Encoding.h revision 303975
176633Stshiozak//===--- Encoding.h - Format C++ code -------------------------------------===// 276633Stshiozak// 376633Stshiozak// The LLVM Compiler Infrastructure 476633Stshiozak// 576633Stshiozak// This file is distributed under the University of Illinois Open Source 676633Stshiozak// License. See LICENSE.TXT for details. 776633Stshiozak// 876633Stshiozak//===----------------------------------------------------------------------===// 976633Stshiozak/// 1076633Stshiozak/// \file 1176633Stshiozak/// \brief Contains functions for text encoding manipulation. Supports UTF-8, 1276633Stshiozak/// 8-bit encodings and escape sequences in C++ string literals. 1376633Stshiozak/// 1476633Stshiozak//===----------------------------------------------------------------------===// 1576633Stshiozak 1676633Stshiozak#ifndef LLVM_CLANG_LIB_FORMAT_ENCODING_H 1776633Stshiozak#define LLVM_CLANG_LIB_FORMAT_ENCODING_H 1876633Stshiozak 1976633Stshiozak#include "clang/Basic/LLVM.h" 2076633Stshiozak#include "llvm/Support/ConvertUTF.h" 2176633Stshiozak#include "llvm/Support/Unicode.h" 2276633Stshiozak 2376633Stshiozaknamespace clang { 2476633Stshiozaknamespace format { 2576633Stshiozaknamespace encoding { 2676633Stshiozak 2779555Sobrienenum Encoding { 2876633Stshiozak Encoding_UTF8, 2976633Stshiozak Encoding_Unknown // We treat all other encodings as 8-bit encodings. 3076633Stshiozak}; 3176633Stshiozak 3276633Stshiozak/// \brief Detects encoding of the Text. If the Text can be decoded using UTF-8, 3376633Stshiozak/// it is considered UTF8, otherwise we treat it as some 8-bit encoding. 3476633Stshiozakinline Encoding detectEncoding(StringRef Text) { 3576633Stshiozak const UTF8 *Ptr = reinterpret_cast<const UTF8 *>(Text.begin()); 3676633Stshiozak const UTF8 *BufEnd = reinterpret_cast<const UTF8 *>(Text.end()); 37101253Stjr if (::isLegalUTF8String(&Ptr, BufEnd)) 38101253Stjr return Encoding_UTF8; 39101313Stjr return Encoding_Unknown; 40101409Stjr} 41101409Stjr 42101313Stjrinline unsigned getCodePointCountUTF8(StringRef Text) { 43101313Stjr unsigned CodePoints = 0; 44101253Stjr for (size_t i = 0, e = Text.size(); i < e; i += getNumBytesForUTF8(Text[i])) { 45101409Stjr ++CodePoints; 46101409Stjr } 47101253Stjr return CodePoints; 48101253Stjr} 49101409Stjr 50101409Stjr/// \brief Gets the number of code points in the Text using the specified 51101409Stjr/// Encoding. 52101409Stjrinline unsigned getCodePointCount(StringRef Text, Encoding Encoding) { 53101409Stjr switch (Encoding) { 5476633Stshiozak case Encoding_UTF8: 5576633Stshiozak return getCodePointCountUTF8(Text); 5676633Stshiozak default: 5776633Stshiozak return Text.size(); 5876633Stshiozak } 5993032Simp} 6093032Simp 6193032Simp/// \brief Returns the number of columns required to display the \p Text on a 6293032Simp/// generic Unicode-capable terminal. Text is assumed to use the specified 63101369Stjr/// \p Encoding. 6493032Simpinline unsigned columnWidth(StringRef Text, Encoding Encoding) { 6593032Simp if (Encoding == Encoding_UTF8) { 6693032Simp int ContentWidth = llvm::sys::unicode::columnWidthUTF8(Text); 6793032Simp // FIXME: Figure out the correct way to handle this in the presence of both 6893032Simp // printable and unprintable multi-byte UTF-8 characters. Falling back to 6993032Simp // returning the number of bytes may cause problems, as columnWidth suddenly 7093032Simp // becomes non-additive. 7193032Simp if (ContentWidth >= 0) 72101369Stjr return ContentWidth; 7393032Simp } 7493032Simp return Text.size(); 75101409Stjr} 76101409Stjr 77101409Stjr/// \brief Returns the number of columns required to display the \p Text, 78101409Stjr/// starting from the \p StartColumn on a terminal with the \p TabWidth. The 79101369Stjr/// text is assumed to use the specified \p Encoding. 80101369Stjrinline unsigned columnWidthWithTabs(StringRef Text, unsigned StartColumn, 81101369Stjr unsigned TabWidth, Encoding Encoding) { 82101369Stjr unsigned TotalWidth = 0; 83101369Stjr StringRef Tail = Text; 84101369Stjr for (;;) { 85101369Stjr StringRef::size_type TabPos = Tail.find('\t'); 86101369Stjr if (TabPos == StringRef::npos) 87101369Stjr return TotalWidth + columnWidth(Tail, Encoding); 88101253Stjr TotalWidth += columnWidth(Tail.substr(0, TabPos), Encoding); 8976633Stshiozak TotalWidth += TabWidth - (TotalWidth + StartColumn) % TabWidth; 9076633Stshiozak Tail = Tail.substr(TabPos + 1); 91101409Stjr } 92101409Stjr} 93101409Stjr 94101409Stjr/// \brief Gets the number of bytes in a sequence representing a single 95101409Stjr/// codepoint and starting with FirstChar in the specified Encoding. 96101409Stjrinline unsigned getCodePointNumBytes(char FirstChar, Encoding Encoding) { 97101409Stjr switch (Encoding) { 98101409Stjr case Encoding_UTF8: 99101409Stjr return getNumBytesForUTF8(FirstChar); 100101409Stjr default: 101101409Stjr return 1; 102101409Stjr } 103101409Stjr} 104101409Stjr 105101409Stjrinline bool isOctDigit(char c) { return '0' <= c && c <= '7'; } 106101253Stjr 107101369Stjrinline bool isHexDigit(char c) { 108101409Stjr return ('0' <= c && c <= '9') || ('a' <= c && c <= 'f') || 109101409Stjr ('A' <= c && c <= 'F'); 110101409Stjr} 111101409Stjr 112101409Stjr/// \brief Gets the length of an escape sequence inside a C++ string literal. 113101409Stjr/// Text should span from the beginning of the escape sequence (starting with a 114101409Stjr/// backslash) to the end of the string literal. 115101369Stjrinline unsigned getEscapeSequenceLength(StringRef Text) { 116101369Stjr assert(Text[0] == '\\'); 11776633Stshiozak if (Text.size() < 2) 118 return 1; 119 120 switch (Text[1]) { 121 case 'u': 122 return 6; 123 case 'U': 124 return 10; 125 case 'x': { 126 unsigned I = 2; // Point after '\x'. 127 while (I < Text.size() && isHexDigit(Text[I])) 128 ++I; 129 return I; 130 } 131 default: 132 if (isOctDigit(Text[1])) { 133 unsigned I = 1; 134 while (I < Text.size() && I < 4 && isOctDigit(Text[I])) 135 ++I; 136 return I; 137 } 138 return 1 + getNumBytesForUTF8(Text[1]); 139 } 140} 141 142} // namespace encoding 143} // namespace format 144} // namespace clang 145 146#endif 147