clang/test/Lexer/char-literal.cpp

   1 // RUN: %clang_cc1 -triple x86_64-apple-darwin -std=c++03 -Wfour-char-constants -fsyntax-only -verify=cxx03,expected %s
   2 // RUN: %clang_cc1 -triple x86_64-apple-darwin -std=c++11 -Wfour-char-constants -fsyntax-only -verify=cxx,expected %s
   3 // RUN: %clang_cc1 -triple x86_64-apple-darwin -std=c++17 -Wfour-char-constants -fsyntax-only -verify=cxx,expected %s
   4 // RUN: %clang_cc1 -triple x86_64-apple-darwin -std=c++20 -Wfour-char-constants -fsyntax-only -verify=cxx,expected %s
   5 // RUN: %clang_cc1 -triple x86_64-apple-darwin -std=c11 -x c -Wfour-char-constants -fsyntax-only -verify=c11,expected %s
   6 // RUN: %clang_cc1 -triple x86_64-apple-darwin -std=c2x -x c -Wfour-char-constants -fsyntax-only -verify=c2x,expected %s
   7
   8 #ifndef __cplusplus
   9 typedef __WCHAR_TYPE__ wchar_t;
  10 typedef __CHAR16_TYPE__ char16_t;
  11 typedef __CHAR32_TYPE__ char32_t;
  12 #endif
  13
  14 int a = 'ab'; // expected-warning {{multi-character character constant}}
  15 int b = '\xFF\xFF'; // expected-warning {{multi-character character constant}}
  16 int c = 'APPS'; // expected-warning {{multi-character character constant}}
  17
  18 char d = '⌘'; // expected-error {{character too large for enclosing character literal type}}
  19 char e = '\u2318'; // expected-error {{character too large for enclosing character literal type}}
  20
  21 #if !defined(__cplusplus) || __cplusplus > 201100L
  22 #ifdef __cplusplus
  23 auto f = '\xE2\x8C\x98'; // expected-warning {{multi-character character constant}}
  24 #endif
  25
  26 char16_t g = u'ab'; // expected-error {{Unicode character literals may not contain multiple characters}}
  27 char16_t h = u'\U0010FFFD'; // expected-error {{character too large for enclosing character literal type}}
  28
  29 wchar_t i = L'ab'; // expected-error {{wide character literals may not contain multiple characters}}
  30
  31 wchar_t j = L'\U0010FFFD';
  32
  33 char32_t k = U'\U0010FFFD';
  34
  35 char l = 'Ø'; // expected-error {{character too large for enclosing character literal type}}
  36 char m = '👿'; // expected-error {{character too large for enclosing character literal type}}
  37
  38 char32_t n = U'ab'; // expected-error {{Unicode character literals may not contain multiple characters}}
  39 char16_t o = '👽'; // expected-error {{character too large for enclosing character literal type}}
  40
  41 char16_t p[2] = u"\U0000FFFF";
  42 char16_t q[2] = u"\U00010000";
  43 #ifdef __cplusplus
  44 // expected-error@-2 {{too long}}
  45 #endif
  46
  47 // UTF-8 character literal code point ranges.
  48 #if __cplusplus >= 201703L || __STDC_VERSION__ >= 201710L
  49 _Static_assert(u8'\U00000000' == 0x00, ""); // c11-error {{universal character name refers to a control character}}
  50 _Static_assert(u8'\U0000007F' == 0x7F, ""); // c11-error {{universal character name refers to a control character}}
  51 _Static_assert(u8'\U00000080', ""); // c11-error {{universal character name refers to a control character}}
  52                                     // cxx-error@-1 {{character too large for enclosing character literal type}}
  53                                     // c2x-error@-2 {{character too large for enclosing character literal type}}
  54 _Static_assert((unsigned char)u8'\xFF' == (unsigned char)0xFF, "");
  55 #endif
  56
  57 // UTF-8 string literal code point ranges.
