[2/2] ntdll: Implement RtlUTF8ToUnicodeN.

[Thomas Faber]

From 50e5976eac7132f3db223d4c055f431e77f21fcf Mon Sep 17 00:00:00 2001
From: Thomas Faber <thomas.faber at reactos.org>
Date: Thu, 21 May 2015 19:29:07 -0400
Subject: ntdll: Implement RtlUTF8ToUnicodeN.

---
 dlls/ntdll/ntdll.spec |    1 +
 dlls/ntdll/rtlstr.c   |  137 +++++++++++++++++++++++++++++++++++++++++++++++++
 2 files changed, 138 insertions(+), 0 deletions(-)

diff --git a/dlls/ntdll/ntdll.spec b/dlls/ntdll/ntdll.spec
index 316f875..9959733 100644
--- a/dlls/ntdll/ntdll.spec
+++ b/dlls/ntdll/ntdll.spec
@@ -935,6 +935,7 @@
 @ stdcall RtlUpperString(ptr ptr)
 @ stub RtlUsageHeap
 @ cdecl -i386 -norelay RtlUshortByteSwap() NTDLL_RtlUshortByteSwap
+@ stdcall RtlUTF8ToUnicodeN(ptr long ptr ptr long)
 @ stdcall RtlValidAcl(ptr)
 # @ stub RtlValidRelativeSecurityDescriptor
 @ stdcall RtlValidSecurityDescriptor(ptr)
diff --git a/dlls/ntdll/rtlstr.c b/dlls/ntdll/rtlstr.c
index 5286b81..4b65c2d 100644
--- a/dlls/ntdll/rtlstr.c
+++ b/dlls/ntdll/rtlstr.c
@@ -2274,3 +2274,140 @@ NTSTATUS WINAPI RtlUnicodeToUTF8N(CHAR *utf8_dest, ULONG utf8_bytes_max,
     *utf8_bytes_written = written;
     return status;
 }
+
+/******************************************************************************
+ * RtlUTF8ToUnicodeN [NTDLL.@]
+ */
+NTSTATUS WINAPI RtlUTF8ToUnicodeN(WCHAR *uni_dest, ULONG uni_bytes_max,
+                                  ULONG *uni_bytes_written,
+                                  const CHAR *utf8_src, ULONG utf8_bytes)
+{
+    NTSTATUS status;
+    ULONG i, j;
+    ULONG written;
+    ULONG ch;
+    ULONG utf8_trail_bytes;
+    WCHAR utf16_ch[3];
+    ULONG utf16_ch_len;
+
+    if (!utf8_src)
+        return STATUS_INVALID_PARAMETER_4;
+    if (!uni_bytes_written)
+        return STATUS_INVALID_PARAMETER;
+
+    written = 0;
+    status = STATUS_SUCCESS;
+
+    for (i = 0; i < utf8_bytes; i++)
+    {
+        /* read UTF-8 lead byte */
+        ch = (BYTE)utf8_src[i];
+        utf8_trail_bytes = 0;
+        if (ch >= 0xf5)
+        {
+            ch = 0xfffd;
+            status = STATUS_SOME_NOT_MAPPED;
+        }
+        else if (ch >= 0xf0)
+        {
+            ch &= 0x07;
+            utf8_trail_bytes = 3;
+        }
+        else if (ch >= 0xe0)
+        {
+            ch &= 0x0f;
+            utf8_trail_bytes = 2;
+        }
+        else if (ch >= 0xc2)
+        {
+            ch &= 0x1f;
+            utf8_trail_bytes = 1;
+        }
+        else if (ch >= 0x80)
+        {
+            /* overlong or trail byte */
+            ch = 0xfffd;
+            status = STATUS_SOME_NOT_MAPPED;
+        }
+
+        /* read UTF-8 trail bytes */
+        if (i + utf8_trail_bytes < utf8_bytes)
+        {
+            for (j = 0; j < utf8_trail_bytes; j++)
+            {
+                if ((utf8_src[i + 1] & 0xc0) == 0x80)
+                {
+                    ch <<= 6;
+                    ch |= utf8_src[i + 1] & 0x3f;
+                    i++;
+                }
+                else
+                {
+                    ch = 0xfffd;
+                    utf8_trail_bytes = 0;
+                    status = STATUS_SOME_NOT_MAPPED;
+                    break;
+                }
+            }
+        }
+        else
+        {
+            ch = 0xfffd;
+            utf8_trail_bytes = 0;
+            status = STATUS_SOME_NOT_MAPPED;
+            i = utf8_bytes;
+        }
+
+        /* encode ch as UTF-16 */
+        if ((ch > 0x10ffff) ||
+            (ch >= 0xd800 && ch <= 0xdfff) ||
+            (utf8_trail_bytes == 2 && ch < 0x00800) ||
+            (utf8_trail_bytes == 3 && ch < 0x10000))
+        {
+            /* invalid codepoint or overlong encoding */
+            utf16_ch[0] = 0xfffd;
+            utf16_ch[1] = 0xfffd;
+            utf16_ch[2] = 0xfffd;
+            utf16_ch_len = utf8_trail_bytes;
+            status = STATUS_SOME_NOT_MAPPED;
+        }
+        else if (ch >= 0x10000)
+        {
+            /* surrogate pair */
+            ch -= 0x010000;
+            utf16_ch[0] = 0xd800 + (ch >> 10 & 0x3ff);
+            utf16_ch[1] = 0xdc00 + (ch >>  0 & 0x3ff);
+            utf16_ch_len = 2;
+        }
+        else
+        {
+            /* single unit */
+            utf16_ch[0] = ch;
+            utf16_ch_len = 1;
+        }
+
+        if (!uni_dest)
+        {
+            written += utf16_ch_len;
+            continue;
+        }
+
+        for (j = 0; j < utf16_ch_len; j++)
+        {
+            if (uni_bytes_max >= sizeof(WCHAR))
+            {
+                *uni_dest++ = utf16_ch[j];
+                uni_bytes_max -= sizeof(WCHAR);
+                written++;
+            }
+            else
+            {
+                uni_bytes_max = 0;
+                status = STATUS_BUFFER_TOO_SMALL;
+            }
+        }
+    }
+
+    *uni_bytes_written = written * sizeof(WCHAR);
+    return status;
+}
-- 
1.7.1