2833 lines
74 KiB
Plaintext
2833 lines
74 KiB
Plaintext
(*
|
|
* This Source Code Form is subject to the terms of the Mozilla Public License,
|
|
* v. 2.0. If a copy of the MPL was not distributed with this file, You can
|
|
* obtain one at https://mozilla.org/MPL/2.0/.
|
|
*)
|
|
|
|
(* Stage 15 compiler. *)
|
|
|
|
type
|
|
LexerAction = (none, accumulate, skip, single, eof, finalize, composite, key_id, integer, delimited);
|
|
|
|
(**
|
|
* Classification table assigns each possible character to a group (class). All
|
|
* characters of the same group a handled equivalently.
|
|
*
|
|
* Transition = record
|
|
* action: TransitionAction;
|
|
* next_state: TransitionState
|
|
* end;
|
|
*)
|
|
LexerClass = (
|
|
invalid,
|
|
digit,
|
|
alpha,
|
|
space,
|
|
colon,
|
|
equals,
|
|
left_paren,
|
|
right_paren,
|
|
asterisk,
|
|
backslash,
|
|
single,
|
|
hex,
|
|
zero,
|
|
x,
|
|
eof,
|
|
dot,
|
|
minus,
|
|
single_quote,
|
|
double_quote,
|
|
greater,
|
|
less,
|
|
other
|
|
);
|
|
LexerState = (
|
|
start,
|
|
colon,
|
|
identifier,
|
|
decimal,
|
|
leading_zero,
|
|
greater,
|
|
minus,
|
|
left_paren,
|
|
less,
|
|
dot,
|
|
comment,
|
|
closing_comment,
|
|
character,
|
|
character_escape,
|
|
string,
|
|
string_escape,
|
|
finish
|
|
);
|
|
LexerTokenKind = (
|
|
identifier,
|
|
_const,
|
|
_var,
|
|
_proc,
|
|
_type,
|
|
_begin,
|
|
_end,
|
|
_if,
|
|
_then,
|
|
_else,
|
|
_elsif,
|
|
_while,
|
|
_do,
|
|
_extern,
|
|
_record,
|
|
_union,
|
|
_true,
|
|
_false,
|
|
null,
|
|
and,
|
|
_or,
|
|
_xor,
|
|
pipe,
|
|
not,
|
|
_return,
|
|
_module,
|
|
_program,
|
|
_import,
|
|
_cast,
|
|
_defer,
|
|
_case,
|
|
_of,
|
|
trait,
|
|
left_paren,
|
|
right_paren,
|
|
left_square,
|
|
right_square,
|
|
shift_left,
|
|
shift_right,
|
|
greater_equal,
|
|
less_equal,
|
|
greater_than,
|
|
less_than,
|
|
not_equal,
|
|
equals,
|
|
semicolon,
|
|
dot,
|
|
comma,
|
|
plus,
|
|
arrow,
|
|
minus,
|
|
multiplication,
|
|
division,
|
|
remainder,
|
|
assignment,
|
|
colon,
|
|
hat,
|
|
at,
|
|
comment,
|
|
string,
|
|
character,
|
|
integer,
|
|
word,
|
|
_goto,
|
|
eof
|
|
);
|
|
NodeKind := (
|
|
integer_literal,
|
|
string_literal,
|
|
character_literal
|
|
);
|
|
|
|
const
|
|
symbol_builtin_name_int := "Int";
|
|
symbol_builtin_name_word := "Word";
|
|
symbol_builtin_name_pointer := "Pointer";
|
|
symbol_builtin_name_char := "Char";
|
|
symbol_builtin_name_array := "Array";
|
|
|
|
(* Every type info starts with a word describing what type it is.
|
|
|
|
PRIMITIVE_TYPE = 1
|
|
ENUMERATION_TYPE = 2
|
|
|
|
Primitive types have only type size. *)
|
|
symbol_builtin_type_int := S(1, 4);
|
|
symbol_builtin_type_word := S(1, 4);
|
|
symbol_builtin_type_pointer := S(1, 4);
|
|
symbol_builtin_type_char := S(1, 1);
|
|
symbol_builtin_type_array := S(1, 4);
|
|
|
|
(* Info objects start with a word describing its type.
|
|
|
|
TYPE_INFO = 1
|
|
PARAMETER_INFO = 2
|
|
TEMPORARY_INFO = 3
|
|
|
|
Type info has the type it belongs to. *)
|
|
symbol_type_info_int := S(1, @symbol_builtin_type_int);
|
|
symbol_type_info_word := S(1, @symbol_builtin_type_word);
|
|
symbol_type_info_pointer := S(1, @symbol_builtin_type_pointer);
|
|
symbol_type_info_char := S(1, @symbol_builtin_type_char);
|
|
symbol_type_info_array := S(1, @symbol_builtin_type_array);
|
|
|
|
var
|
|
source_code: Array;
|
|
compiler_strings: Array;
|
|
symbol_table_global: Array;
|
|
symbol_table_local: Array;
|
|
classification: Array;
|
|
|
|
(* To reserve memory just add the value of needed bytes to the memory_free_pointer_variable. *)
|
|
memory: Array;
|
|
|
|
compiler_strings_position: Pointer := @compiler_strings;
|
|
compiler_strings_length: Word := 0;
|
|
label_counter: Word := 0;
|
|
|
|
(* Points to a segment of free memory. *)
|
|
memory_free_pointer: Word := @memory;
|
|
|
|
(**
|
|
* Calculates and returns the string token length between quotes, including the
|
|
* escaping slash characters.
|
|
*
|
|
* Parameters:
|
|
* string - String token pointer.
|
|
*
|
|
* Returns the length in a0.
|
|
*)
|
|
proc _string_length(string: Word);
|
|
var
|
|
counter: Word;
|
|
begin
|
|
(* Reset the counter. *)
|
|
counter := 0;
|
|
|
|
.string_length_loop;
|
|
string := string + 1;
|
|
|
|
if _load_byte(string) <> '"' then
|
|
counter := counter + 1;
|
|
goto string_length_loop
|
|
end;
|
|
|
|
return counter
|
|
end;
|
|
|
|
(**
|
|
* Adds a string to the global, read-only string storage.
|
|
*
|
|
* Parameters:
|
|
* string - String token.
|
|
*
|
|
* Returns the offset from the beginning of the storage to the new string in a0.
|
|
*)
|
|
proc _add_string(string: Word);
|
|
var
|
|
contents: Word;
|
|
result: Word;
|
|
current_byte: Word;
|
|
begin
|
|
contents := string + 1;
|
|
result := compiler_strings_length;
|
|
|
|
.add_string_loop;
|
|
if _load_byte(contents) <> '"' then
|
|
current_byte := _load_byte(contents);
|
|
_store_byte(current_byte, compiler_strings_position);
|
|
compiler_strings_position := compiler_strings_position + 1;
|
|
contents := contents + 1;
|
|
|
|
if current_byte <> '\\' then
|
|
compiler_strings_length := compiler_strings_length + 1
|
|
end;
|
|
goto add_string_loop
|
|
end;
|
|
|
|
return result
|
|
end;
|
|
|
|
(**
|
|
* Reads standard input into a buffer.
