ustring.h

linux下开源浏览器WebKit的源码,市面上的很多商用浏览器都是移植自WebKit

typedef UChar (U_CALLCONV *UNESCAPE_CHAR_AT)(int32_t offset, void *context);
U_CDECL_END

/**
 * Unescape a single sequence. The character at offset-1 is assumed
 * (without checking) to be a backslash.  This method takes a callback
 * pointer to a function that returns the UChar at a given offset.  By
 * varying this callback, ICU functions are able to unescape char*
 * strings, UnicodeString objects, and UFILE pointers.
 *
 * If offset is out of range, or if the escape sequence is ill-formed,
 * (UChar32)0xFFFFFFFF is returned.  See documentation of u_unescape()
 * for a list of recognized sequences.
 *
 * @param charAt callback function that returns a UChar of the source
 * text given an offset and a context pointer.
 * @param offset pointer to the offset that will be passed to charAt.
 * The offset value will be updated upon return to point after the
 * last parsed character of the escape sequence.  On error the offset
 * is unchanged.
 * @param length the number of characters in the source text.  The
 * last character of the source text is considered to be at offset
 * length-1.
 * @param context an opaque pointer passed directly into charAt.
 * @return the character represented by the escape sequence at
 * offset, or (UChar32)0xFFFFFFFF on error.
 * @see u_unescape()
 * @see UnicodeString#unescape()
 * @see UnicodeString#unescapeAt()
 * @stable ICU 2.0
 */
U_STABLE UChar32 U_EXPORT2
u_unescapeAt(UNESCAPE_CHAR_AT charAt,
             int32_t *offset,
             int32_t length,
             void *context);

/**
 * Uppercase the characters in a string.
 * Casing is locale-dependent and context-sensitive.
 * The result may be longer or shorter than the original.
 * The source string and the destination buffer are allowed to overlap.
 *
 * @param dest      A buffer for the result string. The result will be zero-terminated if
 *                  the buffer is large enough.
 * @param destCapacity The size of the buffer (number of UChars). If it is 0, then
 *                  dest may be NULL and the function will only return the length of the result
 *                  without writing any of the result string.
 * @param src       The original string
 * @param srcLength The length of the original string. If -1, then src must be zero-terminated.
 * @param locale    The locale to consider, or "" for the root locale or NULL for the default locale.
 * @param pErrorCode Must be a valid pointer to an error code value,
 *                  which must not indicate a failure before the function call.
 * @return The length of the result string. It may be greater than destCapacity. In that case,
 *         only some of the result was written to the destination buffer.
 * @stable ICU 2.0
 */
U_STABLE int32_t U_EXPORT2
u_strToUpper(UChar *dest, int32_t destCapacity,
             const UChar *src, int32_t srcLength,
             const char *locale,
             UErrorCode *pErrorCode);

/**
 * Lowercase the characters in a string.
 * Casing is locale-dependent and context-sensitive.
 * The result may be longer or shorter than the original.
 * The source string and the destination buffer are allowed to overlap.
 *
 * @param dest      A buffer for the result string. The result will be zero-terminated if
 *                  the buffer is large enough.
 * @param destCapacity The size of the buffer (number of UChars). If it is 0, then
 *                  dest may be NULL and the function will only return the length of the result
 *                  without writing any of the result string.
 * @param src       The original string
 * @param srcLength The length of the original string. If -1, then src must be zero-terminated.
 * @param locale    The locale to consider, or "" for the root locale or NULL for the default locale.
 * @param pErrorCode Must be a valid pointer to an error code value,
 *                  which must not indicate a failure before the function call.
 * @return The length of the result string. It may be greater than destCapacity. In that case,
 *         only some of the result was written to the destination buffer.
 * @stable ICU 2.0
 */
U_STABLE int32_t U_EXPORT2
u_strToLower(UChar *dest, int32_t destCapacity,
             const UChar *src, int32_t srcLength,
             const char *locale,
             UErrorCode *pErrorCode);

