directives needed by code using these
* to differentiate UNICODE from non-UNICODE builds.
*
* In the case of a non-UNICODE build the tchar strings are char strings that
* are duplicated via strdup and remain in whatever the passed in encoding is,
* which is assumed to be UTF-8 but may be other encoding. Therefore the
* significance of the conversion functions is primarily for UNICODE builds.
*
* Allocated memory should be free'd with curlx_unicodefree().
*
* Note: Because these are curlx functions their memory usage is not tracked
* by the curl memory tracker memdebug. You'll notice that curlx function-like
* macros call free and strdup in parentheses, eg (strdup)(ptr), and that's to
* ensure that the curl memdebug override macros do not replace them.
*/
#if defined(UNICODE) && defined(WIN32)
#define curlx_convert_UTF8_to_tchar(ptr) curlx_convert_UTF8_to_wchar((ptr))
#define curlx_convert_tchar_to_UTF8(ptr) curlx_convert_wchar_to_UTF8((ptr))
typedef union {
unsigned short *tchar_ptr;
const unsigned short *const_tchar_ptr;
unsigned short *tbyte_ptr;
const unsigned short *const_tbyte_ptr;
} xcharp_u;
#else
#define curlx_convert_UTF8_to_tchar(ptr) (strdup)(ptr)
#define curlx_convert_tchar_to_UTF8(ptr) (strdup)(ptr)
typedef union {
char *tchar_ptr;
const char *const_tchar_ptr;
unsigned char *tbyte_ptr;
const unsigned char *const_tbyte_ptr;
} xcharp_u;
#endif /* UNICODE && WIN32 */
#define curlx_unicodefree(ptr) \
do { \
if(ptr) { \
(free)(ptr); \
(ptr) = NULL; \
} \
} while(0)
#endif /* HEADER_CURL_MULTIBYTE_H */