|
|
*
|
|
* Parameters:
|
|
* buffer - Buffer pointer.
|
|
* size - Buffer size.
|
|
*
|
|
* Returns the amount of bytes written in a0.
|
|
*)
|
|
proc _read_file(buffer: Word, size: Word);
|
|
return _syscall(0, buffer, size, 0, 0, 0, 63)
|
|
end;
|
|
|
|
(**
|
|
* Writes to the standard output.
|
|
*
|
|
* Parameters:
|
|
* buffer - Buffer.
|
|
* size - Buffer length.
|
|
*)
|
|
proc _write_s(buffer: Word, size: Word);
|
|
begin
|
|
_syscall(1, buffer, size, 0, 0, 0, 64)
|
|
end;
|
|
|
|
(**
|
|
* Writes a number to a string buffer.
|
|
*
|
|
* Parameters:
|
|
* number - Whole number.
|
|
* output_buffer - Buffer pointer.
|
|
*
|
|
* Sets a0 to the length of the written number.
|
|
*)
|
|
proc _print_i(number: Word, output_buffer: Word);
|
|
var
|
|
local_buffer: Word;
|
|
is_negative: Word;
|
|
current_character: Word;
|
|
result: Word;
|
|
begin
|
|
local_buffer := @result + 11;
|
|
|
|
if number >= 0 then
|
|
is_negative := 0
|
|
else
|
|
number = -number;
|
|
is_negative := 1
|
|
end;
|
|
|
|
.print_i_digit10;
|
|
current_character := number % 10;
|
|
_store_byte(current_character + '0', local_buffer);
|
|
|
|
number := number / 10;
|
|
local_buffer := local_buffer - 1;
|
|
|
|
if number <> 0 then
|
|
goto print_i_digit10
|
|
end;
|
|
if is_negative = 1 then
|
|
_store_byte('-', local_buffer);
|
|
local_buffer := local_buffer - 1
|
|
end;
|
|
result := @result + 11;
|
|
result := result - local_buffer;
|
|
_memcpy(output_buffer, local_buffer + 1, result);
|
|
|
|
return result
|
|
end;
|
|
|
|
(**
|
|
* Writes a number to the standard output.
|
|
*
|
|
* Parameters:
|
|
* number - Whole number.
|
|
*)
|
|
proc _write_i(number: Word);
|
|
var
|
|
local_buffer: Word;
|
|
length: Word;
|
|
begin
|
|
length := _print_i(number, @local_buffer);
|
|
_write_s(@local_buffer, length)
|
|
end;
|
|
|
|
(**
|
|
* Writes a character from a0 into the standard output.
|
|
*
|
|
* Parameters:
|
|
* character - Character to write.
|
|
*)
|
|
proc _write_c(character: Word);
|
|
begin
|
|
_write_s(@character, 1)
|
|
end;
|
|
|
|
(**
|
|
* Write null terminated string.
|
|
*
|
|
* Parameters:
|
|
* string - String.
|
|
*)
|
|
proc _write_z(string: Word);
|
|
var
|
|
next_byte: Word;
|
|
begin
|
|
(* Check for 0 character. *)
|
|
next_byte := _load_byte(string);
|
|
|
|
if next_byte <> 0 then
|
|
(* Print a character. *)
|
|
_write_c(next_byte);
|
|
|
|
(* Advance the input string by one byte. *)
|
|
_write_z(string + 1)
|
|
end
|
|
end;
|
|
|
|
(**
|
|
* Detects if a0 is an uppercase character. Sets a0 to 1 if so, otherwise to 0.
|
|
*)
|
|
proc _is_upper(character: Word);
|
|
var
|
|
lhs: Word;
|
|
rhs: Word;
|
|
begin
|
|
lhs := character >= 'A';
|
|
rhs := character <= 'Z';
|
|
|
|
return lhs & rhs
|
|
|
|
end;
|
|
|
|
(**
|
|
* Detects if a0 is an lowercase character. Sets a0 to 1 if so, otherwise to 0.
|
|
*)
|
|
proc _is_lower(character: Word);
|
|
var
|
|
lhs: Word;
|
|
rhs: Word;
|
|
begin
|
|
lhs := character >= 'a';
|
|
rhs := character <= 'z';
|
|
|
|
return lhs & rhs
|
|
end;
|
|
|
|
(**
|
|
* Detects if the passed character is a 7-bit alpha character or an underscore.
|
|
*
|
|
* Paramters:
|
|
* character - Tested character.
|
|
*
|
|
* Sets a0 to 1 if the character is an alpha character or underscore, sets it to 0 otherwise.
|
|
*)
|
|
proc _is_alpha(character: Word);
|
|
var
|
|
is_upper_result: Word;
|
|
is_lower_result: Word;
|
|
is_alpha_result: Word;
|
|
is_underscore: Word;
|
|
begin
|
|
is_upper_result := _is_upper(character);
|
|
is_lower_result := _is_lower(character);
|
|
is_underscore := character = '_';
|
|
|
|
is_alpha_result := is_lower_result or is_upper_result;
|
|
return is_alpha_result or is_underscore
|
|
end;
|
|
|
|
(**
|
|
* Detects whether the passed character is a digit (a value between 0 and 9).
|
|
*
|
|
* Parameters:
|
|
* character - Exemined value.
|
|
*
|
|
* Sets a0 to 1 if it is a digit, to 0 otherwise.
|
|
*)
|
|
proc _is_digit(character: Word);
|
|
var
|
|
lhs: Word;
|
|
rhs: Word;
|
|
begin
|
|
lhs := character >= '0';
|
|
rhs := character <= '9';
|
|
|
|
return lhs & rhs
|
|
end;
|
|
|
|
proc _is_alnum(character: Word);
|
|
var
|
|
lhs: Word;
|
|
rhs: Word;
|
|
begin
|
|
lhs := _is_alpha(character);
|
|
rhs := _is_digit(character);
|
|
|
|
return lhs or rhs
|
|
end;
|
|
|
|
(**
|
|
* Parameters:
|
|
* lhs - First pointer.
|
|
* rhs - Second pointer.
|
|
* count - The length to compare.
|
|
*
|
|
* Returns 0 if memory regions are equal.
|
|
*)
|
|
proc _memcmp(lhs: Word, rhs: Word, count: Word);
|
|
var
|
|
lhs_byte: Word;
|
|
rhs_byte: Word;
|
|
result: Word;
|
|
begin
|
|
result := 0;
|
|
|
|
.memcmp_loop;
|
|
if count <> 0 then
|
|
lhs_byte := _load_byte(lhs);
|
|
rhs_byte := _load_byte(rhs);
|
|
result := lhs_byte - rhs_byte;
|
|
|
|
lhs := lhs + 1;
|
|
rhs := rhs + 1;
|
|
count := count - 1;
|
|
|
|
if result = 0 then
|
|
goto memcmp_loop
|
|
end
|
|
end;
|
|
|
|
return result
|
|
end;
|
|
|
|
(**
|
|
* Copies memory.
