diff options
Diffstat (limited to 'src')
| -rw-r--r-- | src/pugixml.cpp | 368 |
1 files changed, 147 insertions, 221 deletions
diff --git a/src/pugixml.cpp b/src/pugixml.cpp index 99ba3f3..d40840e 100644 --- a/src/pugixml.cpp +++ b/src/pugixml.cpp @@ -1591,28 +1591,11 @@ PUGI__NS_BEGIN } }; - template <size_t size> struct wchar_selector; - - template <> struct wchar_selector<2> - { - typedef uint16_t type; - typedef utf16_counter counter; - typedef utf16_writer writer; - }; - - template <> struct wchar_selector<4> + struct utf8_decoder { - typedef uint32_t type; - typedef utf32_counter counter; - typedef utf32_writer writer; - }; + typedef uint8_t type; - typedef wchar_selector<sizeof(wchar_t)>::counter wchar_counter; - typedef wchar_selector<sizeof(wchar_t)>::writer wchar_writer; - - template <typename Traits, typename opt_swap = opt_false> struct utf_decoder - { - static inline typename Traits::value_type decode_utf8_block(const uint8_t* data, size_t size, typename Traits::value_type result) + template <typename Traits> static inline typename Traits::value_type process(const uint8_t* data, size_t size, typename Traits::value_type result, Traits) { const uint8_t utf8_byte_mask = 0x3f; @@ -1673,29 +1656,34 @@ PUGI__NS_BEGIN return result; } + }; - static inline typename Traits::value_type decode_utf16_block(const uint16_t* data, size_t size, typename Traits::value_type result) - { - const uint16_t* end = data + size; + template <typename opt_swap> struct utf16_decoder + { + typedef uint16_t type; - while (data < end) + template <typename Traits> static inline typename Traits::value_type process(const uint16_t* data, size_t size, typename Traits::value_type result, Traits) + { + while (size) { - unsigned int lead = opt_swap::value ? endian_swap(*data) : *data; + uint16_t lead = opt_swap::value ? endian_swap(*data) : *data; // U+0000..U+D7FF if (lead < 0xD800) { result = Traits::low(result, lead); data += 1; + size -= 1; } // U+E000..U+FFFF else if (static_cast<unsigned int>(lead - 0xE000) < 0x2000) { result = Traits::low(result, lead); data += 1; + size -= 1; } // surrogate pair lead - else if (static_cast<unsigned int>(lead - 0xD800) < 0x400 && data + 1 < end) + else if (static_cast<unsigned int>(lead - 0xD800) < 0x400 && size >= 2) { uint16_t next = opt_swap::value ? endian_swap(data[1]) : data[1]; @@ -1703,26 +1691,32 @@ PUGI__NS_BEGIN { result = Traits::high(result, 0x10000 + ((lead & 0x3ff) << 10) + (next & 0x3ff)); data += 2; + size -= 2; } else { data += 1; + size -= 1; } } else { data += 1; + size -= 1; } } return result; } + }; - static inline typename Traits::value_type decode_utf32_block(const uint32_t* data, size_t size, typename Traits::value_type result) - { - const uint32_t* end = data + size; + template <typename opt_swap> struct utf32_decoder + { + typedef uint32_t type; - while (data < end) + template <typename Traits> static inline typename Traits::value_type process(const uint32_t* data, size_t size, typename Traits::value_type result, Traits) + { + while (size) { uint32_t lead = opt_swap::value ? endian_swap(*data) : *data; @@ -1731,41 +1725,67 @@ PUGI__NS_BEGIN { result = Traits::low(result, lead); data += 1; + size -= 1; } // U+10000..U+10FFFF else { result = Traits::high(result, lead); data += 1; + size -= 1; } } return result; } + }; + + struct latin1_decoder + { + typedef uint8_t type; - static inline typename Traits::value_type decode_latin1_block(const uint8_t* data, size_t size, typename Traits::value_type result) + template <typename Traits> static inline typename Traits::value_type process(const uint8_t* data, size_t size, typename Traits::value_type result, Traits) { - for (size_t i = 0; i < size; ++i) + while (size) { - result = Traits::low(result, data[i]); + result = Traits::low(result, *data); + data += 1; + size -= 1; } return result; } + }; - static inline typename Traits::value_type decode_wchar_block_impl(const uint16_t* data, size_t size, typename Traits::value_type result) - { - return decode_utf16_block(data, size, result); - } + template <size_t size> struct wchar_selector; - static inline typename Traits::value_type decode_wchar_block_impl(const uint32_t* data, size_t size, typename Traits::value_type result) - { - return decode_utf32_block(data, size, result); - } + template <> struct wchar_selector<2> + { + typedef uint16_t type; + typedef utf16_counter counter; + typedef utf16_writer writer; + typedef utf16_decoder<opt_false> decoder; + }; + + template <> struct wchar_selector<4> + { + typedef uint32_t type; + typedef utf32_counter counter; + typedef utf32_writer writer; + typedef utf32_decoder<opt_false> decoder; + }; + + typedef wchar_selector<sizeof(wchar_t)>::counter wchar_counter; + typedef wchar_selector<sizeof(wchar_t)>::writer wchar_writer; + + struct wchar_decoder + { + typedef wchar_t type; - static inline typename Traits::value_type decode_wchar_block(const wchar_t* data, size_t size, typename Traits::value_type result) + template <typename Traits> static inline typename Traits::value_type process(const wchar_t* data, size_t size, typename Traits::value_type result, Traits traits) { - return decode_wchar_block_impl(reinterpret_cast<const wchar_selector<sizeof(wchar_t)>::type*>(data), size, result); + typedef wchar_selector<sizeof(wchar_t)>::decoder decoder; + return decoder::process(reinterpret_cast<const typename decoder::type*>(data), size, result, traits); } }; @@ -1986,38 +2006,13 @@ PUGI__NS_BEGIN return true; } - PUGI__FN bool convert_buffer_utf8(char_t*& out_buffer, size_t& out_length, const void* contents, size_t size) + template <typename D> PUGI__FN bool convert_buffer_generic(char_t*& out_buffer, size_t& out_length, const void* contents, size_t size, D) { - const uint8_t* data = static_cast<const uint8_t*>(contents); - size_t data_length = size; + const typename D::type* data = static_cast<const typename D::type*>(contents); + size_t data_length = size / sizeof(typename D::type); // first pass: get length in wchar_t units - size_t length = utf_decoder<wchar_counter>::decode_utf8_block(data, data_length, 0); - - // allocate buffer of suitable length - char_t* buffer = static_cast<char_t*>(xml_memory::allocate((length + 1) * sizeof(char_t))); - if (!buffer) return false; - - // second pass: convert utf8 input to wchar_t - wchar_writer::value_type obegin = reinterpret_cast<wchar_writer::value_type>(buffer); - wchar_writer::value_type oend = utf_decoder<wchar_writer>::decode_utf8_block(data, data_length, obegin); - - assert(oend == obegin + length); - *oend = 0; - - out_buffer = buffer; - out_length = length + 1; - - return true; - } - - template <typename opt_swap> PUGI__FN bool convert_buffer_utf16(char_t*& out_buffer, size_t& out_length, const void* contents, size_t size, opt_swap) - { - const uint16_t* data = static_cast<const uint16_t*>(contents); - size_t data_length = size / sizeof(uint16_t); - - // first pass: get length in wchar_t units - size_t length = utf_decoder<wchar_counter, opt_swap>::decode_utf16_block(data, data_length, 0); + size_t length = D::process(data, data_length, 0, wchar_counter()); // allocate buffer of suitable length char_t* buffer = static_cast<char_t*>(xml_memory::allocate((length + 1) * sizeof(char_t))); @@ -2025,57 +2020,7 @@ PUGI__NS_BEGIN // second pass: convert utf16 input to wchar_t wchar_writer::value_type obegin = reinterpret_cast<wchar_writer::value_type>(buffer); - wchar_writer::value_type oend = utf_decoder<wchar_writer, opt_swap>::decode_utf16_block(data, data_length, obegin); - - assert(oend == obegin + length); - *oend = 