/*
- * Copyright (C) 2009-2012 the libgit2 contributors
+ * Copyright (C) the libgit2 contributors. All rights reserved.
*
* This file is part of libgit2, distributed under the GNU GPL v2 with
* a Linking Exception. For full terms see the included COPYING file.
*/
-#include "common.h"
#include "utf-conv.h"
-/*
- * Default codepage value
- */
-static int _active_codepage = CP_UTF8;
-
-void gitwin_set_codepage(unsigned int codepage)
+GIT_INLINE(void) git__set_errno(void)
{
- _active_codepage = codepage;
+ if (GetLastError() == ERROR_INSUFFICIENT_BUFFER)
+ errno = ENAMETOOLONG;
+ else
+ errno = EINVAL;
}
-unsigned int gitwin_get_codepage(void)
+/**
+ * Converts a UTF-8 string to wide characters.
+ *
+ * @param dest The buffer to receive the wide string.
+ * @param dest_size The size of the buffer, in characters.
+ * @param src The UTF-8 string to convert.
+ * @return The length of the wide string, in characters (not counting the NULL terminator), or < 0 for failure
+ */
+int git__utf8_to_16(wchar_t *dest, size_t dest_size, const char *src)
{
- return _active_codepage;
+ int len;
+
+ /* Length of -1 indicates NULL termination of the input string. Subtract 1 from the result to
+ * turn 0 into -1 (an error code) and to not count the NULL terminator as part of the string's
+ * length. MultiByteToWideChar never returns int's minvalue, so underflow is not possible */
+ if ((len = MultiByteToWideChar(CP_UTF8, MB_ERR_INVALID_CHARS, src, -1, dest, (int)dest_size) - 1) < 0)
+ git__set_errno();
+
+ return len;
}
-void gitwin_set_utf8(void)
+/**
+ * Converts a wide string to UTF-8.
+ *
+ * @param dest The buffer to receive the UTF-8 string.
+ * @param dest_size The size of the buffer, in bytes.
+ * @param src The wide string to convert.
+ * @return The length of the UTF-8 string, in bytes (not counting the NULL terminator), or < 0 for failure
+ */
+int git__utf16_to_8(char *dest, size_t dest_size, const wchar_t *src)
{
- _active_codepage = CP_UTF8;
+ int len;
+
+ /* Length of -1 indicates NULL termination of the input string. Subtract 1 from the result to
+ * turn 0 into -1 (an error code) and to not count the NULL terminator as part of the string's
+ * length. WideCharToMultiByte never returns int's minvalue, so underflow is not possible */
+ if ((len = WideCharToMultiByte(CP_UTF8, WC_ERR_INVALID_CHARS, src, -1, dest, (int)dest_size, NULL, NULL) - 1) < 0)
+ git__set_errno();
+
+ return len;
}
-wchar_t* gitwin_to_utf16(const char* str)
+/**
+ * Converts a UTF-8 string to wide characters.
+ * Memory is allocated to hold the converted string.
+ * The caller is responsible for freeing the string with git__free.
+ *
+ * @param dest Receives a pointer to the wide string.
+ * @param src The UTF-8 string to convert.
+ * @return The length of the wide string, in characters (not counting the NULL terminator), or < 0 for failure
+ */
+int git__utf8_to_16_alloc(wchar_t **dest, const char *src)
{
- wchar_t* ret;
- int cb;
+ int utf16_size;
+
+ *dest = NULL;
- if (!str) {
- return NULL;
+ /* Length of -1 indicates NULL termination of the input string */
+ utf16_size = MultiByteToWideChar(CP_UTF8, MB_ERR_INVALID_CHARS, src, -1, NULL, 0);
+
+ if (!utf16_size) {
+ git__set_errno();
+ return -1;
}
- cb = strlen(str) * sizeof(wchar_t);
- if (cb == 0) {
- ret = (wchar_t*)git__malloc(sizeof(wchar_t));
- ret[0] = 0;
- return ret;
+ if (!(*dest = git__mallocarray(utf16_size, sizeof(wchar_t)))) {
+ errno = ENOMEM;
+ return -1;
}
- /* Add space for null terminator */
- cb += sizeof(wchar_t);
+ utf16_size = MultiByteToWideChar(CP_UTF8, MB_ERR_INVALID_CHARS, src, -1, *dest, utf16_size);
- ret = (wchar_t*)git__malloc(cb);
+ if (!utf16_size) {
+ git__set_errno();
- if (MultiByteToWideChar(_active_codepage, 0, str, -1, ret, cb) == 0) {
- git__free(ret);
- ret = NULL;
+ git__free(*dest);
+ *dest = NULL;
}
- return ret;
+ /* Subtract 1 from the result to turn 0 into -1 (an error code) and to not count the NULL
+ * terminator as part of the string's length. MultiByteToWideChar never returns int's minvalue,
+ * so underflow is not possible */
+ return utf16_size - 1;
}
-int gitwin_append_utf16(wchar_t *buffer, const char *str, size_t len)
+/**
+ * Converts a wide string to UTF-8.
+ * Memory is allocated to hold the converted string.
+ * The caller is responsible for freeing the string with git__free.
+ *
+ * @param dest Receives a pointer to the UTF-8 string.
+ * @param src The wide string to convert.
+ * @return The length of the UTF-8 string, in bytes (not counting the NULL terminator), or < 0 for failure
+ */
+int git__utf16_to_8_alloc(char **dest, const wchar_t *src)
{
- return MultiByteToWideChar(_active_codepage, 0, str, -1, buffer, len);
-}
+ int utf8_size;
-char* gitwin_from_utf16(const wchar_t* str)
-{
- char* ret;
- int cb;
+ *dest = NULL;
+
+ /* Length of -1 indicates NULL termination of the input string */
+ utf8_size = WideCharToMultiByte(CP_UTF8, WC_ERR_INVALID_CHARS, src, -1, NULL, 0, NULL, NULL);
- if (!str) {
- return NULL;
+ if (!utf8_size) {
+ git__set_errno();
+ return -1;
}
- cb = wcslen(str) * sizeof(char);
- if (cb == 0) {
- ret = (char*)git__malloc(sizeof(char));
- ret[0] = 0;
- return ret;
+ *dest = git__malloc(utf8_size);
+
+ if (!*dest) {
+ errno = ENOMEM;
+ return -1;
}
- /* Add space for null terminator */
- cb += sizeof(char);
+ utf8_size = WideCharToMultiByte(CP_UTF8, WC_ERR_INVALID_CHARS, src, -1, *dest, utf8_size, NULL, NULL);
- ret = (char*)git__malloc(cb);
+ if (!utf8_size) {
+ git__set_errno();
- if (WideCharToMultiByte(_active_codepage, 0, str, -1, ret, cb, NULL, NULL) == 0) {
- git__free(ret);
- ret = NULL;
+ git__free(*dest);
+ *dest = NULL;
}
- return ret;
-
+ /* Subtract 1 from the result to turn 0 into -1 (an error code) and to not count the NULL
+ * terminator as part of the string's length. MultiByteToWideChar never returns int's minvalue,
+ * so underflow is not possible */
+ return utf8_size - 1;
}
projects / libgit2.git / blobdiff