|
|
*
|
|
* Parameters:
|
|
* destination - Destination.
|
|
* source - Source.
|
|
* count - Size.
|
|
*
|
|
* Returns the destination.
|
|
*)
|
|
proc _memcpy(destination: Word, source: Word, count: Word);
|
|
var
|
|
current_byte: Word;
|
|
begin
|
|
.memcpy_loop;
|
|
if count <> 0 then
|
|
current_byte := _load_byte(source);
|
|
_store_byte(current_byte, destination);
|
|
|
|
destination := destination + 1;
|
|
source := source + 1;
|
|
count := count - 1;
|
|
goto memcpy_loop
|
|
end;
|
|
|
|
return destination
|
|
end;
|
|
|
|
proc _node_get_kind(this: Word);
|
|
return this^
|
|
end;
|
|
|
|
proc _node_set_kind(this: Word, kind: Word);
|
|
begin
|
|
this^ := kind
|
|
end;
|
|
|
|
proc _integer_literal_node_size();
|
|
return 12
|
|
end;
|
|
|
|
proc _integer_literal_node_get_value(this: Word);
|
|
return _load_word(this + 4)
|
|
end;
|
|
|
|
proc _integer_literal_node_set_value(this: Word, value: Word);
|
|
begin
|
|
_store_word(value, this + 4)
|
|
end;
|
|
|
|
proc _integer_literal_node_get_length(this: Word);
|
|
return _load_word(this + 8)
|
|
end;
|
|
|
|
proc _integer_literal_node_set_length(this: Word, value: Word);
|
|
begin
|
|
_store_word(value, this + 8)
|
|
end;
|
|
|
|
proc _parse_integer_literal();
|
|
var
|
|
integer_token: Word;
|
|
integer_length: Word;
|
|
result: Word;
|
|
begin
|
|
result := memory_free_pointer;
|
|
memory_free_pointer := memory_free_pointer + 12;
|
|
|
|
integer_token := _lexer_global_get_start();
|
|
integer_length := _lexer_global_get_end() - integer_token;
|
|
_lexer_skip_token();
|
|
|
|
_node_set_kind(result, NodeKind.integer_literal);
|
|
_integer_literal_node_set_value(result, integer_token);
|
|
_integer_literal_node_set_length(result, integer_length);
|
|
|
|
return result
|
|
end;
|
|
|
|
proc _compile_integer_literal(integer_literal_node: Word);
|
|
var
|
|
integer_token: Word;
|
|
integer_length: Word;
|
|
token_kind: Word;
|
|
begin
|
|
_write_z("\tli t0, \0");
|
|
|
|
integer_token := _integer_literal_node_get_value(integer_literal_node);
|
|
integer_length := _integer_literal_node_get_length(integer_literal_node);
|
|
|
|
_write_s(integer_token, integer_length);
|
|
_write_c('\n')
|
|
end;
|
|
|
|
proc _character_literal_node_get_value(this: Word);
|
|
return _load_word(this + 4)
|
|
end;
|
|
|
|
proc _character_literal_node_set_value(this: Word, value: Word);
|
|
begin
|
|
_store_word(value, this + 4)
|
|
end;
|
|
|
|
proc _character_literal_node_get_length(this: Word);
|
|
return _load_word(this + 8)
|
|
end;
|
|
|
|
proc _character_literal_node_set_length(this: Word, value: Word);
|
|
begin
|
|
_store_word(value, this + 8)
|
|
end;
|
|
|
|
proc _parse_character_literal();
|
|
var
|
|
character: Word;
|
|
character_length: Word;
|
|
result: Word;
|
|
begin
|
|
result := memory_free_pointer;
|
|
memory_free_pointer := memory_free_pointer + 12;
|
|
|
|
character := _lexer_global_get_start();
|
|
character_length := _lexer_global_get_end() - character;
|
|
_lexer_skip_token();
|
|
|
|
_node_set_kind(result, NodeKind.character_literal);
|
|
_integer_literal_node_set_value(result, character);
|
|
_integer_literal_node_set_length(result, character_length);
|
|
|
|
return result
|
|
end;
|
|
|
|
proc _compile_character_literal(character_literal_node: Word);
|
|
var
|
|
character: Word;
|
|
character_length: Word;
|
|
begin
|
|
character := _character_literal_node_get_value(character_literal_node);
|
|
character_length := _character_literal_node_get_length(character_literal_node);;
|
|
|
|
_write_z("\tli t0, \0");
|
|
_write_s(character, character_length);
|
|
_write_c('\n');
|
|
end;
|
|
|
|
proc _compile_variable_expression();
|
|
var
|
|
name: Word;
|
|
lookup_result: Word;
|
|
name_token: Word;
|
|
begin
|
|
name := _lexer_global_get_start();
|
|
name_token := _lexer_global_get_end() - name;
|
|
lookup_result := _symbol_table_lookup(@symbol_table_global, name, name_token);
|
|
|
|
if lookup_result <> 0 then
|
|
_compile_enumeration_value(lookup_result)
|
|
else
|
|
_compile_designator();
|
|
_write_z("\tlw t0, (t0)\n\0")
|
|
end
|
|
end;
|
|
|
|
(**
|
|
* Compiled take address expression, starting with an "@" sign.
|
|
*)
|
|
proc _compile_address_expression();
|
|
begin
|
|
_lexer_skip_token();
|
|
_compile_designator()
|
|
end;
|
|
|
|
(**
|
|
* Compile unary negation, "-" sign.
