Method wchar_width

src/screen.cpp:345–426

return character display width, 0 (non-spacing or invalid character), 1 (single width) or 2 (double width)

// return character display width, 0 (non-spacing or invalid character), 1 (single width) or 2 (double width)
int Screen::wchar_width(uint32_t wc)
{
  // ignore invisible characters, such as invalid UTF-8
  if (wc == 0)
    return 0;

  // control characters are double width to display them e.g. as \t or ^I
  if (wc < 0x20 || wc == 0x7f)
    return 2;

  /* Based on https://www.cl.cam.ac.uk/~mgk25/ucs/wcwidth.c with full table
     generated by "uniset +cat=Me +cat=Mn +cat=Cf -00AD +1160-11FF +200B c"
     from https://www.unicode.org/Public/UCD/latest/ucd/EastAsianWidth.txt
     this is a compressed table of combining character ranges, stored as
     range.first << 8 | (range.last - range.first)

     We don't use wcwidth() because it is broken on some systems, e.g. MacOS
     wcwidth(0x1f600) returns -1 for the emoticon U+1F600
  */
  static const uint32_t combining[] = {
    0x3006f, 0x48303, 0x48801, 0x5912c, 0x5bf00, 0x5c101, 0x5c401, 0x5c700,
    0x60003, 0x61005, 0x64b13, 0x67000, 0x6d60e, 0x6e701, 0x6ea03, 0x70f00,
    0x71100, 0x7301a, 0x7a60a, 0x7eb08, 0x90101, 0x93c00, 0x94107, 0x94d00,
    0x95103, 0x96201, 0x98100, 0x9bc00, 0x9c103, 0x9cd00, 0x9e201, 0xa0101,
    0xa3c00, 0xa4101, 0xa4701, 0xa4b02, 0xa7001, 0xa8101, 0xabc00, 0xac104,
    0xac701, 0xacd00, 0xae201, 0xb0100, 0xb3c00, 0xb3f00, 0xb4102, 0xb4d00,
    0xb5600, 0xb8200, 0xbc000, 0xbcd00, 0xc3e02, 0xc4602, 0xc4a03, 0xc5501,
    0xcbc00, 0xcbf00, 0xcc600, 0xccc01, 0xce201, 0xd4102, 0xd4d00, 0xdca00,
    0xdd202, 0xdd600, 0xe3100, 0xe3406, 0xe4707, 0xeb100, 0xeb405, 0xebb01,
    0xec805, 0xf1801, 0xf3500, 0xf3700, 0xf3900, 0xf710d, 0xf8004, 0xf8601,
    0xf9007, 0xf9923, 0xfc600,
    0x102d03, 0x103200, 0x103601, 0x103900, 0x105801, 0x11609f, 0x135f00,
    0x171202, 0x173202, 0x175201, 0x177201, 0x17b401, 0x17b706, 0x17c600,
    0x17c90a, 0x17dd00, 0x180b02, 0x18a900, 0x192002, 0x192701, 0x193200,
    0x193902, 0x1a1701, 0x1b0003, 0x1b3400, 0x1b3604, 0x1b3c00, 0x1b4200,
    0x1b6b08, 0x1dc00a, 0x1dfe01, 0x200b04, 0x202a04, 0x206003, 0x206a05,
    0x20d01f, 0x302a05, 0x309901, 0xa80600, 0xa80b00, 0xa82501, 0xfb1e00,
    0xfe000f, 0xfe2003, 0xfeff00, 0xfff902,
    0x10a0102, 0x10a0501, 0x10a0c03, 0x10a3802, 0x10a3f00, 0x1d16702, 0x1d1730f,
    0x1d18506, 0x1d1aa03, 0x1d24202, 0xe000100, 0xe00205f, 0xe0100ef,
  };

  int min = 0;
  int max = sizeof(combining) / sizeof(uint32_t) - 1;

  // binary search in table of non-spacing characters
  if (wc >= (combining[0] >> 8) && wc <= (combining[max] >> 8) + (combining[max] & 0xff))
  {
    while (max >= min)
    {
      int mid = (min + max) / 2;
      if (wc < (combining[mid] >> 8))
        max = mid - 1;
      else if (wc > (combining[mid] >> 8) + (combining[mid] & 0xff))
        min = mid + 1;
      else
        return 0;
    }