flang/lib/Evaluate/character.h

   1 //===-- lib/Evaluate/character.h --------------------------------*- C++ -*-===//
   2 //
   3 // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
   4 // See https://llvm.org/LICENSE.txt for license information.
   5 // SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
   6 //
   7 //===----------------------------------------------------------------------===//
   8
   9 #ifndef FORTRAN_EVALUATE_CHARACTER_H_
  10 #define FORTRAN_EVALUATE_CHARACTER_H_
  11
  12 #include "flang/Evaluate/type.h"
  13 #include <string>
  14
  15 // Provides implementations of intrinsic functions operating on character
  16 // scalars. No assumption is made regarding character encodings other than they
  17 // must be compatible with ASCII (else, NEW_LINE, ACHAR and IACHAR need to be
  18 // adapted).
  19
  20 namespace Fortran::evaluate {
  21
  22 template <int KIND> class CharacterUtils {
  23   using Character = Scalar<Type<TypeCategory::Character, KIND>>;
  24   using CharT = typename Character::value_type;
  25
  26 public:
  27   // CHAR also implements ACHAR under assumption that character encodings
  28   // contain ASCII
  29   static Character CHAR(std::uint64_t code) {
  30     return Character{{static_cast<CharT>(code)}};
  31   }
  32
  33   // ICHAR also implements IACHAR under assumption that character encodings
  34   // contain ASCII
  35   static std::int64_t ICHAR(const Character &c) {
  36     CHECK(c.length() == 1);
  37     if constexpr (std::is_same_v<CharT, char>) {
  38       // char may be signed, so cast it first to unsigned to avoid having
  39       // ichar(char(128_4)) returning -128
  40       return static_cast<unsigned char>(c[0]);
  41     } else {
  42       return c[0];
  43     }
  44   }
  45
  46   static Character NEW_LINE() { return Character{{NewLine()}}; }
  47
  48   static Character ADJUSTL(const Character &str) {
  49     auto pos{str.find_first_not_of(Space())};
  50     if (pos != Character::npos && pos != 0) {
  51       return Character{str.substr(pos) + Character(pos, Space())};
  52     }
  53     // else empty or only spaces, or no leading spaces
  54     return str;
  55   }
  56
  57   static Character ADJUSTR(const Character &str) {
  58     auto pos{str.find_last_not_of(Space())};
  59     if (pos != Character::npos && pos != str.length() - 1) {
  60       auto delta{str.length() - 1 - pos};
  61       return Character{Character(delta, Space()) + str.substr(0, pos + 1)};
  62     }
  63     // else empty or only spaces, or no trailing spaces
  64     return str;
  65   }
  66
  67   static ConstantSubscript INDEX(
  68       const Character &str, const Character &substr, bool back = false) {
  69     auto pos{back ? str.rfind(substr) : str.find(substr)};
  70     return static_cast<ConstantSubscript>(pos == str.npos ? 0 : pos + 1);
  71   }
  72
  73   static ConstantSubscript SCAN(
  74       const Character &str, const Character &set, bool back = false) {
  75     auto pos{back ? str.find_last_of(set) : str.find_first_of(set)};
  76     return static_cast<ConstantSubscript>(pos == str.npos ? 0 : pos + 1);
  77   }
  78
  79   static ConstantSubscript VERIFY(
  80       const Character &str, const Character &set, bool back = false) {
  81     auto pos{back ? str.find_last_not_of(set) : str.find_first_not_of(set)};
  82     return static_cast<ConstantSubscript>(pos == str.npos ? 0 : pos + 1);
  83   }
  84
  85   // Resize adds spaces on the right if the new size is bigger than the
  86   // original, or by trimming the rightmost characters otherwise.
  87   static Character Resize(const Character &str, std::size_t newLength) {
  88     auto oldLength{str.length()};
  89     if (newLength > oldLength) {
  90       return str + Character(newLength - oldLength, Space());
  91     } else {
  92       return str.substr(0, newLength);
  93     }
  94   }
  95
  96   static ConstantSubscript LEN_TRIM(const Character &str) {
  97     return VERIFY(str, Character{' '}, true);
  98   }
  99
 100   static Character REPEAT(const Character &str, ConstantSubscript ncopies) {
 101     Character result;
 102     if (!str.empty()) {
 103       while (ncopies-- > 0) {
 104         result += str;
 105       }
 106     }
 107     return result;
 108   }
 109
 110   static Character TRIM(const Character &str) {
 111     return str.substr(0, LEN_TRIM(str));
 112   }
 113
 114 private:
 115   // Following helpers assume that character encodings contain ASCII
 116   static constexpr CharT Space() { return 0x20; }
 117   static constexpr CharT NewLine() { return 0x0a; }
 118 };
 119
 120 } // namespace Fortran::evaluate
 121
 122 #endif // FORTRAN_EVALUATE_CHARACTER_H_