|
|
*)
|
|
proc _compile_negate_expression();
|
|
begin
|
|
_lexer_skip_token();
|
|
_compile_term();
|
|
_write_z("\tneg t0, t0\n\0")
|
|
end;
|
|
|
|
(* Compile unary negation, "~" sign. *)
|
|
proc _compile_not_expression();
|
|
var
|
|
token_kind: Word;
|
|
begin
|
|
_lexer_read_token(@token_kind);
|
|
_lexer_skip_token();
|
|
_compile_term();
|
|
_write_z("\tnot t0, t0\n\0")
|
|
end;
|
|
|
|
proc _string_literal_node_size();
|
|
return 12
|
|
end;
|
|
|
|
proc _string_literal_node_get_value(this: Word);
|
|
return _load_word(this + 4)
|
|
end;
|
|
|
|
proc _string_literal_node_set_value(this: Word, value: Word);
|
|
begin
|
|
_store_word(value, this + 4)
|
|
end;
|
|
|
|
proc _string_literal_node_get_length(this: Word);
|
|
return _load_word(this + 8)
|
|
end;
|
|
|
|
proc _string_literal_node_set_length(this: Word, value: Word);
|
|
begin
|
|
_store_word(value, this + 8)
|
|
end;
|
|
|
|
proc _parse_string_literal();
|
|
var
|
|
length: Word;
|
|
token_start: Word;
|
|
result: Word;
|
|
begin
|
|
result := memory_free_pointer;
|
|
memory_free_pointer := memory_free_pointer + _string_literal_node_size();
|
|
|
|
token_start := _lexer_global_get_start();
|
|
length := _string_length(token_start);
|
|
_lexer_skip_token();
|
|
|
|
_string_literal_node_set_value(result, token_start);
|
|
_string_literal_node_set_length(result, length);
|
|
|
|
return result
|
|
end;
|
|
|
|
proc _compile_string_literal(string_literal_node: Word);
|
|
var
|
|
token_start: Word;
|
|
length: Word;
|
|
offset: Word;
|
|
begin
|
|
token_start := _string_literal_node_get_value(string_literal_node);
|
|
length := _string_literal_node_get_length(string_literal_node);
|
|
offset := _add_string(token_start);
|
|
|
|
_write_z("\tla t0, strings\n\0");
|
|
|
|
_write_z("\tli t1, \0");
|
|
_write_i(offset);
|
|
_write_c('\n');
|
|
|
|
_write_z("\tadd t0, t0, t1\n\0")
|
|
end;
|
|
|
|
proc _compile_term();
|
|
var
|
|
current_character: Word;
|
|
token_kind: Word;
|
|
parser_node: Word;
|
|
begin
|
|
_lexer_read_token(@token_kind);
|
|
|
|
if token_kind = LexerTokenKind.character then
|
|
parser_node := _parse_character_literal();
|
|
_compile_character_literal(parser_node)
|
|
elsif token_kind = LexerTokenKind.string then
|
|
parser_node := _parse_string_literal();
|
|
_compile_string_literal(parser_node)
|
|
elsif token_kind = LexerTokenKind.integer then
|
|
parser_node := _parse_integer_literal();
|
|
_compile_integer_literal(parser_node)
|
|
elsif token_kind = LexerTokenKind.at then
|
|
_compile_address_expression()
|
|
elsif token_kind = LexerTokenKind.minus then
|
|
_compile_negate_expression()
|
|
elsif token_kind = LexerTokenKind.not then
|
|
_compile_not_expression()
|
|
elsif token_kind = LexerTokenKind.identifier then
|
|
current_character := _lexer_global_get_start();
|
|
current_character := _load_byte(current_character);
|
|
|
|
(* This is a call if the statement starts with an underscore. *)
|
|
if current_character = '_' then
|
|
_compile_call();
|
|
_write_z("\tmv t0, a0\n\0")
|
|
else
|
|
_compile_variable_expression()
|
|
end
|
|
end
|
|
end;
|
|
|
|
proc _compile_binary_rhs();
|
|
begin
|
|
(* Save the value of the left expression on the stack. *)
|
|
_write_z("\tsw t0, 64(sp)\n\0");
|
|
_compile_term();
|
|
|
|
(* Load the left expression from the stack; *)
|
|
_write_z("\tlw t1, 64(sp)\n\0")
|
|
end;
|
|
|
|
proc _compile_expression();
|
|
var
|
|
token_kind: Word;
|
|
begin
|
|
_compile_term();
|
|
|
|
_lexer_read_token(@token_kind);
|
|
|
|
if token_kind = LexerTokenKind.plus then
|
|
_lexer_skip_token();
|
|
_compile_binary_rhs();
|
|
|
|
(* Execute the operation. *)
|
|
_write_z("\tadd t0, t0, t1\n\0")
|
|
elsif token_kind = LexerTokenKind.minus then
|
|
_lexer_skip_token();
|
|
_compile_binary_rhs();
|
|
|
|
(* Execute the operation. *)
|
|
_write_z("\tsub t0, t1, t0\n\0");
|
|
elsif token_kind = LexerTokenKind.multiplication then
|
|
_lexer_skip_token();
|
|
_compile_binary_rhs();
|
|
|
|
(* Execute the operation. *)
|
|
_write_z("\tmul t0, t0, t1\n\0")
|
|
elsif token_kind = LexerTokenKind.and then
|
|
_lexer_skip_token();
|
|
_compile_binary_rhs();
|
|
|
|
(* Execute the operation. *)
|
|
_write_z("\tand t0, t0, t1\n\0")
|
|
elsif token_kind = LexerTokenKind._or then
|
|
_lexer_skip_token();
|
|
_compile_binary_rhs();
|
|
|
|
(* Execute the operation. *)
|
|
_write_z("\tor t0, t0, t1\n\0")
|
|
elsif token_kind = LexerTokenKind._xor then
|
|
_lexer_skip_token();
|
|
_compile_binary_rhs();
|
|
|
|
(* Execute the operation. *)
|
|
_write_z("\txor t0, t0, t1\n\0")
|
|
elsif token_kind = LexerTokenKind.equals then
|
|
_lexer_skip_token();
|
|
_compile_binary_rhs();
|
|
|
|
(* Execute the operation. *)
|
|
_write_z("\txor t0, t0, t1\n\tseqz t0, t0\n\0")
|
|
elsif token_kind = LexerTokenKind.remainder then
|
|
_lexer_skip_token();
|
|
_compile_binary_rhs();
|
|
|
|
(* Execute the operation. *)
|
|
_write_z("\trem t0, t1, t0\n\0")
|
|
elsif token_kind = LexerTokenKind.division then
|
|
_lexer_skip_token();
|
|
_compile_binary_rhs();
|
|
|
|
(* Execute the operation. *)
|
|
_write_z("\tdiv t0, t1, t0\n\0")
|
|
elsif token_kind = LexerTokenKind.less_than then
|
|
_lexer_skip_token();
|
|
_compile_binary_rhs();
|
|
|
|
(* Execute the operation. *)
|
|
_write_z("\tslt t0, t1, t0\n\0")
|
|
elsif token_kind = LexerTokenKind.greater_than then
|
|
_lexer_skip_token();
|
|
_compile_binary_rhs();
|
|
|
|
(* Execute the operation. *)
|
|
_write_z("\tslt t0, t0, t1\n\0")
|
|
elsif token_kind = LexerTokenKind.less_equal then
|
|
_lexer_skip_token();
|
|
_compile_binary_rhs();
|
|
|
|
(* Execute the operation. *)
|
|
_write_z("\tslt t0, t0, t1\n\txori t0, t0, 1\n\0")
|
|
elsif token_kind = LexerTokenKind.not_equal then
|
|
_lexer_skip_token();
|
|
_compile_binary_rhs();
|
|
|
|
(* Execute the operation. *)
|
|
_write_z("\txor t0, t0, t1\n\tsnez t0, t0\n\0")
|
|
elsif token_kind = LexerTokenKind.greater_equal then
|
|
_lexer_skip_token();
