__ntapi_tt_array_convert_utf8_to_utf16(): initial implementation.

1 files changed, 241 insertions, 9 deletions

diff --git a/src/argv/ntapi_tt_array_utf8.c b/src/argv/ntapi_tt_array_utf8.c
index b1c95b5..a2f6d0f 100644
--- a/src/argv/ntapi_tt_array_utf8.c
+++ b/src/argv/ntapi_tt_array_utf8.c
@@ -10,6 +10,137 @@
 #include <ntapi/ntapi.h>
 #include "ntapi_impl.h"
 
+typedef struct ___two_bytes {
+	unsigned char	low;
+	unsigned char	high;
+} __two_bytes;
+
+
+typedef struct ___three_bytes {
+	unsigned char	low;
+	unsigned char	middle;
+	unsigned char	high;
+} __three_bytes;
+
+static void __utf8_to_utf16_handler_1byte_or_null_termination(wchar16_t * dst, const unsigned char * ch)
+{
+	/***************************/
+	/* from: 0xxxxxxx          */
+	/* to:   00000000 0xxxxxxx */
+	/***************************/
+
+	*dst = *ch;
+}
+
+
+static void __utf8_to_utf16_handler_2bytes(wchar16_t * dst, const unsigned char * ch)
+{
+	/***************************/
+	/* from: 110yyyyy 10xxxxxx */
+	/* to:   00000yyy yyxxxxxx */
+	/***************************/
+
+	__two_bytes * src; /* big endian */
+
+	src = (__two_bytes *)ch;
+
+	/* yyyyy */
+	*dst   = (src->low ^ 0xC0);
+	*dst <<= 6;
+
+	/* xxxxxx */
+	*dst |= (src->high ^ 0x80);
+}
+
+
+static void __utf8_to_utf16_handler_3bytes(wchar16_t * dst, const unsigned char * ch)
+{
+	/************************************/
+	/* from: 1110zzzz 10yyyyyy 10xxxxxx */
+	/* to:   zzzzyyyy yyxxxxxx          */
+	/************************************/
+
+	__three_bytes *	src; /* big endian */
+	wchar16_t	yyyyy;
+
+	src = (__three_bytes *)ch;
+
+	/* zzzz */
+	*dst   = (src->low ^ 0xE0);
+	*dst <<= 12;
+
+	/* yyyyy */
+	yyyyy   = (src->middle ^ 0x80);
+	yyyyy <<= 6;
+	*dst |= yyyyy;
+
+	/* xxxxxx */
+	*dst |= (src->high ^ 0x80);
+}
+
+
+static void __utf8_to_utf16_handler_4bytes(wchar16_t * dst, const unsigned char * ch)
+{
+	/*************************************************/
+	/* from: 11110uuu  10uuzzzz  10yyyyyy  10xxxxxx  */
+	/* to:   110110ww  wwzzzzyy  110111yy  yyxxxxxx  */
+	/*************************************************/
+
+	__two_bytes *		src_low;	/* big endian */
+	__two_bytes *		src_high;	/* big endian */
+	wchar16_t *		dst_lead;
+	wchar16_t *		dst_trail;
+
+	wchar16_t		wwww;
+	wchar16_t		lead;
+	wchar16_t		trail;
+	unsigned char		ulow;
+	unsigned char		uhigh;
+	unsigned char		yy;
+	unsigned char		yyyy;
+	unsigned char		zzzz;
+
+	dst_lead = dst_trail = (wchar16_t *)dst;
+	dst_trail++;
+
+	src_low	= src_high = (__two_bytes *)ch;
+	src_high++;
+
+	/* uuuuu */
+	ulow  = src_low->low  ^ 0xF0;
+	uhigh = src_low->high ^ 0x80;
+
+	ulow  <<= 2;
+	uhigh >>= 4;
+
+	/* wwww */
+	wwww   = (ulow | uhigh) - 1;
+	wwww <<= 6;
+
+	/* 110110ww wwzzzzyy */
+	yy     = src_high->low ^ 0x80;
+	yy   >>= 4;
+
+	zzzz   = src_low->high;
+	zzzz <<= 4;
+	zzzz >>= 2;
+
+	lead   = 0xD800;
+	lead  |= wwww;
+	lead  |= zzzz;
+	lead  |= yy;
+
+	/* 110111yy  yyxxxxxx */
+	yyyy   = src_high->low << 4;
+	trail  = yyyy << 2;
+	trail |= src_high->high ^ 0x80;
+	trail |= 0xDC00;
+
+	/* write */
+	*dst_lead  = lead;
+	*dst_trail = trail;
+}
+
 int32_t __stdcall __ntapi_tt_array_copy_utf8(
 	__out	int *			argc,
 	__in	const char **		argv,
@@ -140,21 +271,122 @@ int32_t __stdcall __ntapi_tt_array_copy_utf8(
 	return NT_STATUS_SUCCESS;
 }
 
