unified function to check if a character is valid in an identifier

[s311354]

Problem


The current implementation of is_ident1 and is_ident2 functions works without errors, but both cases obviously handle valid identifier characters dynamically based on whether it’s checking the first or subsequent characters.

// This function returns true if a given character is acceptable as the first character of an identifier.

bool is_ident1(uint32_t c) {
  static uint32_t range[] = {
    '_', '_', 'a', 'z', 'A', 'Z', '$', '$',
    0x00A8, 0x00A8, 0x00AA, 0x00AA, 0x00AD, 0x00AD, 0x00AF, 0x00AF,
    0x00B2, 0x00B5, 0x00B7, 0x00BA, 0x00BC, 0x00BE, 0x00C0, 0x00D6,
    0x00D8, 0x00F6, 0x00F8, 0x00FF, 0x0100, 0x02FF, 0x0370, 0x167F,
    0x1681, 0x180D, 0x180F, 0x1DBF, 0x1E00, 0x1FFF, 0x200B, 0x200D,
    0x202A, 0x202E, 0x203F, 0x2040, 0x2054, 0x2054, 0x2060, 0x206F,
    0x2070, 0x20CF, 0x2100, 0x218F, 0x2460, 0x24FF, 0x2776, 0x2793,
    0x2C00, 0x2DFF, 0x2E80, 0x2FFF, 0x3004, 0x3007, 0x3021, 0x302F,
    0x3031, 0x303F, 0x3040, 0xD7FF, 0xF900, 0xFD3D, 0xFD40, 0xFDCF,
    0xFDF0, 0xFE1F, 0xFE30, 0xFE44, 0xFE47, 0xFFFD,
    0x10000, 0x1FFFD, 0x20000, 0x2FFFD, 0x30000, 0x3FFFD, 0x40000, 0x4FFFD,
    0x50000, 0x5FFFD, 0x60000, 0x6FFFD, 0x70000, 0x7FFFD, 0x80000, 0x8FFFD,
    0x90000, 0x9FFFD, 0xA0000, 0xAFFFD, 0xB0000, 0xBFFFD, 0xC0000, 0xCFFFD,
    0xD0000, 0xDFFFD, 0xE0000, 0xEFFFD, -1,
  };

  return in_range(range, c);
}

// Check based on the context (first or subsequent character)

bool is_ident2(uint32_t c) {
  static uint32_t range[] = {
    '0', '9', '$', '$', 0x0300, 0x036F, 0x1DC0, 0x1DFF, 0x20D0, 0x20FF,
    0xFE20, 0xFE2F, -1,
  };

  return is_ident1(c) || in_range(range, c);
}

Proposal

I combined exactly is_ident1 and is_ident2 chunks of code into a single logic block in tokenize.c. This change would streamline the process of identifying valid characters while reading identifiers.


Retaining the functionality to distinguish valid identifier characters is also a critical part, and this will reduce code duplication and improve maintainability. we should probably enhance caching efficiency and CPU branch prediction due to consolidated logic, too. I would even propose to eliminate redundant calls to separate functions, but this really depends on your focus in this project.