|
|
_compile_binary_rhs();
|
|
|
|
(* Execute the operation. *)
|
|
_write_z("\tslt t0, t1, t0\n\txori t0, t0, 1\n\0")
|
|
end;
|
|
|
|
.compile_expression_end;
|
|
end;
|
|
|
|
proc _compile_call();
|
|
var
|
|
name_length: Word;
|
|
name: Word;
|
|
argument_count: Word;
|
|
stack_offset: Word;
|
|
token_kind: Word;
|
|
begin
|
|
_lexer_read_token(@token_kind);
|
|
name := _lexer_global_get_start();
|
|
name_length := _lexer_global_get_end() - name;
|
|
argument_count := 0;
|
|
|
|
(* Skip the identifier and left paren. *)
|
|
_lexer_skip_token();
|
|
_lexer_read_token(@token_kind);
|
|
_lexer_skip_token();
|
|
|
|
_lexer_read_token(@token_kind);
|
|
if token_kind = LexerTokenKind.right_paren then
|
|
goto compile_call_finalize
|
|
end;
|
|
.compile_call_loop;
|
|
_compile_expression();
|
|
|
|
(* Save the argument on the stack. *)
|
|
_write_z("\tsw t0, \0");
|
|
|
|
(* Calculate the stack offset: 116 - (4 * argument_counter) *)
|
|
stack_offset := argument_count * 4;
|
|
_write_i(116 - stack_offset);
|
|
|
|
_write_z("(sp)\n\0");
|
|
|
|
(* Add one to the argument counter. *)
|
|
argument_count := argument_count + 1;
|
|
|
|
_lexer_read_token(@token_kind);
|
|
|
|
if token_kind <> LexerTokenKind.comma then
|
|
goto compile_call_finalize
|
|
end;
|
|
_lexer_skip_token();
|
|
goto compile_call_loop;
|
|
|
|
.compile_call_finalize;
|
|
(* Load the argument from the stack. *)
|
|
if argument_count <> 0 then
|
|
(* Decrement the argument counter. *)
|
|
argument_count := argument_count - 1;
|
|
|
|
_write_z("\tlw a\0");
|
|
_write_i(argument_count);
|
|
|
|
_write_z(", \0");
|
|
|
|
(* Calculate the stack offset: 116 - (4 * argument_counter) *)
|
|
stack_offset := argument_count * 4;
|
|
_write_i(116 - stack_offset);
|
|
|
|
_write_z("(sp)\n\0");
|
|
|
|
goto compile_call_finalize
|
|
end;
|
|
|
|
.compile_call_end;
|
|
_write_z("\tcall \0");
|
|
_write_s(name, name_length);
|
|
_write_c('\n');
|
|
|
|
(* Skip the right paren. *)
|
|
_lexer_read_token(@token_kind);
|
|
_lexer_skip_token()
|
|
end;
|
|
|
|
proc _compile_goto();
|
|
var
|
|
next_token: Word;
|
|
next_length: Word;
|
|
token_kind: Word;
|
|
begin
|
|
_lexer_skip_token();
|
|
_lexer_read_token(@token_kind);
|
|
|
|
if token_kind = LexerTokenKind.dot then
|
|
_lexer_skip_token();
|
|
_lexer_read_token(@token_kind)
|
|
end;
|
|
next_token := _lexer_global_get_start();
|
|
next_length := _lexer_global_get_end() - next_token;
|
|
|
|
_write_z("\tj .\0");
|
|
|
|
_write_s(next_token, next_length);
|
|
_lexer_skip_token()
|
|
end;
|
|
|
|
proc _compile_local_designator(symbol: Word);
|
|
var
|
|
variable_offset: Word;
|
|
begin
|
|
_write_z("\taddi t0, sp, \0");
|
|
variable_offset := _parameter_info_get_offset(symbol);
|
|
_write_i(variable_offset);
|
|
_write_c('\n');
|
|
_lexer_skip_token()
|
|
end;
|
|
|
|
proc _compile_global_designator();
|
|
var
|
|
name: Word;
|
|
token_kind: Word;
|
|
token_length: Word;
|
|
begin
|
|
_write_z("\tla t0, \0");
|
|
|
|
_lexer_read_token(@token_kind);
|
|
name := _lexer_global_get_start();
|
|
token_length := _lexer_global_get_end() - name;
|
|
_write_s(name, token_length);
|
|
_lexer_skip_token();
|
|
|
|
_write_c('\n')
|
|
end;
|
|
|
|
proc _compile_enumeration_value(symbol: Word);
|
|
var
|
|
enumeration_type: Word;
|
|
members: Word;
|
|
members_length: Word;
|
|
token_type: Word;
|
|
value_name: Word;
|
|
name_length: Word;
|
|
member_name: Word;
|
|
member_length: Word;
|
|
counter: Word;
|
|
begin
|
|
enumeration_type := _type_info_get_type(symbol);
|
|
members := _enumeration_type_get_members(enumeration_type);
|
|
members_length := _enumeration_type_get_length(enumeration_type);
|
|
|
|
(* Skip enumeration type name and dot. Read the enumeration value. *)
|
|
_lexer_skip_token();
|
|
_lexer_read_token(@token_type);
|
|
_lexer_skip_token();
|
|
_lexer_read_token(@token_type);
|
|
|
|
value_name := _lexer_global_get_start();
|
|
name_length := _lexer_global_get_end() - value_name;
|
|
_lexer_skip_token();
|
|
counter := 1;
|
|
|
|
.compile_enumeration_value_members;
|
|
if members_length > 0 then
|
|
member_name := members^;
|
|
member_length := _load_word(members + 4);
|
|
|
|
if _lexer_compare_keyword(value_name, name_length, member_name, member_length) = 0 then
|
|
members_length := members_length - 1;
|
|
members := members + 8;
|
|
counter := counter + 1;
|
|
goto compile_enumeration_value_members
|
|
end;
|
|
_write_z("\tli t0, \0");
|
|
_write_i(counter);
|
|
_write_c('\n')
|
|
end
|
|
end;
|
|
|
|
proc _compile_designator();
|
|
var
|
|
name_token: Word;
|
|
lookup_result: Word;
|
|
token_kind: Word;
|
|
name: Word;
|
|
begin
|
|
_lexer_read_token(@token_kind);
|
|
name := _lexer_global_get_start();
|
|
name_token := _lexer_global_get_end() - name;
|
|
lookup_result := _symbol_table_lookup(@symbol_table_local, name, name_token);
|
|
|
|
if lookup_result <> 0 then
|
|
_compile_local_designator(lookup_result)
|
|
else
|
|
_compile_global_designator()
|
|
end;
|
|
_lexer_read_token(@token_kind);
|
|
if token_kind = LexerTokenKind.hat then
|
|
_lexer_skip_token();
|
|
_write_z("\tlw t0, (t0)\n\0")
|
|
end
|
|
end;
|
|
|
|
proc _compile_assignment();
|
|
var
|
|
token_kind: Word;
|
|
begin
|
|
_compile_designator();
|
|
|
|
(* Save the assignee address on the stack. *)
|
|
_write_z("\tsw t0, 60(sp)\n\0");
|
|
|
|
(* Skip the assignment sign (:=) with surrounding whitespaces. *)
|
|
_lexer_read_token(@token_kind);
|
|
_lexer_skip_token();
|
|
|
|
(* Compile the assignment. *)
|
|
_compile_expression();
|
|
|
|
_write_z("\tlw t1, 60(sp)\n\tsw t0, (t1)\n\0")
|
|
end;
|
|
|
|
proc _compile_return_statement();
|
|
var
|
|
token_kind: Word;
|
|
begin
|
|
(* Skip "return" keyword and whitespace after it. *)
|
|
_lexer_read_token(@token_kind);
|
|
_lexer_skip_token();
|
|
|
|
_compile_expression();
|
|
_write_z("\tmv a0, t0\n\0")
|
|
end;
|
|
|
|
(**
|
|
* Writes a label, .Ln, where n is a unique number.
