Convert C struct into Kaitai Struct spec (.ksy)

c-struct-to-ksy.rb

#!/usr/bin/env ruby

require 'yaml'

TYPE_TO_KSY = {
  'uint8_t' => 'u1',
  'uint16_t' => 'u2',
  'uint32_t' => 'u4',
  'uint64_t' => 'u8',

  'int8_t' => 's1',
  'int16_t' => 's2',
  'int32_t' => 's4',
  'int64_t' => 's8',

  'char' => 's1',
  'short' => 's2',
  'int' => 's4',
  'uint' => 'u4',
  'long' => 's8',
  'float' => 'f4',
  'double' => 'f8',

  'WORD' => 'u2',
  'DWORD' => 'u4',

  '__s32' => 's4',
  '__s64' => 's8',

  'off_t' => 'u8',  # __int64_t
}

def parse_body(body)
  s = []
  body.gsub(/\s*(?<type>(?:[A-Za-z0-9_]+|(?:(?:un)?signed\s+)?(?:char|(?:(?:short|(?:long\s+)?long)\s+)?(int))?|(?:float|(?:long\s+)?double)))\s+(?<pointer>\*)*(?<name>[A-Za-z0-9_]+)(?:\s*\[(?<array>[0-9]+)\])?(?:\s*:\s*(?<bit>[0-8]))?\s*;(\s*\/\*\s*(?<comment_body>.*?)\s*\*\/)?/) {
    type, pointer, name, array, bit, comment, comment_body = $1, $2, $3, $4, $5, $6, $7
    h = {'id' => name, 'type' => TYPE_TO_KSY[type] || type}
    if pointer and type == 'char' and array
      h['type'] = 'str'
      h['encoding'] = 'ASCII'
      h['size'] = array.to_i
    end
    if bit
      h['type'] = 'b' + bit
    end
    if array and type != 'char'
      h['repeat'] = 'expr' 
      h['repeat-expr'] = array.to_i
    end
    h['doc'] = comment_body unless comment_body.nil? or comment_body.empty?
    s << h
  }
  s
  r = {'seq' => s}  
end
r = {}

$stdin.read.gsub(/struct\s+([A-Za-z0-9_]+)?\s*\{(.*?)\}\s*(?:[A-Z_]+\s*)?(\w*?);/m) {
  tag, body, name = $1, $2, $3
  name = name.split(/\s*,\s*/)[0]
  name = tag if name.nil? or name.empty?
  r[name] = parse_body(body)
}

puts({'types' => r}.to_yaml)