0; - - out_buffer = buffer; - out_length = length + 1; - - return true; - } - - template <typename opt_swap> PUGI__FN bool convert_buffer_utf32(char_t*& out_buffer, size_t& out_length, const void* contents, size_t size, opt_swap) - { - const uint32_t* data = static_cast<const uint32_t*>(contents); - size_t data_length = size / sizeof(uint32_t); - - // first pass: get length in wchar_t units - size_t length = utf_decoder<wchar_counter, opt_swap>::decode_utf32_block(data, data_length, 0); - - // allocate buffer of suitable length - char_t* buffer = static_cast<char_t*>(xml_memory::allocate((length + 1) * sizeof(char_t))); - if (!buffer) return false; - - // second pass: convert utf32 input to wchar_t - wchar_writer::value_type obegin = reinterpret_cast<wchar_writer::value_type>(buffer); - wchar_writer::value_type oend = utf_decoder<wchar_writer, opt_swap>::decode_utf32_block(data, data_length, obegin); - - assert(oend == obegin + length); - *oend = 0; - - out_buffer = buffer; - out_length = length + 1; - - return true; - } - - PUGI__FN bool convert_buffer_latin1(char_t*& out_buffer, size_t& out_length, const void* contents, size_t size) - { - const uint8_t* data = static_cast<const uint8_t*>(contents); - size_t data_length = size; - - // get length in wchar_t units - size_t length = data_length; - - // allocate buffer of suitable length - char_t* buffer = static_cast<char_t*>(xml_memory::allocate((length + 1) * sizeof(char_t))); - if (!buffer) return false; - - // convert latin1 input to wchar_t - wchar_writer::value_type obegin = reinterpret_cast<wchar_writer::value_type>(buffer); - wchar_writer::value_type oend = utf_decoder<wchar_writer>::decode_latin1_block(data, data_length, obegin); + wchar_writer::value_type oend = D::process(data, data_length, obegin, wchar_writer()); assert(oend == obegin + length); *oend = 0; @@ -2092,13 +2037,16 @@ PUGI__NS_BEGIN xml_encoding wchar_encoding = get_wchar_encoding(); // fast path: no conversion required - if (encoding == wchar_encoding) return get_mutable_buffer(out_buffer, out_length, contents, size, is_mutable); + if (encoding == wchar_encoding) + return get_mutable_buffer(out_buffer, out_length, contents, size, is_mutable); // only endian-swapping is required - if (need_endian_swap_utf(encoding, wchar_encoding)) return convert_buffer_endian_swap(out_buffer, out_length, contents, size, is_mutable); + if (need_endian_swap_utf(encoding, wchar_encoding)) + return convert_buffer_endian_swap(out_buffer, out_length, contents, size, is_mutable); // source encoding is utf8 - if (encoding == encoding_utf8) return convert_buffer_utf8(out_buffer, out_length, contents, size); + if (encoding == encoding_utf8) + return convert_buffer_generic(out_buffer, out_length, contents, size, utf8_decoder()); // source encoding is utf16 if (encoding == encoding_utf16_be || encoding == encoding_utf16_le) @@ -2106,8 +2054,8 @@ PUGI__NS_BEGIN xml_encoding native_encoding = is_little_endian() ? encoding_utf16_le : encoding_utf16_be; return (native_encoding == encoding) ? - convert_buffer_utf16(out_buffer, out_length, contents, size, opt_false()) : - convert_buffer_utf16(out_buffer, out_length, contents, size, opt_true()); + convert_buffer_generic(out_buffer, out_length, contents, size, utf16_decoder<opt_false>()) : + convert_buffer_generic(out_buffer, out_length, contents, size, utf16_decoder<opt_true>()); } // source encoding is utf32 @@ -2116,24 +2064,25 @@ PUGI__NS_BEGIN xml_encoding native_encoding = is_little_endian() ? encoding_utf32_le : encoding_utf32_be; return (native_encoding == encoding) ? - convert_buffer_utf32(out_buffer, out_length, contents, size, opt_false()) : - convert_buffer_utf32(out_buffer, out_length, contents, size, opt_true()); + convert_buffer_generic(out_buffer, out_length, contents, size, utf32_decoder<opt_false>()) : + convert_buffer_generic(out_buffer, out_length, contents, size, utf32_decoder<opt_true>()); } // source