  58 _Static_assert(u8"\U00000000"[0] == 0x00, ""); // c11-error {{universal character name refers to a control character}}
  59 _Static_assert(u8"\U0000007F"[0] == 0x7F, ""); // c11-error {{universal character name refers to a control character}}
  60 _Static_assert((unsigned char)u8"\U00000080"[0] == (unsigned char)0xC2, ""); // c11-error {{universal character name refers to a control character}}
  61 _Static_assert((unsigned char)u8"\U00000080"[1] == (unsigned char)0x80, ""); // c11-error {{universal character name refers to a control character}}
  62 _Static_assert((unsigned char)u8"\U000007FF"[0] == (unsigned char)0xDF, "");
  63 _Static_assert((unsigned char)u8"\U000007FF"[1] == (unsigned char)0xBF, "");
  64 _Static_assert((unsigned char)u8"\U00000800"[0] == (unsigned char)0xE0, "");
  65 _Static_assert((unsigned char)u8"\U00000800"[1] == (unsigned char)0xA0, "");
  66 _Static_assert((unsigned char)u8"\U00000800"[2] == (unsigned char)0x80, "");
  67 _Static_assert(u8"\U0000D800"[0], ""); // expected-error {{invalid universal character}}
  68 _Static_assert(u8"\U0000DFFF"[0], ""); // expected-error {{invalid universal character}}
  69 _Static_assert((unsigned char)u8"\U0000FFFF"[0] == (unsigned char)0xEF, "");
  70 _Static_assert((unsigned char)u8"\U0000FFFF"[1] == (unsigned char)0xBF, "");
  71 _Static_assert((unsigned char)u8"\U0000FFFF"[2] == (unsigned char)0xBF, "");
  72 _Static_assert((unsigned char)u8"\U00010000"[0] == (unsigned char)0xF0, "");
  73 _Static_assert((unsigned char)u8"\U00010000"[1] == (unsigned char)0x90, "");
  74 _Static_assert((unsigned char)u8"\U00010000"[2] == (unsigned char)0x80, "");
  75 _Static_assert((unsigned char)u8"\U00010000"[3] == (unsigned char)0x80, "");
  76 _Static_assert((unsigned char)u8"\U0010FFFF"[0] == (unsigned char)0xF4, "");
  77 _Static_assert((unsigned char)u8"\U0010FFFF"[1] == (unsigned char)0x8F, "");
  78 _Static_assert((unsigned char)u8"\U0010FFFF"[2] == (unsigned char)0xBF, "");
  79 _Static_assert((unsigned char)u8"\U0010FFFF"[3] == (unsigned char)0xBF, "");
  80 _Static_assert(u8"\U00110000"[0], ""); // expected-error {{invalid universal character}}
  81
  82 #if !defined(__STDC_UTF_16__)
  83 #error __STDC_UTF_16__ is not defined.
  84 #endif
  85 #if __STDC_UTF_16__ != 1
  86 #error __STDC_UTF_16__ has the wrong value.
  87 #endif
  88
  89 // UTF-16 character literal code point ranges.
  90 _Static_assert(u'\U00000000' == 0x0000, ""); // c11-error {{universal character name refers to a control character}}
  91 _Static_assert(u'\U0000D800', ""); // expected-error {{invalid universal character}}
  92 _Static_assert(u'\U0000DFFF', ""); // expected-error {{invalid universal character}}
  93 _Static_assert(u'\U0000FFFF' == 0xFFFF, "");
  94 _Static_assert(u'\U00010000', ""); // expected-error {{character too large for enclosing character literal type}}
  95
  96 // UTF-16 string literal code point ranges.
  97 _Static_assert(u"\U00000000"[0] == 0x0000, ""); // c11-error {{universal character name refers to a control character}}
  98 _Static_assert(u"\U0000D800"[0], ""); // expected-error {{invalid universal character}}
  99 _Static_assert(u"\U0000DFFF"[0], ""); // expected-error {{invalid universal character}}
 100 _Static_assert(u"\U0000FFFF"[0] == 0xFFFF, "");
 101 _Static_assert(u"\U00010000"[0] == 0xD800, "");
 102 _Static_assert(u"\U00010000"[1] == 0xDC00, "");
 103 _Static_assert(u"\U0010FFFF"[0] == 0xDBFF, "");
 104 _Static_assert(u"\U0010FFFF"[1] == 0xDFFF, "");
 105 _Static_assert(u"\U00110000"[0], ""); // expected-error {{invalid universal character}}
 106
 107 #if !defined(__STDC_UTF_32__)
 108 #error __STDC_UTF_32__ is not defined.
 109 #endif
 110 #if __STDC_UTF_32__ != 1
 111 #error __STDC_UTF_32__ has the wrong value.
 112 #endif
 113
 114 // UTF-32 character literal code point ranges.
 115 _Static_assert(U'\U00000000' == 0x00000000, ""); // c11-error {{universal character name refers to a control character}}
 116 _Static_assert(U'\U0010FFFF' == 0x0010FFFF, "");
 117 _Static_assert(U'\U00110000', ""); // expected-error {{invalid universal character}}
 118
 119 // UTF-32 string literal code point ranges.
 120 _Static_assert(U"\U00000000"[0] == 0x00000000, ""); // c11-error {{universal character name refers to a control character}}
 121 _Static_assert(U"\U0000D800"[0], ""); // expected-error {{invalid universal character}}
 122 _Static_assert(U"\U0000DFFF"[0], ""); // expected-error {{invalid universal character}}
 123 _Static_assert(U"\U0010FFFF"[0] == 0x0010FFFF, "");
 124 _Static_assert(U"\U00110000"[0], ""); // expected-error {{invalid universal character}}
 125
 126 #endif // !defined(__cplusplus) || __cplusplus > 201100L
 127
 128 _Static_assert('\u0024' == '$', "");
 129 _Static_assert('\u0040' == '@', "");
 130 _Static_assert('\u0060' == '`', "");
 131
 132 _Static_assert('\u0061' == 'a', "");  // c11-error {{character 'a' cannot be specified by a universal character name}} \
 133                                       // cxx03-error {{character 'a' cannot be specified by a universal character name}}
 134 _Static_assert('\u0000' == '\0', ""); // c11-error {{universal character name refers to a control character}} \
 135                                       // cxx03-error {{universal character name refers to a control character}}