#if !UCONFIG_NO_BREAK_ITERATION

/**
 * Titlecase a string.
 * Casing is locale-dependent and context-sensitive.
 * Titlecasing uses a break iterator to find the first characters of words
 * that are to be titlecased. It titlecases those characters and lowercases
 * all others.
 *
 * The titlecase break iterator can be provided to customize for arbitrary
 * styles, using rules and dictionaries beyond the standard iterators.
 * It may be more efficient to always provide an iterator to avoid
 * opening and closing one for each string.
 * The standard titlecase iterator for the root locale implements the
 * algorithm of Unicode TR 21.
 *
 * This function uses only the first() and next() methods of the
 * provided break iterator.
 *
 * The result may be longer or shorter than the original.
 * The source string and the destination buffer are allowed to overlap.
 *
 * @param dest      A buffer for the result string. The result will be zero-terminated if
 *                  the buffer is large enough.
 * @param destCapacity The size of the buffer (number of UChars). If it is 0, then
 *                  dest may be NULL and the function will only return the length of the result
 *                  without writing any of the result string.
 * @param src       The original string
 * @param srcLength The length of the original string. If -1, then src must be zero-terminated.
 * @param titleIter A break iterator to find the first characters of words
 *                  that are to be titlecased.
 *                  If none is provided (NULL), then a standard titlecase
 *                  break iterator is opened.
 * @param locale    The locale to consider, or "" for the root locale or NULL for the default locale.
 * @param pErrorCode Must be a valid pointer to an error code value,
 *                  which must not indicate a failure before the function call.
 * @return The length of the result string. It may be greater than destCapacity. In that case,
 *         only some of the result was written to the destination buffer.
 * @stable ICU 2.1
 */
U_STABLE int32_t U_EXPORT2
u_strToTitle(UChar *dest, int32_t destCapacity,
             const UChar *src, int32_t srcLength,
             UBreakIterator *titleIter,
             const char *locale,
             UErrorCode *pErrorCode);

#endif

/**
 * Case-fold the characters in a string.
 * Case-folding is locale-independent and not context-sensitive,
 * but there is an option for whether to include or exclude mappings for dotted I
 * and dotless i that are marked with 'I' in CaseFolding.txt.
 * The result may be longer or shorter than the original.
 * The source string and the destination buffer are allowed to overlap.
 *
 * @param dest      A buffer for the result string. The result will be zero-terminated if
 *                  the buffer is large enough.
 * @param destCapacity The size of the buffer (number of UChars). If it is 0, then
 *                  dest may be NULL and the function will only return the length of the result
 *                  without writing any of the result string.
 * @param src       The original string
 * @param srcLength The length of the original string. If -1, then src must be zero-terminated.
 * @param options   Either U_FOLD_CASE_DEFAULT or U_FOLD_CASE_EXCLUDE_SPECIAL_I
 * @param pErrorCode Must be a valid pointer to an error code value,
 *                  which must not indicate a failure before the function call.
 * @return The length of the result string. It may be greater than destCapacity. In that case,
 *         only some of the result was written to the destination buffer.
 * @stable ICU 2.0
 */
U_STABLE int32_t U_EXPORT2
u_strFoldCase(UChar *dest, int32_t destCapacity,
              const UChar *src, int32_t srcLength,
              uint32_t options,
              UErrorCode *pErrorCode);

/**
 * Converts a sequence of UChars to wchar_t units.
 *
 * @param dest          A buffer for the result string. The result will be zero-terminated if
 *                      the buffer is large enough.
 * @param destCapacity  The size of the buffer (number of wchar_t's). If it is 0, then
 *                      dest may be NULL and the function will only return the length of the 
 *                      result without writing any of the result string (pre-flighting).
 * @param pDestLength   A pointer to receive the number of units written to the destination. If 
 *                      pDestLength!=NULL then *pDestLength is always set to the 
 *                      number of output units corresponding to the transformation of 
 *                      all the input units, even in case of a buffer overflow.
 * @param src           The original source string
 * @param srcLength     The length of the original string. If -1, then src must be zero-terminated.
 * @param pErrorCode    Must be a valid pointer to an error code value,
 *                      which must not indicate a failure before the function call.
 * @return The pointer to destination buffer.
 * @stable ICU 2.0
 */
U_STABLE wchar_t* U_EXPORT2
u_strToWCS(wchar_t *dest, 
           int32_t destCapacity,
           int32_t *pDestLength,
           const UChar *src, 
           int32_t srcLength,
           UErrorCode *pErrorCode);
/**
 * Converts a sequence of wchar_t units to UChars
 *
 * @param dest          A buffer for the result string. The result will be zero-terminated if
 *                      the buffer is large enough.
 * @param destCapacity  The size of the buffer (number of UChars). If it is 0, then
 *                      dest may be NULL and the function will only return the length of the 
 *                      result without writing any of the result string (pre-flighting).
 * @param pDestLength   A pointer to receive the number of units written to the destination. If 
 *                      pDestLength!=NULL then *pDestLength is always set to the 
 *                      number of output units corresponding to the transformation of 
 *                      all the input units, even in case of a buffer overflow.
 * @param src           The original source string
 * @param srcLength     The length of the original string. If -1, then src must be zero-terminated.
 * @param pErrorCode    Must be a valid pointer to an error code value,
 *                      which must not indicate a failure before the function call.
 * @return The pointer to destination buffer.
 * @stable ICU 2.0
 */
U_STABLE UChar* U_EXPORT2
u_strFromWCS(UChar   *dest,
             int32_t destCapacity, 
             int32_t *pDestLength,
             const wchar_t *src,
             int32_t srcLength,
             UErrorCode *pErrorCode);
/**
 * Converts a sequence of UChars (UTF-16) to UTF-8 bytes
 *
 * @param dest          A buffer for the result string. The result will be zero-terminated if
 *                      the buffer is large enough.
 * @param destCapacity  The size of the buffer (number of chars). If it is 0, then
 *                      dest may be NULL and the function will only return the length of the 
 *                      result without writing any of the result string (pre-flighting).
 * @param pDestLength   A pointer to receive the number of units written to the destination. If 
 *                      pDestLength!=NULL then *pDestLength is always set to the 
 *                      number of output units corresponding to the transformation of 
 *                      all the input units, even in case of a buffer overflow.
 * @param src           The original source string
 * @param srcLength     The length of the original string. If -1, then src must be zero-terminated.
 * @param pErrorCode    Must be a valid pointer to an error code value,
 *                      which must not indicate a failure before the function call.
 * @return The pointer to destination buffer.
 * @stable ICU 2.0
 */
U_STABLE char* U_EXPORT2 
u_strToUTF8(char *dest,           
            int32_t destCapacity,
            int32_t *pDestLength,
            const UChar *src, 
            int32_t srcLength,
            UErrorCode *pErrorCode);

