uni.h revision 3e6c6998d1dfeded8b9a23f5aa94ad63e9a681d9
a2760a4d593af57a19ec32cdfd1723b68495df8cvboxsync * innotek Portable Runtime - Unicode Code Points.
a2760a4d593af57a19ec32cdfd1723b68495df8cvboxsync * Copyright (C) 2006-2007 innotek GmbH
a2760a4d593af57a19ec32cdfd1723b68495df8cvboxsync * This file is part of VirtualBox Open Source Edition (OSE), as
a2760a4d593af57a19ec32cdfd1723b68495df8cvboxsync * available from http://www.virtualbox.org. This file is free software;
a2760a4d593af57a19ec32cdfd1723b68495df8cvboxsync * you can redistribute it and/or modify it under the terms of the GNU
a2760a4d593af57a19ec32cdfd1723b68495df8cvboxsync * General Public License as published by the Free Software Foundation,
a2760a4d593af57a19ec32cdfd1723b68495df8cvboxsync * in version 2 as it comes in the "COPYING" file of the VirtualBox OSE
a2760a4d593af57a19ec32cdfd1723b68495df8cvboxsync * distribution. VirtualBox OSE is distributed in the hope that it will
a2760a4d593af57a19ec32cdfd1723b68495df8cvboxsync * be useful, but WITHOUT ANY WARRANTY of any kind.
a2760a4d593af57a19ec32cdfd1723b68495df8cvboxsync/** @defgroup grp_rt_uni RTUniCp - Unicode Code Points
a2760a4d593af57a19ec32cdfd1723b68495df8cvboxsync * @ingroup grp_rt
a2760a4d593af57a19ec32cdfd1723b68495df8cvboxsync/** @def RTUNI_USE_WCTYPE
a2760a4d593af57a19ec32cdfd1723b68495df8cvboxsync * Define RTUNI_USE_WCTYPE to not use the IPRT unicode data but the
a2760a4d593af57a19ec32cdfd1723b68495df8cvboxsync * data which the C runtime library provides. */
a2760a4d593af57a19ec32cdfd1723b68495df8cvboxsync/** Max value a RTUNICP type can hold. */
a2760a4d593af57a19ec32cdfd1723b68495df8cvboxsync/** Invalid code point.
a2760a4d593af57a19ec32cdfd1723b68495df8cvboxsync * This is returned when encountered invalid encodings or invalid
a2760a4d593af57a19ec32cdfd1723b68495df8cvboxsync * unicode code points. */
a2760a4d593af57a19ec32cdfd1723b68495df8cvboxsync * A unicode flags range.
a2760a4d593af57a19ec32cdfd1723b68495df8cvboxsync * @internal
a2760a4d593af57a19ec32cdfd1723b68495df8cvboxsync /** The first code point of the range. */
a2760a4d593af57a19ec32cdfd1723b68495df8cvboxsync /** The last + 1 code point of the range. */
a2760a4d593af57a19ec32cdfd1723b68495df8cvboxsync /** Pointer to the array of case folded code points. */
a2760a4d593af57a19ec32cdfd1723b68495df8cvboxsync/** Pointer to a flags range.
a2760a4d593af57a19ec32cdfd1723b68495df8cvboxsync * @internal */
a2760a4d593af57a19ec32cdfd1723b68495df8cvboxsync/** Pointer to a const flags range.
a2760a4d593af57a19ec32cdfd1723b68495df8cvboxsync * @internal */
a2760a4d593af57a19ec32cdfd1723b68495df8cvboxsync * A unicode case folded range.
a2760a4d593af57a19ec32cdfd1723b68495df8cvboxsync * @internal
a2760a4d593af57a19ec32cdfd1723b68495df8cvboxsync /** The first code point of the range. */
a2760a4d593af57a19ec32cdfd1723b68495df8cvboxsync /** The last + 1 code point of the range. */
a2760a4d593af57a19ec32cdfd1723b68495df8cvboxsync /** Pointer to the array of case folded code points. */
a2760a4d593af57a19ec32cdfd1723b68495df8cvboxsync/** Pointer to a case folded range.
a2760a4d593af57a19ec32cdfd1723b68495df8cvboxsync * @internal */
a2760a4d593af57a19ec32cdfd1723b68495df8cvboxsync/** Pointer to a const case folded range.
a2760a4d593af57a19ec32cdfd1723b68495df8cvboxsync * @internal */
a2760a4d593af57a19ec32cdfd1723b68495df8cvboxsync/** @name Unicode Code Point Flags.
a2760a4d593af57a19ec32cdfd1723b68495df8cvboxsync * @internal
a2760a4d593af57a19ec32cdfd1723b68495df8cvboxsync/*#define RTUNI_BSPACE RT_BIT(6) - later */
a2760a4d593af57a19ec32cdfd1723b68495df8cvboxsync * Array of flags ranges.
a2760a4d593af57a19ec32cdfd1723b68495df8cvboxsync * @internal
a2760a4d593af57a19ec32cdfd1723b68495df8cvboxsyncextern RTDATADECL(const RTUNIFLAGSRANGE) g_aRTUniFlagsRanges[];
a2760a4d593af57a19ec32cdfd1723b68495df8cvboxsync * Gets the flags for a unicode code point.
a2760a4d593af57a19ec32cdfd1723b68495df8cvboxsync * @returns The flag mask. (RTUNI_*)
a2760a4d593af57a19ec32cdfd1723b68495df8cvboxsync * @param CodePoint The unicode code point.
