1341825Sdim//===--- Encoding.h - Format C++ code ---------------------------*- C++ -*-===// 2259701Sdim// 3353358Sdim// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions. 4353358Sdim// See https://llvm.org/LICENSE.txt for license information. 5353358Sdim// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception 6259701Sdim// 7259701Sdim//===----------------------------------------------------------------------===// 8259701Sdim/// 9259701Sdim/// \file 10341825Sdim/// Contains functions for text encoding manipulation. Supports UTF-8, 11259701Sdim/// 8-bit encodings and escape sequences in C++ string literals. 12259701Sdim/// 13259701Sdim//===----------------------------------------------------------------------===// 14259701Sdim 15280031Sdim#ifndef LLVM_CLANG_LIB_FORMAT_ENCODING_H 16280031Sdim#define LLVM_CLANG_LIB_FORMAT_ENCODING_H 17259701Sdim 18259701Sdim#include "clang/Basic/LLVM.h" 19309124Sdim#include "llvm/ADT/StringRef.h" 20259701Sdim#include "llvm/Support/ConvertUTF.h" 21259701Sdim#include "llvm/Support/Unicode.h" 22259701Sdim 23259701Sdimnamespace clang { 24259701Sdimnamespace format { 25259701Sdimnamespace encoding { 26259701Sdim 27259701Sdimenum Encoding { 28259701Sdim Encoding_UTF8, 29259701Sdim Encoding_Unknown // We treat all other encodings as 8-bit encodings. 30259701Sdim}; 31259701Sdim 32341825Sdim/// Detects encoding of the Text. If the Text can be decoded using UTF-8, 33259701Sdim/// it is considered UTF8, otherwise we treat it as some 8-bit encoding. 34259701Sdiminline Encoding detectEncoding(StringRef Text) { 35314564Sdim const llvm::UTF8 *Ptr = reinterpret_cast<const llvm::UTF8 *>(Text.begin()); 36314564Sdim const llvm::UTF8 *BufEnd = reinterpret_cast<const llvm::UTF8 *>(Text.end()); 37314564Sdim if (llvm::isLegalUTF8String(&Ptr, BufEnd)) 38259701Sdim return Encoding_UTF8; 39259701Sdim return Encoding_Unknown; 40259701Sdim} 41259701Sdim 42341825Sdim/// Returns the number of columns required to display the \p Text on a 43259701Sdim/// generic Unicode-capable terminal. Text is assumed to use the specified 44259701Sdim/// \p Encoding. 45259701Sdiminline unsigned columnWidth(StringRef Text, Encoding Encoding) { 46259701Sdim if (Encoding == Encoding_UTF8) { 47259701Sdim int ContentWidth = llvm::sys::unicode::columnWidthUTF8(Text); 48276479Sdim // FIXME: Figure out the correct way to handle this in the presence of both 49276479Sdim // printable and unprintable multi-byte UTF-8 characters. Falling back to 50276479Sdim // returning the number of bytes may cause problems, as columnWidth suddenly 51276479Sdim // becomes non-additive. 52259701Sdim if (ContentWidth >= 0) 53259701Sdim return ContentWidth; 54259701Sdim } 55259701Sdim return Text.size(); 56259701Sdim} 57259701Sdim 58341825Sdim/// Returns the number of columns required to display the \p Text, 59259701Sdim/// starting from the \p StartColumn on a terminal with the \p TabWidth. The 60259701Sdim/// text is assumed to use the specified \p Encoding. 61259701Sdiminline unsigned columnWidthWithTabs(StringRef Text, unsigned StartColumn, 62259701Sdim unsigned TabWidth, Encoding Encoding) { 63259701Sdim unsigned TotalWidth = 0; 64259701Sdim StringRef Tail = Text; 65259701Sdim for (;;) { 66259701Sdim StringRef::size_type TabPos = Tail.find('\t'); 67259701Sdim if (TabPos == StringRef::npos) 68259701Sdim return TotalWidth + columnWidth(Tail, Encoding); 69276479Sdim TotalWidth += columnWidth(Tail.substr(0, TabPos), Encoding); 70360784Sdim if (TabWidth) 71360784Sdim TotalWidth += TabWidth - (TotalWidth + StartColumn) % TabWidth; 72259701Sdim Tail = Tail.substr(TabPos + 1); 73259701Sdim } 74259701Sdim} 75259701Sdim 76341825Sdim/// Gets the number of bytes in a sequence representing a single 77259701Sdim/// codepoint and starting with FirstChar in the specified Encoding. 78259701Sdiminline unsigned getCodePointNumBytes(char FirstChar, Encoding Encoding) { 79259701Sdim switch (Encoding) { 80259701Sdim case Encoding_UTF8: 81314564Sdim return llvm::getNumBytesForUTF8(FirstChar); 82259701Sdim default: 83259701Sdim return 1; 84259701Sdim } 85259701Sdim} 86259701Sdim 87259701Sdiminline bool isOctDigit(char c) { return '0' <= c && c <= '7'; } 88259701Sdim 89259701Sdiminline bool isHexDigit(char c) { 90259701Sdim return ('0' <= c && c <= '9') || ('a' <= c && c <= 'f') || 91259701Sdim ('A' <= c && c <= 'F'); 92259701Sdim} 93259701Sdim 94341825Sdim/// Gets the length of an escape sequence inside a C++ string literal. 95259701Sdim/// Text should span from the beginning of the escape sequence (starting with a 96259701Sdim/// backslash) to the end of the string literal. 97259701Sdiminline unsigned getEscapeSequenceLength(StringRef Text) { 98259701Sdim assert(Text[0] == '\\'); 99259701Sdim if (Text.size() < 2) 100259701Sdim return 1; 101259701Sdim 102259701Sdim switch (Text[1]) { 103259701Sdim case 'u': 104259701Sdim return 6; 105259701Sdim case 'U': 106259701Sdim return 10; 107259701Sdim case 'x': { 108259701Sdim unsigned I = 2; // Point after '\x'. 109259701Sdim while (I < Text.size() && isHexDigit(Text[I])) 110259701Sdim ++I; 111259701Sdim return I; 112259701Sdim } 113259701Sdim default: 114259701Sdim if (isOctDigit(Text[1])) { 115259701Sdim unsigned I = 1; 116259701Sdim while (I < Text.size() && I < 4 && isOctDigit(Text[I])) 117259701Sdim ++I; 118259701Sdim return I; 119259701Sdim } 120314564Sdim return 1 + llvm::getNumBytesForUTF8(Text[1]); 121259701Sdim } 122259701Sdim} 123259701Sdim 124259701Sdim} // namespace encoding 125259701Sdim} // namespace format 126259701Sdim} // namespace clang 127259701Sdim 128280031Sdim#endif 129