encoding is latin1 - if (encoding == encoding_latin1) return convert_buffer_latin1(out_buffer, out_length, contents, size); + if (encoding == encoding_latin1) + return convert_buffer_generic(out_buffer, out_length, contents, size, latin1_decoder()); assert(!"Invalid encoding"); return false; } #else - template <typename opt_swap> PUGI__FN bool convert_buffer_utf16(char_t*& out_buffer, size_t& out_length, const void* contents, size_t size, opt_swap) + template <typename D> PUGI__FN bool convert_buffer_generic(char_t*& out_buffer, size_t& out_length, const void* contents, size_t size, D) { - const uint16_t* data = static_cast<const uint16_t*>(contents); - size_t data_length = size / sizeof(uint16_t); + const typename D::type* data = static_cast<const typename D::type*>(contents); + size_t data_length = size / sizeof(typename D::type); // first pass: get length in utf8 units - size_t length = utf_decoder<utf8_counter, opt_swap>::decode_utf16_block(data, data_length, 0); + size_t length = D::process(data, data_length, 0, utf8_counter()); // allocate buffer of suitable length char_t* buffer = static_cast<char_t*>(xml_memory::allocate((length + 1) * sizeof(char_t))); @@ -2141,32 +2090,7 @@ PUGI__NS_BEGIN // second pass: convert utf16 input to utf8 uint8_t* obegin = reinterpret_cast<uint8_t*>(buffer); - uint8_t* oend = utf_decoder<utf8_writer, opt_swap>::decode_utf16_block(data, data_length, obegin); - - assert(oend == obegin + length); - *oend = 0; - - out_buffer = buffer; - out_length = length + 1; - - return true; - } - - template <typename opt_swap> PUGI__FN bool convert_buffer_utf32(char_t*& out_buffer, size_t& out_length, const void* contents, size_t size, opt_swap) - { - const uint32_t* data = static_cast<const uint32_t*>(contents); - size_t data_length = size / sizeof(uint32_t); - - // first pass: get length in utf8 units - size_t length = utf_decoder<utf8_counter, opt_swap>::decode_utf32_block(data, data_length, 0); - - // allocate buffer of suitable length - char_t* buffer = static_cast<char_t*>(xml_memory::allocate((length + 1) * sizeof(char_t))); - if (!buffer) return false; - - // second pass: convert utf32 input to utf8 - uint8_t* obegin = reinterpret_cast<uint8_t*>(buffer); - uint8_t* oend = utf_decoder<utf8_writer, opt_swap>::decode_utf32_block(data, data_length, obegin); + uint8_t* oend = D::process(data, data_length, obegin, utf8_writer()); assert(oend == obegin + length); *oend = 0; @@ -2202,7 +2126,7 @@ PUGI__NS_BEGIN if (postfix_length == 0) return get_mutable_buffer(out_buffer, out_length, contents, size, is_mutable); // first pass: get length in utf8 units - size_t length = prefix_length + utf_decoder<utf8_counter>::decode_latin1_block(postfix, postfix_length, 0); + size_t length = prefix_length + latin1_decoder::process(postfix, postfix_length, 0, utf8_counter()); // allocate buffer of suitable length char_t* buffer = static_cast<char_t*>(xml_memory::allocate((length + 1) * sizeof(char_t))); @@ -2212,7 +2136,7 @@ PUGI__NS_BEGIN memcpy(buffer, data, prefix_length); uint8_t* obegin = reinterpret_cast<uint8_t*>(buffer); - uint8_t* oend = utf_decoder<utf8_writer>::decode_latin1_block(postfix, postfix_length, obegin + prefix_length); + uint8_t* oend = latin1_decoder::process(postfix, postfix_length, obegin + prefix_length, utf8_writer()); assert(oend == obegin + length); *oend = 0; @@ -2226,7 +2150,8 @@ PUGI__NS_BEGIN PUGI__FN bool convert_buffer(char_t*& out_buffer, size_t& out_length, xml_encoding encoding, const void* contents, size_t size, bool is_mutable) { // fast path: no conversion required - if (encoding == encoding_utf8) return get_mutable_buffer(out_buffer, out_length, contents, size, is_mutable); + if (encoding == encoding_utf8) + return get_mutable_buffer(out_buffer, out_length, contents, size, is_mutable); // source encoding is utf16 if (encoding == encoding_utf16_be || encoding == encoding_utf16_le) @@ -2234,8 +2159,8 @@ PUGI__NS_BEGIN xml_encoding native_encoding = is_little_endian() ? encoding_utf16_le : encoding_utf16_be; return (native_encoding == encoding) ? - convert_buffer_utf16(out_buffer, out_length, contents, size, opt_false()) : - convert_buffer_utf16(out_buffer, out_length, contents, size, opt_true()); + convert_buffer_generic(out_buffer, out_length, contents, size, utf16_decoder<opt_false>()) : + convert_buffer_generic(out_buffer, out_length, contents, size, utf16_decoder<opt_true>()); } // source encoding is utf32 @@ -2244,12 +2169,13 @@ PUGI__NS_BEGIN xml_encoding native_encoding = is_little_endian() ? encoding_utf32_le : encoding_utf32_be; return (native_encoding == encoding) ? - convert_buffer_utf32(out_buffer, out_length, contents, size, opt_false()) : - convert_buffer_utf32(out_buffer, out_length, contents, size, opt_true()); + convert_buffer_generic(out_buffer, out_length, contents, size, utf32_decoder<opt_false>()) : + convert_buffer_generic(out_buffer, out_length, contents, size, utf32_decoder<opt_true>()); } // source encoding is latin1 - if (encoding == encoding_latin1) return convert_buffer_latin1(out_buffer, out_length, contents, size, is_mutable); + if (encoding == encoding_latin1) + return convert_buffer_latin1(out_buffer, out_length, contents, size, is_mutable); assert(!"Invalid encoding"); return false; @@ -2259,14 +2185,14 @@ PUGI__NS_BEGIN PUGI__FN size_t as_utf8_begin(const wchar_t* str, size_t length) { // get length in utf8 characters - return utf_decoder<utf8_counter>::decode_wchar_block(str, length, 0); + return wchar_decoder::process(str, length, 0, utf8_counter()); } PUGI__FN void as_utf8_end(char* buffer, size_t size, const wchar_t* str, size_t length) { // convert to utf8 uint8_t* begin = reinterpret_cast<uint8_t*>(buffer); - uint8_t* end = utf_decoder<utf8_writer>::decode_wchar_block(str, length, begin); + uint8_t* end = wchar_decoder::process(str, length, begin, utf8_writer()); assert(begin + size == end); (void)!end; @@ -2294,7 +2220,7 @@ PUGI__NS_BEGIN const uint8_t* data = reinterpret_cast<const uint8_t*>(str); // first pass: get length in wchar_t units - size_t length = utf_decoder<wchar_counter>::decode_utf8_block(data, size, 0); + size_t length = utf8_decoder::process(data, size, 0, wchar_counter()); // allocate resulting string std::basic_string<wchar_t> result; @@ -2304,7 +2230,7 @@ PUGI__NS_BEGIN if (length > 0) { wchar_writer::value_type begin = reinterpret_cast<wchar_writer::value_type>(&result[0]); - wchar_writer::value_type end = utf_decoder<wchar_writer>::decode_utf8_block(data, size, begin); + wchar_writer::value_type end = utf8_decoder::process(data, size, begin, wchar_writer()); assert(begin + length == end); (void)!end; @@ -3576,7 +3502,7 @@ PUGI__NS_BEGIN if (encoding == encoding_utf8) { uint8_t* dest = r_u8; - uint8_t* end = utf_decoder<utf8_writer>::decode_wchar_block(data, length, dest); + uint8_t* end = wchar_decoder::process(data, length, dest, utf8_writer()); return static_cast<size_t>(end - dest); } @@ -3587,7 +3513,7 @@ PUGI__NS_BEGIN uint16_t* dest = r_u16; // convert to native utf16 - uint16_t* end = utf_decoder<utf16_writer>::decode_wchar_block(data, length, dest); + uint16_t* end = wchar_decoder::process(data, length, dest, utf16_writer()); // swap if necessary xml_encoding native_encoding = is_little_endian() ? encoding_utf16_le : encoding_utf16_be; @@ -3603,7 +3529,7 @@ PUGI__NS_BEGIN uint32_t* dest = r_u32; // convert to native utf32 - uint32_t* end = utf_decoder<utf32_writer>::decode_wchar_block(data, length, dest); + uint32_t* end = wchar_decoder::process(data, length, dest, utf32_writer()); // swap if necessary xml_encoding native_encoding = is_little_endian() ? encoding_utf32_le : encoding_utf32_be; @@ -3617,7 +3543,7 @@ PUGI__NS_BEGIN if (encoding == encoding_latin1) { uint8_t* dest = r_u8; - uint8_t* end = utf_decoder<latin1_writer>::decode_wchar_block(data, length, dest); + uint8_t* end = wchar_decoder::process(data, length, dest, latin1_writer()); return static_cast<size_t>(end - dest); } @@ -3649,7 +3575,7 @@ PUGI__NS_BEGIN uint16_t* dest = r_u16; // convert to native utf16 - uint16_t* end = utf_decoder<utf16_writer>::decode_utf8_block(reinterpret_cast<const uint8_t*>(data), length, dest); + uint16_t* end = utf8_decoder::process(reinterpret_cast<const uint8_t*>(data), length, dest, utf16_writer()); // swap if necessary xml_encoding native_encoding = is_little_endian() ? encoding_utf16_le : encoding_utf16_be; @@ -3664,7 +3590,7 @@ PUGI__NS_BEGIN uint32_t* dest = r_u32; // convert to native utf32 - uint32_t* end = utf_decoder<utf32_writer>::decode_utf8_block(reinterpret_cast<const uint8_t*>(data), length, dest); + uint32_t* end = utf8_decoder::process(reinterpret_cast<const uint8_t*>(data), length, dest, utf32_writer()); // swap if necessary xml_encoding native_encoding = is_little_endian() ? encoding_utf32_le : encoding_utf32_be; @@ -3677,7 +3603,7 @@ PUGI__NS_BEGIN if (encoding == encoding_latin1) { uint8_t* dest = r_u8; - uint8_t* end = utf_decoder<latin1_writer>::decode_utf8_block(reinterpret_cast<const uint8_t*>(data), length, dest); + uint8_t* end = utf8_decoder::process(reinterpret_cast<const uint8_t*>(data), length, dest, latin1_writer()); return static_cast<size_t>(end - dest); } @@ -4649,6 +4575,38 @@ PUGI__NS_BEGIN } #endif + PUGI__FN xml_parse_result load_buffer_impl(xml_document_struct* doc, xml_node_struct* root, void* contents, size_t size, unsigned int options, xml_encoding encoding, bool is_mutable, bool own, char_t** out_buffer) + { + // check input buffer + if (!contents && size) return make_parse_result(status_io_error); + + // get actual encoding + xml_encoding buffer_encoding = impl::get_buffer_encoding(encoding, contents, size); + + // get private buffer + char_t* buffer = 0; + size_t length = 0; + + if (!impl::convert_buffer(buffer, length, buffer_encoding, contents, size, is_mutable)) return impl::make_parse_result(status_out_of_memory); + + // delete original buffer if we performed a conversion + if (own && buffer != contents && contents) impl::xml_memory::deallocate(contents); + + // grab onto buffer if it's our buffer, user is responsible for deallocating contents himself + if (own || buffer != contents) *out_buffer = buffer; + + // store buffer for offset_debug + doc->buffer = buffer; + + // parse + xml_parse_result res = impl::xml_parser::parse(buffer, length, doc, root, options); + + // remember encoding + res.encoding = buffer_encoding; + + return res; + } + // we need to get length of entire file to load it in memory; the only (relatively) sane way to do it is via seek/tell trick PUGI__FN xml_parse_status get_file_size(FILE* file, size_t& out_result) { @@ -4714,7 +4672,7 @@ PUGI__NS_BEGIN return size; } - PUGI__FN xml_parse_result load_file_impl(xml_document& doc, FILE* file, unsigned int options, xml_encoding encoding) + PUGI__FN xml_parse_result load_file_impl(xml_document_struct* doc, FILE* file, unsigned int options, xml_encoding encoding, char_t** out_buffer) { if (!file) return make_parse_result(status_file_not_found); @@ -4739,8 +4697,8 @@ PUGI__NS_BEGIN } xml_encoding real_encoding = get_buffer_encoding(encoding, contents, size); - - return doc.load_buffer_inplace_own(contents, zero_terminate_buffer(contents, size, real_encoding), options, real_encoding); + + return