a2760a4d593af57a19ec32cdfd1723b68495df8cvboxsync * @internal
a2760a4d593af57a19ec32cdfd1723b68495df8cvboxsync CodePoint = pCur->pafFlags[CodePoint - pCur->BeginCP];
a2760a4d593af57a19ec32cdfd1723b68495df8cvboxsync * Checks if a unicode code point is upper case.
a2760a4d593af57a19ec32cdfd1723b68495df8cvboxsync * @returns true if it is.
a2760a4d593af57a19ec32cdfd1723b68495df8cvboxsync * @returns false if it isn't.
a2760a4d593af57a19ec32cdfd1723b68495df8cvboxsync * @param CodePoint The code point.
a2760a4d593af57a19ec32cdfd1723b68495df8cvboxsync return (rtUniCpFlags(CodePoint) & RTUNI_UPPER) != 0;
a2760a4d593af57a19ec32cdfd1723b68495df8cvboxsync * Checks if a unicode code point is lower case.
a2760a4d593af57a19ec32cdfd1723b68495df8cvboxsync * @returns true if it is.
a2760a4d593af57a19ec32cdfd1723b68495df8cvboxsync * @returns false if it isn't.
a2760a4d593af57a19ec32cdfd1723b68495df8cvboxsync * @param CodePoint The code point.
a2760a4d593af57a19ec32cdfd1723b68495df8cvboxsync return (rtUniCpFlags(CodePoint) & RTUNI_LOWER) != 0;
a2760a4d593af57a19ec32cdfd1723b68495df8cvboxsync * Checks if a unicode code point is alphabetic.
a2760a4d593af57a19ec32cdfd1723b68495df8cvboxsync * @returns true if it is.
a2760a4d593af57a19ec32cdfd1723b68495df8cvboxsync * @returns false if it isn't.
a2760a4d593af57a19ec32cdfd1723b68495df8cvboxsync * @param CodePoint The code point.
a2760a4d593af57a19ec32cdfd1723b68495df8cvboxsyncDECLINLINE(bool) RTUniCpIsAlphabetic(RTUNICP CodePoint)
a2760a4d593af57a19ec32cdfd1723b68495df8cvboxsync return (rtUniCpFlags(CodePoint) & RTUNI_ALPHA) != 0;
a2760a4d593af57a19ec32cdfd1723b68495df8cvboxsync * Checks if a unicode code point is a decimal digit.
a2760a4d593af57a19ec32cdfd1723b68495df8cvboxsync * @returns true if it is.
a2760a4d593af57a19ec32cdfd1723b68495df8cvboxsync * @returns false if it isn't.
a2760a4d593af57a19ec32cdfd1723b68495df8cvboxsync * @param CodePoint The code point.
a2760a4d593af57a19ec32cdfd1723b68495df8cvboxsyncDECLINLINE(bool) RTUniCpIsDecDigit(RTUNICP CodePoint)
a2760a4d593af57a19ec32cdfd1723b68495df8cvboxsync return (rtUniCpFlags(CodePoint) & RTUNI_DDIGIT) != 0;
a2760a4d593af57a19ec32cdfd1723b68495df8cvboxsync * Checks if a unicode code point is a hexadecimal digit.
a2760a4d593af57a19ec32cdfd1723b68495df8cvboxsync * @returns true if it is.
a2760a4d593af57a19ec32cdfd1723b68495df8cvboxsync * @returns false if it isn't.
a2760a4d593af57a19ec32cdfd1723b68495df8cvboxsync * @param CodePoint The code point.
a2760a4d593af57a19ec32cdfd1723b68495df8cvboxsyncDECLINLINE(bool) RTUniCpIsHexDigit(RTUNICP CodePoint)
a2760a4d593af57a19ec32cdfd1723b68495df8cvboxsync return (rtUniCpFlags(CodePoint) & RTUNI_XDIGIT) != 0;
a2760a4d593af57a19ec32cdfd1723b68495df8cvboxsync * Checks if a unicode code point is white space.
a2760a4d593af57a19ec32cdfd1723b68495df8cvboxsync * @returns true if it is.
a2760a4d593af57a19ec32cdfd1723b68495df8cvboxsync * @returns false if it isn't.
a2760a4d593af57a19ec32cdfd1723b68495df8cvboxsync * @param CodePoint The code point.
a2760a4d593af57a19ec32cdfd1723b68495df8cvboxsync return (rtUniCpFlags(CodePoint) & RTUNI_WSPACE) != 0;
a2760a4d593af57a19ec32cdfd1723b68495df8cvboxsync * Array of uppercase ranges.
a2760a4d593af57a19ec32cdfd1723b68495df8cvboxsync * @internal
a2760a4d593af57a19ec32cdfd1723b68495df8cvboxsyncextern RTDATADECL(const RTUNICASERANGE) g_aRTUniUpperRanges[];
a2760a4d593af57a19ec32cdfd1723b68495df8cvboxsync * Array of lowercase ranges.
a2760a4d593af57a19ec32cdfd1723b68495df8cvboxsync * @internal
a2760a4d593af57a19ec32cdfd1723b68495df8cvboxsyncextern RTDATADECL(const RTUNICASERANGE) g_aRTUniLowerRanges[];
a2760a4d593af57a19ec32cdfd1723b68495df8cvboxsync * Folds a unicode code point using the specified range array.
pCur++;
return CodePoint;