<% if @grammar.modulename %> module <%= @grammar.modulename %>; <% end %> class <%= classname %> { class Decoder { enum { CODE_POINT_INVALID = 0xFFFFFFFE, CODE_POINT_EOF = 0xFFFFFFFF, } static uint decode_code_point(const(ubyte) * input, size_t input_length, size_t * code_point_length) { if (input_length == 0u) { return CODE_POINT_EOF; } ubyte c = *input; uint result; if ((c & 0x80u) == 0u) { result = c; *code_point_length = 1u; } else { ubyte following_bytes; if ((c & 0xE0u) == 0xC0u) { result = c & 0x1Fu; following_bytes = 1u; } else if ((c & 0xF0u) == 0xE0u) { result = c & 0x0Fu; following_bytes = 2u; } else if ((c & 0xF8u) == 0xF0u) { result = c & 0x07u; following_bytes = 3u; } else if ((c & 0xFCu) == 0xF8u) { result = c & 0x03u; following_bytes = 4u; } else if ((c & 0xFEu) == 0xFCu) { result = c & 0x01u; following_bytes = 5u; } if (input_length <= following_bytes) { return CODE_POINT_INVALID; } *code_point_length = following_bytes + 1u; while (following_bytes-- > 0u) { input++; result <<= 6u; result |= *input & 0x3Fu; } } return result; } } class Lexer { enum { <% @grammar.tokens.each_with_index do |token, index| %> <% if token.name %> TOKEN_<%= token.c_name %> = <%= index %>, <% end %> <% end %> TOKEN_EOF = <%= TOKEN_EOF %>, TOKEN_DECODE_ERROR = <%= TOKEN_DECODE_ERROR %>, TOKEN_DROP = <%= TOKEN_DROP %>, TOKEN_NONE = <%= TOKEN_NONE %>, } static immutable string TokenNames[] = [ <% @grammar.tokens.each_with_index do |token, index| %> <% if token.name %> "<%= token.name %>", <% else %> null, <% end %> <% end %> ]; private struct Transition { uint first; uint last; uint destination; } private struct State { uint transition_table_index; uint n_transitions; uint accepts; } <% transition_table, state_table = lexer_dfa.build_tables %> private static const Transition transitions[] = [ <% transition_table.each do |transition_table_entry| %> Transition(<%= transition_table_entry[:first] %>u, <%= transition_table_entry[:last] %>u, <%= transition_table_entry[:destination] %>u), <% end %> ]; private static const State states[] = [ <% state_table.each do |state_table_entry| %> State(<%= state_table_entry[:transition_table_index] %>u, <%= state_table_entry[:n_transitions] %>u, <%= state_table_entry[:accepts] %>u), <% end %> ]; struct LexedToken { size_t row; size_t col; uint token; } static string[] lex(const(ubyte)[] input) { string[] tokens; string token = lex_token(input.ptr, input.length); if (token !is null) { tokens ~= token; } return tokens; } private static string lex_token(const(ubyte) * input, size_t input_length) { size_t code_point_length; uint code_point = Decoder.decode_code_point(input, input_length, &code_point_length); return null; } private const(ubyte) * m_input; private size_t m_input_length; private size_t m_input_position; private size_t m_input_row; private size_t m_input_col; this(const(ubyte) * input, size_t input_length) { m_input = input; m_input_length = input_length; } LexedToken lex_token() { LexedToken lt = LexedToken(m_input_row, m_input_col, TOKEN_NONE); size_t token_length; size_t token_attempt_length; uint current_state; for (;;) { size_t code_point_length; uint code_point = Decoder.decode_code_point(&m_input[m_input_position], m_input_length - m_input_position, &code_point_length); if (code_point == Decoder.CODE_POINT_INVALID) { lt.token = TOKEN_DECODE_ERROR; return lt; } bool lex_continue = false; if (code_point != Decoder.CODE_POINT_EOF) { uint dest = transition(current_state, code_point); } } return lt; } private uint transition(uint current_state, uint code_point) { uint transition_table_index = states[current_state].transition_table_index; for (uint i = 0u; i < states[current_state].n_transitions; i++) { if ((transitions[transition_table_index].first <= code_point) && (code_point <= transitions[transition_table_index].last)) { return transitions[transition_table_index].destination; } } return cast(uint)-1; } } }