diff options
Diffstat (limited to 'src/mongo/util/text.cpp')
-rw-r--r-- | src/mongo/util/text.cpp | 115 |
1 files changed, 115 insertions, 0 deletions
diff --git a/src/mongo/util/text.cpp b/src/mongo/util/text.cpp new file mode 100644 index 00000000000..51a2556afdc --- /dev/null +++ b/src/mongo/util/text.cpp @@ -0,0 +1,115 @@ +// text.cpp + +/* Copyright 2009 10gen Inc. + * + * Licensed under the Apache License, Version 2.0 (the "License"); + * you may not use this file except in compliance with the License. + * You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + +#include "pch.h" +#include "text.h" +#include "unittest.h" + +namespace mongo { + + inline int leadingOnes(unsigned char c) { + if (c < 0x80) return 0; + static const char _leadingOnes[128] = { + 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, // 0x80 - 0x8F + 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, // 0x90 - 0x99 + 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, // 0xA0 - 0xA9 + 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, // 0xB0 - 0xB9 + 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, // 0xC0 - 0xC9 + 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, // 0xD0 - 0xD9 + 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, // 0xE0 - 0xE9 + 4, 4, 4, 4, 4, 4, 4, 4, // 0xF0 - 0xF7 + 5, 5, 5, 5, // 0xF8 - 0xFB + 6, 6, // 0xFC - 0xFD + 7, // 0xFE + 8, // 0xFF + }; + return _leadingOnes[c & 0x7f]; + + } + + bool isValidUTF8(const char *s) { + int left = 0; // how many bytes are left in the current codepoint + while (*s) { + const unsigned char c = (unsigned char) *(s++); + const int ones = leadingOnes(c); + if (left) { + if (ones != 1) return false; // should be a continuation byte + left--; + } + else { + if (ones == 0) continue; // ASCII byte + if (ones == 1) return false; // unexpected continuation byte + if (c > 0xF4) return false; // codepoint too large (< 0x10FFFF) + if (c == 0xC0 || c == 0xC1) return false; // codepoints <= 0x7F shouldn't be 2 bytes + + // still valid + left = ones-1; + } + } + if (left!=0) return false; // string ended mid-codepoint + return true; + } + +#if defined(_WIN32) + + std::string toUtf8String(const std::wstring& wide) { + if (wide.size() > boost::integer_traits<int>::const_max) + throw std::length_error( + "Wide string cannot be more than INT_MAX characters long."); + if (wide.size() == 0) + return ""; + + // Calculate necessary buffer size + int len = ::WideCharToMultiByte( + CP_UTF8, 0, wide.c_str(), static_cast<int>(wide.size()), + NULL, 0, NULL, NULL); + + // Perform actual conversion + if (len > 0) { + std::vector<char> buffer(len); + len = ::WideCharToMultiByte( + CP_UTF8, 0, wide.c_str(), static_cast<int>(wide.size()), + &buffer[0], static_cast<int>(buffer.size()), NULL, NULL); + if (len > 0) { + assert(len == static_cast<int>(buffer.size())); + return std::string(&buffer[0], buffer.size()); + } + } + + throw boost::system::system_error( + ::GetLastError(), boost::system::system_category); + } + +#if defined(_UNICODE) + std::wstring toWideString(const char *s) { + std::basic_ostringstream<TCHAR> buf; + buf << s; + return buf.str(); + } +#endif + +#endif + + struct TextUnitTest : public UnitTest { + void run() { + assert( parseLL("123") == 123 ); + assert( parseLL("-123000000000") == -123000000000LL ); + } + } textUnitTest; + +} + |