load_buffer_impl(doc, doc, contents, zero_terminate_buffer(contents, size, real_encoding), options, real_encoding, true, true, out_buffer); } #ifndef PUGIXML_NO_STL @@ -4867,7 +4825,7 @@ PUGI__NS_BEGIN return status_ok; } - template <typename T> PUGI__FN xml_parse_result load_stream_impl(xml_document& doc, std::basic_istream<T>& stream, unsigned int options, xml_encoding encoding) + template <typename T> PUGI__FN xml_parse_result load_stream_impl(xml_document_struct* doc, std::basic_istream<T>& stream, unsigned int options, xml_encoding encoding, char_t** out_buffer) { void* buffer = 0; size_t size = 0; @@ -4889,7 +4847,7 @@ PUGI__NS_BEGIN xml_encoding real_encoding = get_buffer_encoding(encoding, buffer, size); - return doc.load_buffer_inplace_own(buffer, zero_terminate_buffer(buffer, size, real_encoding), options, real_encoding); + return load_buffer_impl(doc, doc, buffer, zero_terminate_buffer(buffer, size, real_encoding), options, real_encoding, true, true, out_buffer); } #endif @@ -4949,38 +4907,6 @@ PUGI__NS_BEGIN return ferror(file) == 0; } - - PUGI__FN xml_parse_result load_buffer_impl(xml_document_struct* doc, xml_node_struct* root, void* contents, size_t size, unsigned int options, xml_encoding encoding, bool is_mutable, bool own, char_t** out_buffer) - { - // check input buffer - if (!contents && size) return make_parse_result(status_io_error); - - // get actual encoding - xml_encoding buffer_encoding = impl::get_buffer_encoding(encoding, contents, size); - - // get private buffer - char_t* buffer = 0; - size_t length = 0; - - if (!impl::convert_buffer(buffer, length, buffer_encoding, contents, size, is_mutable)) return impl::make_parse_result(status_out_of_memory); - - // delete original buffer if we performed a conversion - if (own && buffer != contents && contents) impl::xml_memory::deallocate(contents); - - // grab onto buffer if it's our buffer, user is responsible for deallocating contents himself - if (own || buffer != contents) *out_buffer = buffer; - - // store buffer for offset_debug - doc->buffer = buffer; - - // parse - xml_parse_result res = impl::xml_parser::parse(buffer, length, doc, root, options); - - // remember encoding - res.encoding = buffer_encoding; - - return res; - } PUGI__NS_END namespace pugi @@ -6837,14 +6763,14 @@ namespace pugi { reset(); - return impl::load_stream_impl(*this, stream, options, encoding); + return impl::load_stream_impl(static_cast<impl::xml_document_struct*>(_root), stream, options, encoding, &_buffer); } PUGI__FN xml_parse_result xml_document::load(std::basic_istream<wchar_t, std::char_traits<wchar_t> >& stream, unsigned int options) { reset(); - return impl::load_stream_impl(*this, stream, options, encoding_wchar); + return impl::load_stream_impl(static_cast<impl::xml_document_struct*>(_root), stream, options, encoding_wchar, &_buffer); } #endif @@ -6872,7 +6798,7 @@ namespace pugi using impl::auto_deleter; // MSVC7 workaround auto_deleter<FILE, int(*)(FILE*)> file(fopen(path_, "rb"), fclose); - return impl::load_file_impl(*this, file.data, options, encoding); + return impl::load_file_impl(static_cast<impl::xml_document_struct*>(_root), file.data, options, encoding, &_buffer); } PUGI__FN xml_parse_result xml_document::load_file(const wchar_t* path_, unsigned int options, xml_encoding encoding) @@ -6882,7 +6808,7 @@ namespace pugi using impl::auto_deleter; // MSVC7 workaround auto_deleter<FILE, int(*)(FILE*)> file(impl::open_file_wide(path_, L"rb"), fclose); - return impl::load_file_impl(*this, file.data, options, encoding); + return impl::load_file_impl(static_cast<impl::xml_document_struct*>(_root), file.data, options, encoding, &_buffer); } PUGI__FN xml_parse_result xml_document::load_buffer(const void* contents, size_t size, unsigned int options, xml_encoding encoding) |