/**
 * Converts a sequence of UTF-8 bytes to UChars (UTF-16).
 *
 * @param dest          A buffer for the result string. The result will be zero-terminated if
 *                      the buffer is large enough.
 * @param destCapacity  The size of the buffer (number of UChars). If it is 0, then
 *                      dest may be NULL and the function will only return the length of the 
 *                      result without writing any of the result string (pre-flighting).
 * @param pDestLength   A pointer to receive the number of units written to the destination. If 
 *                      pDestLength!=NULL then *pDestLength is always set to the 
 *                      number of output units corresponding to the transformation of 
 *                      all the input units, even in case of a buffer overflow.
 * @param src           The original source string
 * @param srcLength     The length of the original string. If -1, then src must be zero-terminated.
 * @param pErrorCode    Must be a valid pointer to an error code value,
 *                      which must not indicate a failure before the function call.
 * @return The pointer to destination buffer.
 * @stable ICU 2.0
 */
U_STABLE UChar* U_EXPORT2
u_strFromUTF8(UChar *dest,             
              int32_t destCapacity,
              int32_t *pDestLength,
              const char *src, 
              int32_t srcLength,
              UErrorCode *pErrorCode);

/**
 * Converts a sequence of UChars (UTF-16) to UTF32 units.
 *
 * @param dest          A buffer for the result string. The result will be zero-terminated if
 *                      the buffer is large enough.
 * @param destCapacity  The size of the buffer (number of UChar32s). If it is 0, then
 *                      dest may be NULL and the function will only return the length of the 
 *                      result without writing any of the result string (pre-flighting).
 * @param pDestLength   A pointer to receive the number of units written to the destination. If 
 *                      pDestLength!=NULL then *pDestLength is always set to the 
 *                      number of output units corresponding to the transformation of 
 *                      all the input units, even in case of a buffer overflow.
 * @param src           The original source string
 * @param srcLength     The length of the original string. If -1, then src must be zero-terminated.
 * @param pErrorCode    Must be a valid pointer to an error code value,
 *                      which must not indicate a failure before the function call.
 * @return The pointer to destination buffer.
 * @stable ICU 2.0
 */
U_STABLE UChar32* U_EXPORT2 
u_strToUTF32(UChar32 *dest, 
             int32_t  destCapacity,
             int32_t  *pDestLength,
             const UChar *src, 
             int32_t  srcLength,
             UErrorCode *pErrorCode);

/**
 * Converts a sequence of UTF32 units to UChars (UTF-16)
 *
 * @param dest          A buffer for the result string. The result will be zero-terminated if
 *                      the buffer is large enough.
 * @param destCapacity  The size of the buffer (number of UChars). If it is 0, then
 *                      dest may be NULL and the function will only return the length of the 
 *                      result without writing any of the result string (pre-flighting).
 * @param pDestLength   A pointer to receive the number of units written to the destination. If 
 *                      pDestLength!=NULL then *pDestLength is always set to the 
 *                      number of output units corresponding to the transformation of 
 *                      all the input units, even in case of a buffer overflow.
 * @param src           The original source string
 * @param srcLength     The length of the original string. If -1, then src must be zero-terminated.
 * @param pErrorCode    Must be a valid pointer to an error code value,
 *                      which must not indicate a failure before the function call.
 * @return The pointer to destination buffer.
 * @stable ICU 2.0
 */
U_STABLE UChar* U_EXPORT2 
u_strFromUTF32(UChar   *dest,
               int32_t destCapacity, 
               int32_t *pDestLength,
               const UChar32 *src,
               int32_t srcLength,
               UErrorCode *pErrorCode);

#endif