|
|
*
|
|
* Parameters:
|
|
* counter - Label counter.
|
|
*)
|
|
proc _write_label(counter: Word);
|
|
begin
|
|
_write_z(".L\0");
|
|
_write_i(counter)
|
|
end;
|
|
|
|
proc _compile_condition(after_end_label: Word);
|
|
var
|
|
condition_label: Word;
|
|
token_kind: Word;
|
|
begin
|
|
(* Compile condition. *)
|
|
_compile_expression();
|
|
(* Skip " then" with newline. *)
|
|
_lexer_read_token(@token_kind);
|
|
_lexer_skip_token();
|
|
|
|
(* condition_label is the label in front of the next elsif condition or end. *)
|
|
condition_label := label_counter;
|
|
label_counter := label_counter + 1;
|
|
|
|
_write_z("\tbeqz t0, \0");
|
|
_write_label(condition_label);
|
|
_write_c('\n');
|
|
|
|
_compile_statement_list();
|
|
|
|
_write_z("\tj \0");
|
|
_write_label(after_end_label);
|
|
_write_c('\n');
|
|
|
|
_write_label(condition_label);
|
|
_write_z(":\n\0")
|
|
end;
|
|
|
|
proc _compile_if();
|
|
var
|
|
after_end_label: Word;
|
|
condition_label: Word;
|
|
token_kind: Word;
|
|
begin
|
|
(* Skip "if ". *)
|
|
_lexer_read_token(@token_kind);
|
|
_lexer_skip_token();
|
|
|
|
after_end_label := label_counter;
|
|
label_counter := label_counter + 1;
|
|
|
|
_compile_condition(after_end_label);
|
|
.compile_if_loop;
|
|
|
|
_lexer_read_token(@token_kind);
|
|
if token_kind = LexerTokenKind._else then
|
|
_lexer_skip_token();
|
|
_compile_statement_list()
|
|
elsif token_kind = LexerTokenKind._elsif then
|
|
_lexer_skip_token();
|
|
_compile_condition(after_end_label);
|
|
|
|
goto compile_if_loop
|
|
end;
|
|
_lexer_skip_token();
|
|
|
|
_write_label(after_end_label);
|
|
_write_z(":\n\0")
|
|
end;
|
|
|
|
proc _compile_label_declaration();
|
|
var
|
|
label_token: Word;
|
|
token_kind: Word;
|
|
name: Word;
|
|
begin
|
|
(* Skip the dot. *)
|
|
_lexer_skip_token();
|
|
_lexer_read_token(@token_kind);
|
|
name := _lexer_global_get_start();
|
|
label_token := _lexer_global_get_end() - name;
|
|
_write_c('.');
|
|
_write_s(name, label_token);
|
|
_write_z(":\n\0");
|
|
_lexer_skip_token()
|
|
end;
|
|
|
|
proc _compile_statement();
|
|
var
|
|
current_byte: Word;
|
|
token_kind: Word;
|
|
begin
|
|
_lexer_read_token(@token_kind);
|
|
|
|
if token_kind = LexerTokenKind._goto then
|
|
_compile_goto()
|
|
elsif token_kind = LexerTokenKind._if then
|
|
_compile_if()
|
|
elsif token_kind = LexerTokenKind._return then
|
|
_compile_return_statement()
|
|
elsif token_kind = LexerTokenKind.dot then
|
|
_compile_label_declaration()
|
|
elsif token_kind = LexerTokenKind.identifier then
|
|
current_byte := _lexer_global_get_start();
|
|
current_byte := _load_byte(current_byte);
|
|
|
|
(* This is a call if the statement starts with an underscore. *)
|
|
if current_byte = '_' then
|
|
_compile_call()
|
|
else
|
|
_compile_assignment()
|
|
end
|
|
end;
|
|
_write_c('\n')
|
|
end;
|
|
|
|
proc _compile_statement_list();
|
|
var
|
|
token_kind: Word;
|
|
begin
|
|
_skip_empty_lines();
|
|
_compile_statement();
|
|
_lexer_read_token(@token_kind);
|
|
|
|
if token_kind = LexerTokenKind.semicolon then
|
|
_lexer_skip_token();
|
|
_compile_statement_list()
|
|
end;
|
|
_skip_empty_lines()
|
|
end;
|
|
|
|
(**
|
|
* Writes a regster name to the standard output.
|
|
*
|
|
* Parameters:
|
|
* register_character - Register character.
|
|
* register_number - Register number.
|
|
*)
|
|
proc _write_register(register_character: Word, register_number: Word);
|
|
begin
|
|
_write_c(register_character);
|
|
_write_c(register_number + '0')
|
|
end;
|
|
|
|
proc _type_get_kind(this: Word);
|
|
return this^
|
|
end;
|
|
|
|
proc _type_set_kind(this: Word, value: Word);
|
|
begin
|
|
_store_word(value, this)
|
|
end;
|
|
|
|
proc _type_get_size(this: Word);
|
|
return _load_word(this + 4)
|
|
end;
|
|
|
|
proc _type_set_size(this: Word, value: Word);
|
|
begin
|
|
_store_word(value, this + 4)
|
|
end;
|
|
|
|
proc _enumeration_type_get_members(this: Word);
|
|
return _load_word(this + 8)
|
|
end;
|
|
|
|
proc _enumeration_type_set_members(this: Word, value: Word);
|
|
begin
|
|
_store_word(value, this + 8)
|
|
end;
|
|
|
|
proc _enumeration_type_get_length(this: Word);
|
|
return _load_word(this + 12)
|
|
end;
|
|
|
|
proc _enumeration_type_set_length(this: Word, value: Word);
|
|
begin
|
|
_store_word(value, this + 12)
|
|
end;
|
|
|
|
(**
|
|
* Reads and creates enumeration type representation.
|
|
*
|
|
* record
|
|
* type_kind: Word;
|
|
* size: Word;
|
|
* members: StringArray;
|
|
* length: Word
|
|
* end;
|
|
*
|
|
* Returns enumeration type description.