+static void (*__utf8_to_utf16_handlers[5])(wchar16_t *, const unsigned char *) = {
+	0,
+	__utf8_to_utf16_handler_1byte_or_null_termination,
+	__utf8_to_utf16_handler_2bytes,
+	__utf8_to_utf16_handler_3bytes,
+	__utf8_to_utf16_handler_4bytes};
+
 int32_t __stdcall __ntapi_tt_array_convert_utf8_to_utf16(
 	__in		char **		arrv,
-	__in		wchar16_t **	arra,
-	__in		void *		base,
-	__in		wchar16_t *	buffer,
+	__out		wchar16_t **	warrv,
+	__out		void *		base,
+	__out		wchar16_t *	buffer,
 	__in		size_t		buffer_len,
 	__out		size_t *	bytes_written)
 {
-	(void)arrv;
-	(void)arra;
-	(void)base;
-	(void)buffer;
-	(void)buffer_len;
+	wchar16_t *	ubound;
+	wchar16_t *	wch;
+	ptrdiff_t	diff;
+	ptrdiff_t	wdiff;
+	char *		ch;
+	const uint8_t *	utf8;
+	uint8_t		byte_count;
+
+	if ((uintptr_t)base % sizeof(wchar16_t))
+		return NT_STATUS_INVALID_PARAMETER_3;
+
+	wch   = buffer;
+	diff  = (ptrdiff_t)base;
+	wdiff = (ptrdiff_t)base / sizeof(wchar16_t);
+
+	ubound  = buffer;
+	ubound += buffer_len / sizeof(wchar16_t);
+	ubound--;
+	ubound--;
+	ubound--;
+
+	for (; arrv && *arrv; arrv++,warrv++) {
+		*warrv = wch - wdiff;
+		ch     = *arrv + diff;
+
+		/* ubound already accounts for null termination, see above */
+		for (; *ch && (wch<ubound); ch++) {
+			utf8       = (const uint8_t *)ch;
+			byte_count = 0;
+
+			/* try one byte */
+			if (utf8[0] <= 0x7F)
+				byte_count = 1;
+
+			/* try two bytes */
+			else if ((++ch)
+					&& (utf8[0] >= 0xC2) && (utf8[0] <= 0xDF)
+					&& (utf8[1] >= 0x80) && (utf8[1] <= 0xBF))
+				byte_count = 2;
+
+			/* try three bytes */
+			else if ((++ch)
+					&& (utf8[0] == 0xE0)
+					&& (utf8[1] >= 0xA0) && (utf8[1] <= 0xBF)
+					&& (utf8[2] >= 0x80) && (utf8[2] <= 0xBF))
+				byte_count = 3;
+
+			else if (
+					(utf8[0] >= 0xE1) && (utf8[0] <= 0xEC)
+					&& (utf8[1] >= 0x80) && (utf8[1] <= 0xBF)
+					&& (utf8[2] >= 0x80) && (utf8[2] <= 0xBF))
+				byte_count = 3;
+
+			else if (
+					(utf8[0] == 0xED)
+					&& (utf8[1] >= 0x80) && (utf8[1] <= 0x9F)
+					&& (utf8[2] >= 0x80) && (utf8[2] <= 0xBF))
+				byte_count = 3;
+
+			else if (
+					(utf8[0] >= 0xEE) && (utf8[0] <= 0xEF)
+					&& (utf8[1] >= 0x80) && (utf8[1] <= 0xBF)
+					&& (utf8[2] >= 0x80) && (utf8[2] <= 0xBF))
+				byte_count = 3;
+
+			/* try four bytes */
+			else if ((++ch)
+					&& (utf8[0] == 0xF0)
+					&& (utf8[1] >= 0x90) && (utf8[1] <= 0xBF)
+					&& (utf8[2] >= 0x80) && (utf8[2] <= 0xBF)
+					&& (utf8[3] >= 0x80) && (utf8[3] <= 0xBF))
+				byte_count = 4;
+
+			else if (
+					(utf8[0] >= 0xF1) && (utf8[0] <= 0xF3)
+					&& (utf8[1] >= 0x80) && (utf8[1] <= 0xBF)
+					&& (utf8[2] >= 0x80) && (utf8[2] <= 0xBF)
+					&& (utf8[3] >= 0x80) && (utf8[3] <= 0xBF))
+				byte_count = 4;
+
+			else if (
+					(utf8[0] == 0xF4)
+					&& (utf8[1] >= 0x80) && (utf8[1] <= 0x8F)
+					&& (utf8[2] >= 0x80) && (utf8[2] <= 0xBF)
+					&& (utf8[3] >= 0x80) && (utf8[3] <= 0xBF))
+				byte_count = 4;
+
+			if (byte_count) {
+				__utf8_to_utf16_handlers[byte_count](wch,utf8);
+				wch = &wch[byte_count >> 3];
+				wch++;
+			} else {
+				return NT_STATUS_ILLEGAL_CHARACTER;
+			}
+		}
+
+		*wch++ = 0;
+	}
 
-	*bytes_written = 0;
+	*warrv = 0;
+	*bytes_written = sizeof(wchar16_t) * (wch - buffer);
 
 	return NT_STATUS_SUCCESS;
 }