Index: src/scanner-character-streams.cc |
diff --git a/src/scanner-character-streams.cc b/src/scanner-character-streams.cc |
index 9ec0ad1008316dfca01264c619d9b5c88655a231..d38b1d440a0f7f234f54d4ad3944c39ea77199af 100644 |
--- a/src/scanner-character-streams.cc |
+++ b/src/scanner-character-streams.cc |
@@ -6,12 +6,43 @@ |
#include "src/scanner-character-streams.h" |
+#include "include/v8.h" |
#include "src/handles.h" |
#include "src/unicode-inl.h" |
namespace v8 { |
namespace internal { |
+namespace { |
+ |
+unsigned CopyCharsHelper( |
+ uint16_t* dest, unsigned length, const char* src, unsigned* src_pos, |
+ unsigned src_length, |
+ ScriptCompiler::ExternalSourceStream::Encoding encoding) { |
+ if (encoding == ScriptCompiler::ExternalSourceStream::UTF8) { |
+ return v8::internal::Utf8ToUtf16CharacterStream::CopyChars( |
+ dest, length, reinterpret_cast<const uint8_t*>(src), src_pos, |
+ src_length); |
+ } |
+ |
+ unsigned to_fill = length; |
+ if (to_fill > src_length - *src_pos) to_fill = src_length - *src_pos; |
+ |
+ if (encoding == ScriptCompiler::ExternalSourceStream::ONE_BYTE) { |
+ v8::internal::CopyChars<uint8_t, uint16_t>( |
+ dest, reinterpret_cast<const uint8_t*>(src + *src_pos), to_fill); |
+ } else { |
+ DCHECK(encoding == ScriptCompiler::ExternalSourceStream::TWO_BYTE); |
+ v8::internal::CopyChars<uint16_t, uint16_t>( |
+ dest, reinterpret_cast<const uint16_t*>(src + *src_pos), to_fill); |
+ } |
+ *src_pos += to_fill; |
+ return to_fill; |
+} |
+ |
+} // namespace |
+ |
+ |
// ---------------------------------------------------------------------------- |
// BufferedUtf16CharacterStreams |
@@ -145,6 +176,35 @@ Utf8ToUtf16CharacterStream::Utf8ToUtf16CharacterStream(const byte* data, |
Utf8ToUtf16CharacterStream::~Utf8ToUtf16CharacterStream() { } |
+unsigned Utf8ToUtf16CharacterStream::CopyChars(uint16_t* dest, unsigned length, |
+ const byte* src, |
+ unsigned* src_pos, |
+ unsigned src_length) { |
+ static const unibrow::uchar kMaxUtf16Character = 0xffff; |
+ unsigned i = 0; |
+ // Because of the UTF-16 lead and trail surrogates, we stop filling the buffer |
+ // one character early (in the normal case), because we need to have at least |
+ // two free spaces in the buffer to be sure that the next character will fit. |
+ while (i < length - 1) { |
+ if (*src_pos == src_length) break; |
+ unibrow::uchar c = src[*src_pos]; |
+ if (c <= unibrow::Utf8::kMaxOneByteChar) { |
+ *src_pos = *src_pos + 1; |
+ } else { |
+ c = unibrow::Utf8::CalculateValue(src + *src_pos, src_length - *src_pos, |
+ src_pos); |
+ } |
+ if (c > kMaxUtf16Character) { |
+ dest[i++] = unibrow::Utf16::LeadSurrogate(c); |
+ dest[i++] = unibrow::Utf16::TrailSurrogate(c); |
+ } else { |
+ dest[i++] = static_cast<uc16>(c); |
+ } |
+ } |
+ return i; |
+} |
+ |
+ |
unsigned Utf8ToUtf16CharacterStream::BufferSeekForward(unsigned delta) { |
unsigned old_pos = pos_; |
unsigned target_pos = pos_ + delta; |
@@ -156,31 +216,14 @@ unsigned Utf8ToUtf16CharacterStream::BufferSeekForward(unsigned delta) { |
unsigned Utf8ToUtf16CharacterStream::FillBuffer(unsigned char_position) { |
- static const unibrow::uchar kMaxUtf16Character = 0xffff; |
SetRawPosition(char_position); |
if (raw_character_position_ != char_position) { |
// char_position was not a valid position in the stream (hit the end |
// while spooling to it). |
return 0u; |
} |
- unsigned i = 0; |
- while (i < kBufferSize - 1) { |
- if (raw_data_pos_ == raw_data_length_) break; |
- unibrow::uchar c = raw_data_[raw_data_pos_]; |
- if (c <= unibrow::Utf8::kMaxOneByteChar) { |
- raw_data_pos_++; |
- } else { |
- c = unibrow::Utf8::CalculateValue(raw_data_ + raw_data_pos_, |
- raw_data_length_ - raw_data_pos_, |
- &raw_data_pos_); |
- } |
- if (c > kMaxUtf16Character) { |
- buffer_[i++] = unibrow::Utf16::LeadSurrogate(c); |
- buffer_[i++] = unibrow::Utf16::TrailSurrogate(c); |
- } else { |
- buffer_[i++] = static_cast<uc16>(c); |
- } |
- } |
+ unsigned i = CopyChars(buffer_, kBufferSize, raw_data_, &raw_data_pos_, |
+ raw_data_length_); |
raw_character_position_ = char_position + i; |
return i; |
} |
@@ -276,6 +319,45 @@ void Utf8ToUtf16CharacterStream::SetRawPosition(unsigned target_position) { |
} |
+unsigned ExternalStreamingStream::FillBuffer(unsigned position) { |
+ // Ignore "position" which is the position in the decoded data. Instead, |
+ // ExternalStreamingStream keeps track of the position in the raw data. |
+ unsigned data_in_buffer = 0; |
+ // Note that the UTF-8 decoder might not be able to fill the buffer |
+ // completely; it will typically leave the last character empty (see |
+ // Utf8ToUtf16CharacterStream::CopyChars). |
+ while (data_in_buffer < kBufferSize - 1) { |
+ if (current_data_ == NULL) { |
+ // GetSomeData will wait until the embedder has enough data. |
+ current_data_length_ = source_stream_->GetMoreData(¤t_data_); |
+ current_data_offset_ = 0; |
+ // Did the data stream end? |
+ if (current_data_length_ == 0) { |
+ return data_in_buffer; |
+ } |
+ } |
+ // Fill the buffer from current_data_. FIXME: this doesn't work if the data |
+ // chunk ends in the middle of an UTF-8 character. |
+ unsigned new_offset = 0; |
+ unsigned new_chars_in_buffer = CopyCharsHelper( |
+ buffer_ + data_in_buffer, kBufferSize - data_in_buffer, |
+ current_data_ + current_data_offset_, &new_offset, |
+ current_data_length_ - current_data_offset_, source_stream_->encoding); |
+ data_in_buffer += new_chars_in_buffer; |
+ current_data_offset_ += new_offset; |
+ DCHECK(data_in_buffer <= kBufferSize); |
+ // Did we use all the data? |
+ if (current_data_offset_ == current_data_length_) { |
+ delete[] current_data_; |
+ current_data_ = NULL; |
+ current_data_length_ = 0; |
+ current_data_offset_ = 0; |
+ } |
+ } |
+ return data_in_buffer; |
+} |
+ |
+ |
// ---------------------------------------------------------------------------- |
// ExternalTwoByteStringUtf16CharacterStream |