From 4d1b9a918f8a189ba1e9887c6a9da04e7392db90 Mon Sep 17 00:00:00 2001 From: Adrian Kummerländer Date: Sat, 5 Oct 2013 12:41:07 +0200 Subject: Initial commit * CodepointIterator is a simple C++ iterator class which iterates through unicode codepoints in a UTF8-encoded string * It is derived from std::iterator and implements the std::bidirectional_iterator_tag * Dereferencing an instance of the class provides the codepoint as char32_t * Tests require Google Test and use UTF8-samples from http://www.columbia.edu/~fdc/utf8/ --- src/codepoint_iterator.cc | 190 ++++++++++++++++++++++++++++++++++++++++++++++ src/codepoint_iterator.h | 49 ++++++++++++ 2 files changed, 239 insertions(+) create mode 100644 src/codepoint_iterator.cc create mode 100644 src/codepoint_iterator.h (limited to 'src') diff --git a/src/codepoint_iterator.cc b/src/codepoint_iterator.cc new file mode 100644 index 0000000..21a8c36 --- /dev/null +++ b/src/codepoint_iterator.cc @@ -0,0 +1,190 @@ +#include "codepoint_iterator.h" + +#include + +namespace { + +enum class CodeUnitType : uint8_t { + CONTINUATION = 128, // 10000000 + LEADING = 64, // 01000000 + THREE = 32, // 00100000 + FOUR = 16, // 00010000 +}; + +enum class CodePoint : uint8_t { + CONTINUATION = 63, // 00111111 + TWO = 31, // 00011111 + THREE = 15, // 00001111 + FOUR = 7, // 00000111 +}; + +inline bool match(const uint8_t& codeUnit, CodeUnitType&& type) { + return codeUnit & static_cast(type); +} + +inline void write(char32_t& codePoint, + const uint8_t& codeUnit, + CodePoint&& mask, + const uint8_t& offset) { + codePoint += (codeUnit & static_cast(mask)) << offset; +} + +} + +namespace UTF8 { + +CodepointIterator::CodepointIterator(std::string::const_iterator iter): + iterator_(iter), + dereferenced_(false), + codepoint_(0) { } + +CodepointIterator::CodepointIterator(const CodepointIterator& src): + iterator_(src.iterator_), + dereferenced_(src.dereferenced_), + codepoint_(src.codepoint_) { } + +CodepointIterator& CodepointIterator::operator=(const CodepointIterator& src) { + this->iterator_ = src.iterator_; + this->dereferenced_ = src.dereferenced_; + this->codepoint_ = src.codepoint_; + + return *this; +} + +bool CodepointIterator::operator==(const CodepointIterator& src) const { + return this->iterator_ == src.iterator_; +} + +bool CodepointIterator::operator!=(const CodepointIterator& src) const { + return this->iterator_ != src.iterator_; +} + +bool CodepointIterator::operator==( + const std::string::const_iterator& src) const { + return this->iterator_ == src; +} + +bool CodepointIterator::operator!=( + const std::string::const_iterator& src) const { + return this->iterator_ != src; +} + +char32_t CodepointIterator::operator*() { + if ( !this->dereferenced_ ) { + uint8_t currByte = *(this->iterator_); + this->dereferenced_ = true; + this->codepoint_ = 0; + + if ( match(currByte, CodeUnitType::CONTINUATION) ) { + if ( match(currByte, CodeUnitType::THREE) ) { + if ( match(currByte, CodeUnitType::FOUR) ) { + write(this->codepoint_, + currByte, + CodePoint::FOUR, + 18); + write(this->codepoint_, + *(this->iterator_ + 1), + CodePoint::CONTINUATION, + 12); + write(this->codepoint_, + *(this->iterator_ + 2), + CodePoint::CONTINUATION, + 6); + write(this->codepoint_, + *(this->iterator_ + 3), + CodePoint::CONTINUATION, + 0); + } else { + write(this->codepoint_, + currByte, + CodePoint::THREE, + 12); + write(this->codepoint_, + *(this->iterator_ + 1), + CodePoint::CONTINUATION, + 6); + write(this->codepoint_, + *(this->iterator_ + 2), + CodePoint::CONTINUATION, + 0); + } + } else { + write(this->codepoint_, + currByte, + CodePoint::TWO, + 6); + write(this->codepoint_, + *(this->iterator_ + 1), + CodePoint::CONTINUATION, + 0); + } + } else { + this->codepoint_ = currByte; + } + } + + return this->codepoint_; +} + +CodepointIterator& CodepointIterator::operator++() { + this->dereferenced_ = false; + uint8_t currByte = *(this->iterator_); + std::string::difference_type offset = 1; + + if ( match(currByte, CodeUnitType::CONTINUATION) ) { + if ( match(currByte, CodeUnitType::THREE) ) { + if ( match(currByte, CodeUnitType::FOUR) ) { + offset = 4; + } else { + offset = 3; + } + } else { + offset = 2; + } + } + + this->iterator_ += offset; + + return *this; +} + +CodepointIterator& CodepointIterator::operator--() { + this->dereferenced_ = false; + --this->iterator_; + + if ( match(*(this->iterator_), CodeUnitType::CONTINUATION) ) { + --this->iterator_; + + if ( !match(*(this->iterator_), CodeUnitType::LEADING) ) { + --this->iterator_; + + if ( !match(*(this->iterator_), CodeUnitType::LEADING) ) { + --this->iterator_; + + if ( !match(*(this->iterator_), CodeUnitType::LEADING) ) { + throw codepoint_invalid(); + } + } + } + } + + return *this; +} + +CodepointIterator CodepointIterator::operator++(int) { + CodepointIterator oldIter(*this); + + ++(*this); + + return oldIter; +} + +CodepointIterator CodepointIterator::operator--(int) { + CodepointIterator oldIter(*this); + + --(*this); + + return oldIter; +} + +} diff --git a/src/codepoint_iterator.h b/src/codepoint_iterator.h new file mode 100644 index 0000000..938f53d --- /dev/null +++ b/src/codepoint_iterator.h @@ -0,0 +1,49 @@ +#ifndef CODEPOINT_ITERATOR_H_ +#define CODEPOINT_ITERATOR_H_ + +#include +#include +#include + +namespace UTF8 { + +class CodepointIterator : public std::iterator { + public: + CodepointIterator(std::string::const_iterator); + CodepointIterator(const CodepointIterator&); + + CodepointIterator& operator=(const CodepointIterator&); + + bool operator==(const CodepointIterator&) const; + bool operator==(const std::string::const_iterator&) const; + + bool operator!=(const CodepointIterator&) const; + bool operator!=(const std::string::const_iterator&) const; + + char32_t operator*(); + + CodepointIterator& operator++(); + CodepointIterator& operator--(); + + CodepointIterator operator++(int); + CodepointIterator operator--(int); + + private: + std::string::const_iterator iterator_; + bool dereferenced_; + char32_t codepoint_; +}; + +class codepoint_invalid: public std::exception { + virtual const char* what() const throw() { + return "codepoint_invalid"; + } +}; + +} + +#endif // CODEPOINT_ITERATOR_H_ -- cgit v1.2.3