public void Utf8ToUtf16_All()
{
// Validate that index conversions on "a*b" map correctly for every Unicode codepoint
byte[] buffer = new byte[10];
String8.Convert("a", ref buffer);
for (int codepoint = 0; codepoint < 0x10FFFF; ++codepoint)
{
// Skip illegal codepoints
if (codepoint >= 0xD800 && codepoint <= 0xDFFF)
{
continue;
}
// Convert the codepoint to UTF16
string value = char.ConvertFromUtf32(codepoint);
// Append it to the String8 after 'a'
String8 value8 = String8.Convert(value, ref buffer, 1);
// Append 'b' after that
String8.Convert("b", ref buffer, value8.Index + value8.Length);
// Map the whole value
String8 whole8 = new String8(buffer, 0, value8.Index + value8.Length + 1);
// 'a' should always map to index 0
Assert.Equal(0, String8.Utf8ToUtf16(0, whole8));
// 'b' should always map to the last .NET char (the length needed for the .NET string + 1 for 'a')
Assert.Equal(1 + value.Length, String8.Utf8ToUtf16(whole8.Length - 1, whole8));
// All indices in between are the middle character (index 1, since it's after 'a')
for (int i = 1; i < whole8.Length - 1; ++i)
{
Assert.Equal(1, String8.Utf8ToUtf16(i, whole8));
}
}
}
public static String8 ConvertExpensively(string value)
{
byte[] buffer = null;
return(String8.Convert(value, ref buffer));
}
public bool Equals(String8 other)
{
return(CompareTo(other) == 0);
}
/// <summary>
/// Translate a UTF-8 index (from an RE2 match) to a UTF-16 match (safe for indexing into a .NET string).
/// Takes an optional previous mapping, so it doesn't have to rescan from the beginning of the file each time.
/// </summary>
/// <remarks>
/// <para>
/// UTF-8:
/// 0xxxxxxx [lt 0x80] - First byte of single byte character
/// 10xxxxxx [lt 0xC0] - Non-first byte of any multi-byte character
/// 110xxxxx [lt 0xE0] - First byte of two byte character
/// 1110xxxx [lt 0xF0] - First byte of three byte character
/// 11110xxx [lt 0xF8] - First byte of four byte character.
/// </para>
/// <para>
/// All Unicode codepoints up to U+FFFF fit in one UTF-16 char.
/// UTF-8 can store 4 + 6 + 6 bits = 16 bits in a three byte encoding,
/// so all UTF-8 three byte and smaller characters will be one UTF-16 character.
/// </para>
/// </remarks>
/// <param name="index">UTF-8 index to translate.</param>
/// <param name="text8">String8 value.</param>
/// <param name="previousUtf8Index">A previous UTF8 index translated, if available.</param>
/// <param name="previousUtf16Index">The UTF-16 equivalent of the previous index, if available.</param>
/// <returns>UTF-16 index corresponding to the UTF-8 index passed in.</returns>
public static int Utf8ToUtf16(int index, String8 text8, int previousUtf8Index = 0, int previousUtf16Index = 0)
{
if (index < 0 || index > text8.Length)
{
throw new ArgumentOutOfRangeException(nameof(index));
}
if (text8.Array == null)
{
throw new ArgumentNullException("text.Array");
}
if (text8.Index < 0 || text8.Length < 0 || text8.Index + text8.Length > text8.Array.Length)
{
throw new ArgumentOutOfRangeException("text");
}
// The first character is always the same
if (index == 0)
{
return(0);
}
int startIndex = 0;
int utf16Index = 0;
// Track from the previous match, if it's available and before the current one
if (previousUtf8Index > 0 && previousUtf8Index <= index)
{
startIndex = previousUtf8Index;
utf16Index = previousUtf16Index;
}
// Count the number of UTF16 characters before text8[index]
int currentLength = 0;
for (int i = startIndex; i <= index - 1; ++i)
{
byte c = text8.Array[i + text8.Index];
if (c >= 0x80 && c < 0xC0)
{
// Continuation Byte; don't count (we figure out the counts from the first byte only)
}
else
{
// Add the previous character length (now that it's done)
utf16Index += currentLength;
// This character will be one UTF-16 char if it's 1-3 bytes, and two if it's 4 bytes (0xF0+)
currentLength = c < 0xF0 ? 1 : 2;
}
}
// If the character being pointed to isn't a continuation character, it's a new index
byte last = index < text8.Length ? text8.Array[index + text8.Index] : (byte)0;
if (!(last >= 0x80 && last < 0xC0))
{
utf16Index += currentLength;
}
return(utf16Index);
}
/// <summary>
/// Initializes a new instance of the <see cref="FlexString"/> class.
/// </summary>
/// <param name="value">String8 value to wrap.</param>
public FlexString(String8 value)
{
_string8 = value;
_isString8Available = true;
}
