Encoding.h revision 341825
190699Srobert//===--- Encoding.h - Format C++ code ---------------------------*- C++ -*-===// 290699Srobert// 390699Srobert// The LLVM Compiler Infrastructure 490699Srobert// 590699Srobert// This file is distributed under the University of Illinois Open Source 690699Srobert// License. See LICENSE.TXT for details. 790699Srobert// 890699Srobert//===----------------------------------------------------------------------===// 990699Srobert/// 1090699Srobert/// \file 1190699Srobert/// Contains functions for text encoding manipulation. Supports UTF-8, 1290699Srobert/// 8-bit encodings and escape sequences in C++ string literals. 1390699Srobert/// 1490699Srobert//===----------------------------------------------------------------------===// 1590699Srobert 1690699Srobert#ifndef LLVM_CLANG_LIB_FORMAT_ENCODING_H 1790699Srobert#define LLVM_CLANG_LIB_FORMAT_ENCODING_H 1890699Srobert 1990699Srobert#include "clang/Basic/LLVM.h" 2090699Srobert#include "llvm/ADT/StringRef.h" 2190699Srobert#include "llvm/Support/ConvertUTF.h" 2290699Srobert#include "llvm/Support/Unicode.h" 2390699Srobert 2490699Srobertnamespace clang { 2590699Srobertnamespace format { 2690699Srobertnamespace encoding { 2790699Srobert 2890699Srobertenum Encoding { 2990699Srobert Encoding_UTF8, 3090699Srobert Encoding_Unknown // We treat all other encodings as 8-bit encodings. 3190699Srobert}; 3290699Srobert 3390699Srobert/// Detects encoding of the Text. If the Text can be decoded using UTF-8, 3490699Srobert/// it is considered UTF8, otherwise we treat it as some 8-bit encoding. 3590699Srobertinline Encoding detectEncoding(StringRef Text) { 3690699Srobert const llvm::UTF8 *Ptr = reinterpret_cast<const llvm::UTF8 *>(Text.begin()); 3790699Srobert const llvm::UTF8 *BufEnd = reinterpret_cast<const llvm::UTF8 *>(Text.end()); 3890699Srobert if (llvm::isLegalUTF8String(&Ptr, BufEnd)) 3990699Srobert return Encoding_UTF8; 4090699Srobert return Encoding_Unknown; 4190699Srobert} 4290699Srobert 4390699Srobert/// Returns the number of columns required to display the \p Text on a 4490699Srobert/// generic Unicode-capable terminal. Text is assumed to use the specified 4590699Srobert/// \p Encoding. 4690699Srobertinline unsigned columnWidth(StringRef Text, Encoding Encoding) { 4790699Srobert if (Encoding == Encoding_UTF8) { 4890699Srobert int ContentWidth = llvm::sys::unicode::columnWidthUTF8(Text); 4990699Srobert // FIXME: Figure out the correct way to handle this in the presence of both 5090699Srobert // printable and unprintable multi-byte UTF-8 characters. Falling back to 5190699Srobert // returning the number of bytes may cause problems, as columnWidth suddenly 5290699Srobert // becomes non-additive. 5390699Srobert if (ContentWidth >= 0) 5490699Srobert return ContentWidth; 5590699Srobert } 5690699Srobert return Text.size(); 5790699Srobert} 5890699Srobert 5990699Srobert/// Returns the number of columns required to display the \p Text, 6090699Srobert/// starting from the \p StartColumn on a terminal with the \p TabWidth. The 6190699Srobert/// text is assumed to use the specified \p Encoding. 6290699Srobertinline unsigned columnWidthWithTabs(StringRef Text, unsigned StartColumn, 6390699Srobert unsigned TabWidth, Encoding Encoding) { 6490699Srobert unsigned TotalWidth = 0; 6590699Srobert StringRef Tail = Text; 6690699Srobert for (;;) { 6790699Srobert StringRef::size_type TabPos = Tail.find('\t'); 6890699Srobert if (TabPos == StringRef::npos) 6990699Srobert return TotalWidth + columnWidth(Tail, Encoding); 7090699Srobert TotalWidth += columnWidth(Tail.substr(0, TabPos), Encoding); 7190699Srobert TotalWidth += TabWidth - (TotalWidth + StartColumn) % TabWidth; 7290699Srobert Tail = Tail.substr(TabPos + 1); 7390699Srobert } 7490699Srobert} 7590699Srobert 7690699Srobert/// Gets the number of bytes in a sequence representing a single 7790699Srobert/// codepoint and starting with FirstChar in the specified Encoding. 7890699Srobertinline unsigned getCodePointNumBytes(char FirstChar, Encoding Encoding) { 7990699Srobert switch (Encoding) { 8090699Srobert case Encoding_UTF8: 8190699Srobert return llvm::getNumBytesForUTF8(FirstChar); 8290699Srobert default: 8390699Srobert return 1; 8490699Srobert } 8590699Srobert} 8690699Srobert 8790699Srobertinline bool isOctDigit(char c) { return '0' <= c && c <= '7'; } 8890699Srobert 8990699Srobertinline bool isHexDigit(char c) { 9090699Srobert return ('0' <= c && c <= '9') || ('a' <= c && c <= 'f') || 9190699Srobert ('A' <= c && c <= 'F'); 9290699Srobert} 9390699Srobert 9490699Srobert/// Gets the length of an escape sequence inside a C++ string literal. 9590699Srobert/// Text should span from the beginning of the escape sequence (starting with a 9690699Srobert/// backslash) to the end of the string literal. 9790699Srobertinline unsigned getEscapeSequenceLength(StringRef Text) { 9890699Srobert assert(Text[0] == '\\'); 9990699Srobert if (Text.size() < 2) 10090699Srobert return 1; 10190699Srobert 10290699Srobert switch (Text[1]) { 10390699Srobert case 'u': 10490699Srobert return 6; 10590699Srobert case 'U': 10690699Srobert return 10; 10790699Srobert case 'x': { 10890699Srobert unsigned I = 2; // Point after '\x'. 10990699Srobert while (I < Text.size() && isHexDigit(Text[I])) 11090699Srobert ++I; 11190699Srobert return I; 11290699Srobert } 11390699Srobert default: 11490699Srobert if (isOctDigit(Text[1])) { 11590699Srobert unsigned I = 1; 11690699Srobert while (I < Text.size() && I < 4 && isOctDigit(Text[I])) 11790699Srobert ++I; 11890699Srobert return I; 11990699Srobert } 12090699Srobert return 1 + llvm::getNumBytesForUTF8(Text[1]); 12190699Srobert } 12290699Srobert} 12390699Srobert 12490699Srobert} // namespace encoding 12590699Srobert} // namespace format 12690699Srobert} // namespace clang 12790699Srobert 12890699Srobert#endif 12990699Srobert