UTF-8 validation now uses UTF-8 decoding code.

This commit is contained in:
Ferdinand Majerech 2014-08-02 01:37:16 +02:00
parent b5da695d6b
commit 7b699c5903
2 changed files with 10 additions and 93 deletions

View file

@ -233,20 +233,12 @@ struct ValidateResult
/// Is the validated string valid? /// Is the validated string valid?
bool valid; bool valid;
/// Number of characters in the string. /// Number of characters in the string.
///
/// If the string is not valid, this is the number of valid characters before
/// hitting the first invalid sequence.
size_t characterCount; size_t characterCount;
/// If the string is not valid, error message with details is here. /// If the string is not valid, error message with details is here.
string msg; string msg;
/// If the string is not valid, the first invalid sequence of bytes is here.
const(uint)[] sequence() @safe pure nothrow const @nogc
{
return sequenceBuffer[0 .. sequenceLength];
}
private:
// Buffer for the invalid sequence of bytes if valid == false.
uint[4] sequenceBuffer;
// Number of used bytes in sequenceBuffer.
size_t sequenceLength;
} }
/// Validate a UTF-8 string, checking if it is well-formed Unicode. /// Validate a UTF-8 string, checking if it is well-formed Unicode.
@ -265,88 +257,12 @@ ValidateResult validateUTF8NoGC(const(char[]) str) @trusted pure nothrow @nogc
continue; continue;
} }
// The following encodings are valid, except for the 5 and 6 byte combinations: auto decoded = decodeUTF8NoGC!(No.validated)(str, index);
// 0xxxxxxx if(decoded.errorMessage !is null)
// 110xxxxx 10xxxxxx
// 1110xxxx 10xxxxxx 10xxxxxx
// 11110xxx 10xxxxxx 10xxxxxx 10xxxxxx
// 111110xx 10xxxxxx 10xxxxxx 10xxxxxx 10xxxxxx
// 1111110x 10xxxxxx 10xxxxxx 10xxxxxx 10xxxxxx 10xxxxxx
// Dchar bitmask for different numbers of UTF-8 code units.
import std.typecons;
enum bitMask = tuple((1 << 7) - 1, (1 << 11) - 1, (1 << 16) - 1, (1 << 21) - 1);
auto pstr = str.ptr + index;
immutable length = str.length - index;
ubyte fst = pstr[0];
static ValidateResult error(const(char[]) str, string msg) @safe pure nothrow @nogc
{ {
ValidateResult result; return ValidateResult(false, characterCount, decoded.errorMessage);
size_t i;
do
{
result.sequenceBuffer[i] = str[i];
} while (++i < str.length && i < 4 && (str[i] & 0xC0) == 0x80);
result.valid = false;
result.msg = msg;
result.sequenceLength = i;
return result;
} }
++characterCount;
static ValidateResult invalidUTF(const(char[]) str) @safe pure nothrow @nogc
{
return error(str, "Invalid UTF-8 sequence");
}
static ValidateResult outOfBounds(const(char[]) str) @safe pure nothrow @nogc
{
return error(str, "Attempted to decode past the end of a string");
}
assert(fst & 0x80);
// starter must have at least 2 first bits set
if((fst & 0b1100_0000) != 0b1100_0000) { return invalidUTF(pstr[0 .. length]); }
ubyte tmp = void;
dchar d = fst; // upper control bits are masked out later
fst <<= 1;
foreach (i; TypeTuple!(1, 2, 3))
{
if(i == length) { return outOfBounds(pstr[0 .. length]); }
tmp = pstr[i];
if ((tmp & 0xC0) != 0x80) { return invalidUTF(pstr[0 .. length]); }
d = (d << 6) | (tmp & 0x3F);
fst <<= 1;
if (!(fst & 0x80)) // no more bytes
{
d &= bitMask[i]; // mask out control bits
// overlong, could have been encoded with i bytes
if ((d & ~bitMask[i - 1]) == 0) { return invalidUTF(pstr[0 .. length]); }
// check for surrogates only needed for 3 bytes
static if(i == 2)
{
if (!isValidDchar(d)) { return invalidUTF(pstr[0 .. length]); }
}
++characterCount;
index += i + 1;
static if(i == 3)
{
if (d > dchar.max) { return invalidUTF(pstr[0 .. length]); }
}
continue outer;
}
}
return invalidUTF(pstr[0 .. length]);
} }
return ValidateResult(true, characterCount); return ValidateResult(true, characterCount);

View file

@ -682,8 +682,9 @@ auto toUTF8(ubyte[] input, const UTFEncoding encoding) @safe pure nothrow
const validateResult = result.utf8.validateUTF8NoGC(); const validateResult = result.utf8.validateUTF8NoGC();
if(!validateResult.valid) if(!validateResult.valid)
{ {
result.errorMessage = "UTF-8 validation error: " ~ validateResult.msg ~ result.errorMessage = "UTF-8 validation error after character #" ~
validateResult.sequence.to!string; validateResult.characterCount.to!string ~ ": " ~
validateResult.msg;
} }
result.characterCount = validateResult.characterCount; result.characterCount = validateResult.characterCount;
break; break;