Index: src/unicode-inl.h |
diff --git a/src/unicode-inl.h b/src/unicode-inl.h |
index f861f9f2d47449945d62a6fbc8044abbcd0b2a2b..1efe782958e534b03dff8b18f0d00f6922450739 100644 |
--- a/src/unicode-inl.h |
+++ b/src/unicode-inl.h |
@@ -107,8 +107,17 @@ unsigned Utf8::EncodeOneByte(char* str, uint8_t c) { |
return 2; |
} |
+// Encode encodes the unicode code point c into the given str buffer. Unless |
+// allow_invalid is set to true, surrogate code points will be replaced with |
+// kReplacementCharacter. The caller is required to combine surrogate pairs |
+// into code points before calling Encode. |
+unsigned Utf8::Encode(char* str, uchar c, bool allow_invalid) { |
dcarney
2014/01/04 15:56:45
the name of the third argument is not the same as
haimuiba
2014/01/06 05:40:18
Thx, will fix.
|
+ if (!allow_invalid && |
+ Utf16::IsLeadSurrogate(c) || |
+ Utf16::IsTrailSurrogate(c)) { |
+ c = kReplacementCharacter; |
+ } |
-unsigned Utf8::Encode(char* str, uchar c, int previous) { |
static const int kMask = ~(1 << 6); |
if (c <= kMaxOneByteChar) { |
str[0] = c; |
@@ -118,13 +127,6 @@ unsigned Utf8::Encode(char* str, uchar c, int previous) { |
str[1] = 0x80 | (c & kMask); |
return 2; |
} else if (c <= kMaxThreeByteChar) { |
dcarney
2014/01/04 15:56:45
can't chop this section out - need backwards compa
haimuiba
2014/01/06 05:40:18
I'd prefer changing the other call sites if needed
|
- if (Utf16::IsTrailSurrogate(c) && |
- Utf16::IsLeadSurrogate(previous)) { |
- const int kUnmatchedSize = kSizeOfUnmatchedSurrogate; |
- return Encode(str - kUnmatchedSize, |
- Utf16::CombineSurrogatePair(previous, c), |
- Utf16::kNoPreviousCharacter) - kUnmatchedSize; |
- } |
str[0] = 0xE0 | (c >> 12); |
str[1] = 0x80 | ((c >> 6) & kMask); |
str[2] = 0x80 | (c & kMask); |