|
|
*)
|
|
proc _read_type_enumeration();
|
|
var
|
|
token_kind: Word;
|
|
enumeration_name: Word;
|
|
name_length: Word;
|
|
memory_start: Word;
|
|
member_count: Word;
|
|
result: Word;
|
|
begin
|
|
_lexer_skip_token();
|
|
memory_start := memory_free_pointer;
|
|
member_count := 0;
|
|
|
|
_lexer_read_token(@token_kind);
|
|
if token_kind = LexerTokenKind.right_paren then
|
|
goto read_type_enumeration_end
|
|
end;
|
|
.read_type_enumeration_loop;
|
|
member_count := member_count + 1;
|
|
|
|
enumeration_name := _lexer_global_get_start();
|
|
name_length := _lexer_global_get_end() - enumeration_name;
|
|
|
|
_store_word(enumeration_name, memory_free_pointer);
|
|
memory_free_pointer := memory_free_pointer + 4;
|
|
|
|
_store_word(name_length, memory_free_pointer);
|
|
memory_free_pointer := memory_free_pointer + 4;
|
|
|
|
(* Skip the identifier. *)
|
|
_lexer_skip_token();
|
|
|
|
_lexer_read_token(@token_kind);
|
|
if token_kind = LexerTokenKind.comma then
|
|
_lexer_skip_token();
|
|
_lexer_read_token(@token_kind);
|
|
goto read_type_enumeration_loop
|
|
end;
|
|
|
|
.read_type_enumeration_end;
|
|
_lexer_skip_token();
|
|
|
|
(* The resulting structure is 16 bytes long. *)
|
|
result := memory_free_pointer;
|
|
memory_free_pointer := memory_free_pointer + 16;
|
|
|
|
(* ENUMERATION_TYPE is 2. *)
|
|
_type_set_kind(result, 2);
|
|
_type_set_size(result, 4);
|
|
_enumeration_type_set_members(result, memory_start);
|
|
_enumeration_type_set_length(result, member_count);
|
|
|
|
return _type_info_create(result)
|
|
end;
|
|
|
|
proc _read_type_expression();
|
|
var
|
|
token_kind: Word;
|
|
type_name: Word;
|
|
name_length: Word;
|
|
result: Word;
|
|
begin
|
|
result := 0;
|
|
_lexer_read_token(@token_kind);
|
|
|
|
if token_kind = LexerTokenKind.identifier then
|
|
(* Named type. *)
|
|
type_name := _lexer_global_get_start();
|
|
name_length := _lexer_global_get_end() - type_name;
|
|
result := _symbol_table_lookup(@symbol_table_global, type_name, name_length);
|
|
result := _type_info_get_type(result);
|
|
|
|
_lexer_skip_token()
|
|
elsif token_kind = LexerTokenKind.left_paren then
|
|
result := _read_type_enumeration()
|
|
end;
|
|
|
|
return result
|
|
end;
|
|
|
|
proc _type_info_get_type(this: Word);
|
|
return _load_word(this + 4)
|
|
end;
|
|
|
|
(**
|
|
* Parameters:
|
|
* parameter_index - Parameter index.
|
|
*)
|
|
proc _parameter_info_create(parameter_index: Word);
|
|
var
|
|
offset: Word;
|
|
current_word: Word;
|
|
result: Word;
|
|
begin
|
|
result := memory_free_pointer;
|
|
current_word := result;
|
|
(* 2 is INFO_PARAMETER *)
|
|
_store_word(2, current_word);
|
|
|
|
current_word := current_word + 4;
|
|
|
|
(* Calculate the stack offset: 88 - (4 * parameter_counter) *)
|
|
offset := parameter_index * 4;
|
|
_store_word(88 - offset, current_word);
|
|
|
|
memory_free_pointer := current_word + 4;
|
|
|
|
return result
|
|
end;
|
|
|
|
proc _parameter_info_get_offset(this: Word);
|
|
begin
|
|
this := this + 4;
|
|
return this^
|
|
end;
|
|
|
|
proc _type_info_create(type_representation: Word);
|
|
var
|
|
result: Word;
|
|
current_word: Word;
|
|
begin
|
|
result := memory_free_pointer;
|
|
current_word := result;
|
|
(* 1 is INFO_TYPE *)
|
|
_store_word(1, current_word);
|
|
|
|
current_word := current_word + 4;
|
|
_store_word(type_representation, current_word);
|
|
|
|
memory_free_pointer := current_word + 4;
|
|
|
|
return result
|
|
end;
|
|
|
|
(**
|
|
* Parameters:
|
|
* temporary_index - Parameter index.
|
|
*)
|
|
proc _temporary_info_create(temporary_index: Word);
|
|
var
|
|
offset: Word;
|
|
current_word: Word;
|
|
result: Word;
|
|
begin
|
|
result := memory_free_pointer;
|
|
current_word := result;
|
|
(* 3 is INFO_TEMPORARY *)
|
|
_store_word(3, current_word);
|
|
|
|
current_word := current_word + 4;
|
|
|
|
(* Calculate the stack offset: 4 * variable_counter. *)
|
|
_store_word(temporary_index * 4, current_word);
|
|
|
|
memory_free_pointer := current_word + 4;
|
|
|
|
return result
|
|
end;
|
|
|
|
proc _temporary_info_get_offset(this: Word);
|
|
begin
|
|
this := this + 4;
|
|
return this^
|
|
end;
|
|
|
|
(**
|
|
* Parameters:
|
|
* parameter_index - Parameter index.
|
|
*)
|
|
proc _read_procedure_parameter(parameter_index: Word);
|
|
var
|
|
name_length: Word;
|
|
info: Word;
|
|
name_position: Word;
|
|
token_kind: Word;
|
|
begin
|
|
(* Read the parameter name. *)
|
|
_lexer_read_token(@token_kind);
|
|
name_position := _lexer_global_get_start();
|
|
name_length := _lexer_global_get_end() - name_position;
|
|
_lexer_skip_token();
|
|
|
|
(* Skip colon and space in front of the type expression. *)
|
|
_lexer_read_token(@token_kind);
|
|
_lexer_skip_token();
|
|
|
|
_read_type_expression();
|
|
|
|
_write_z("\tsw a\0");
|
|
_write_i(parameter_index);
|
|
_write_z(", \0");
|
|
|
|
info := _parameter_info_create(parameter_index);
|
|
_symbol_table_enter(@symbol_table_local, name_position, name_length, info);
|
|
|
|
info := _parameter_info_get_offset(info);
|
|
_write_i(info);
|
|
|
|
_write_z("(sp)\n\0")
|
|
end;
|
|
|
|
proc _read_procedure_parameters();
|
|
var
|
|
parameter_counter: Word;
|
|
token_kind: Word;
|
|
begin
|
|
(* Skip open paren. *)
|
|
_lexer_read_token(@token_kind);
|
|
_lexer_skip_token();
|
|
parameter_counter := 0;
|
|
|
|
.compile_procedure_prologue_skip;
|
|
_lexer_read_token(@token_kind);
|
|
|
|
if token_kind <> LexerTokenKind.right_paren then
|
|
_read_procedure_parameter(parameter_counter);
|
|
parameter_counter := parameter_counter + 1;
|
|
_lexer_read_token(@token_kind);
|
|
|
|
if token_kind = LexerTokenKind.comma then
|
|
_lexer_skip_token();
|
|
goto compile_procedure_prologue_skip
|
|
end
|
|
end;
|
|
(* Skip close paren. *)
|
|
_lexer_skip_token()
|
|
end;
|
|
|
|
(**
|
|
* Parameters:
|
|
* variable_index - Variable index.
|
|
*)
|
|
proc _read_procedure_temporary(variable_index: Word);
|
|
var
|
|
name_length: Word;
|
|
info: Word;
|
|
name_position: Word;
|
|
token_kind: Word;
|
|
begin
|
|
_lexer_read_token(@token_kind);
|
|
name_position := _lexer_global_get_start();
|
|
name_length := _lexer_global_get_end() - name_position;
|
|
_lexer_skip_token();
|
|
|
|
(* Read and skip variable name, colon and the space *)
|
|
_lexer_read_token(@token_kind);
|
|
_lexer_skip_token();
|
|
|
|
_read_type_expression();
|
|
|
|
info := _temporary_info_create(variable_index);
|
|
_symbol_table_enter(@symbol_table_local, name_position, name_length, info);
|
|
|
|
(* Skip semicolon and newline after the variable declaration *)
|
|
_lexer_read_token(@token_kind);
|
|
_lexer_skip_token()
|
|
end;
|
|
|
|
proc _read_procedure_temporaries();
|
|
var
|
|
temporary_counter: Word;
|
|
token_kind: Word;
|
|
begin
|
|
_lexer_read_token(@token_kind);
|
|
|
|
if token_kind = LexerTokenKind._var then
|
|
_lexer_skip_token();
|
|
temporary_counter := 0;
|
|
|
|
.read_local_variables_loop;
|
|
_lexer_read_token(@token_kind);
|
|
|
|
if token_kind = LexerTokenKind.identifier then
|
|
_read_procedure_temporary(temporary_counter);
|
|
|
|
temporary_counter := temporary_counter + 1;
|
|
goto read_local_variables_loop
|
|
end
|
|
end
|
|
end;
|
|
|
|
proc _compile_procedure();
|
|
var
|
|
name_pointer: Word;
|
|
name_length: Word;
|
|
token_kind: Word;
|
|
begin
|
|
(* Skip "proc ". *)
|
|
_lexer_read_token(@token_kind);
|
|
_lexer_skip_token();
|
|
|
|
(* Clear local symbol table. *)
|
|
_store_word(0, @symbol_table_local);
|
|
|
|
_lexer_read_token(@token_kind);
|
|
name_pointer := _lexer_global_get_start();
|
|
name_length := _lexer_global_get_end() - name_pointer;
|
|
|
|
(* Write .type _procedure_name, @function. *)
|
|
_write_z(".type \0");
|
|
|
|
_write_s(name_pointer, name_length);
|
|
_write_z(", @function\n\0");
|
|
|
|
(* Write procedure label, _procedure_name: *)
|
|
_write_s(name_pointer, name_length);
|
|
_write_z(":\n\0");
|
|
|
|
(* Skip procedure name. *)
|
|
_lexer_skip_token();
|
|
_write_z("\taddi sp, sp, -128\n\tsw ra, 124(sp)\n\tsw s0, 120(sp)\n\taddi s0, sp, 128\n\0");
|
|
_read_procedure_parameters();
|
|
|
|
(* Skip semicolon and newline. *)
|
|
_lexer_read_token(@token_kind);
|
|
_lexer_skip_token();
|
|
_read_procedure_temporaries();
|
|
|
|
(* Skip semicolon, "begin" and newline. *)
|
|
_lexer_read_token(@token_kind);
|
|
if token_kind = LexerTokenKind._begin then
|
|
_lexer_skip_token();
|
|
_compile_statement_list()
|
|
elsif token_kind = LexerTokenKind._return then
|
|
_compile_return_statement()
|
|
end;
|
|
|
|
(* Write the epilogue. *)
|
|
_write_z("\tlw ra, 124(sp)\n\tlw s0, 120(sp)\n\taddi sp, sp, 128\n\tret\n\0");
|
|
|
|
(* Skip the "end" keyword, semicolon and newline. *)
|
|
_lexer_read_token(@token_kind);
|
|
_lexer_skip_token();
|
|
_lexer_read_token(@token_kind);
|
|
_lexer_skip_token()
|
|
end;
|
|
|
|
(**
|
|
* Skips comments.
|
|
*)
|
|
proc _skip_empty_lines();
|
|
var
|
|
token_kind: Word;
|
|
begin
|
|
.skip_empty_lines_rerun;
|
|
|
|
_lexer_read_token(@token_kind);
|
|
|
|
if token_kind = LexerTokenKind.comment then
|
|
_lexer_skip_token();
|
|
goto skip_empty_lines_rerun
|
|
end
|
|
end;
|
|
|
|
(**
|
|
* Compile global variable initializer.
|
|
*)
|
|
proc _compile_global_initializer();
|
|
var
|
|
current_byte: Word;
|
|
length: Word;
|
|
token_kind: Word;
|
|
token_start: Word;
|
|
begin
|
|
_lexer_read_token(@token_kind);
|
|
token_start := _lexer_global_get_start();
|
|
current_byte := _load_byte(token_start);
|
|
|
|
if token_kind = LexerTokenKind.string then
|
|
_write_z("\n\t.word strings + \0");
|
|
length := _string_length(token_start);
|
|
|
|
_add_string(token_start);
|
|
_write_i();
|
|
|
|
(* Skip the quoted string. *)
|
|
_lexer_skip_token();
|
|
|
|
goto compile_global_initializer_end
|
|
elsif current_byte = 'S' then
|
|
(* Skip "S(". *)
|
|
_lexer_skip_token();
|
|
_lexer_read_token(@token_kind);
|
|
_lexer_skip_token();
|
|
_lexer_read_token(@token_kind);
|
|
|
|
if token_kind = LexerTokenKind.right_paren then
|
|
goto compile_global_initializer_closing
|
|
end;
|
|
goto compile_global_initializer_loop
|
|
elsif token_kind = LexerTokenKind.at then
|
|
(* Skip @. *)
|
|
_lexer_skip_token();
|
|
_write_z("\n\t.word \0");
|
|
_lexer_read_token(@token_kind);
|
|
token_start := _lexer_global_get_start();
|
|
_write_s(token_start, _lexer_global_get_end() - token_start);
|
|
_lexer_skip_token();
|
|
|
|
goto compile_global_initializer_end
|
|
elsif token_kind = LexerTokenKind.integer then
|
|
_write_z("\n\t.word \0");
|
|
_write_s(token_start, _lexer_global_get_end() - token_start);
|
|
_lexer_skip_token();
|
|
|
|
goto compile_global_initializer_end
|
|
end;
|
|
|
|
.compile_global_initializer_loop;
|
|
_compile_global_initializer();
|
|
|
|
_lexer_read_token(@token_kind);
|
|
if token_kind <> LexerTokenKind.right_paren then
|
|
(* Skip comma and whitespace after it. *)
|
|
_lexer_skip_token();
|
|
|
|
goto compile_global_initializer_loop
|
|
end;
|
|
|
|
.compile_global_initializer_closing;
|
|
(* Skip ")" *)
|
|
_lexer_skip_token();
|
|
|
|
.compile_global_initializer_end
|
|
end;
|
|
|
|
proc _compile_constant_declaration();
|
|
var
|
|
name: Word;
|
|
name_length: Word;
|
|
token_kind: Word;
|
|
begin
|
|
name := _lexer_global_get_start();
|
|
name_length := _lexer_global_get_end() - name;
|
|
|
|
_write_z(".type \0");
|
|
_write_s(name, name_length);
|
|
_write_z(", @object\n\0");
|
|
|
|
_write_s(name, name_length);
|
|
_write_c(':');
|
|
|
|
(* Skip the constant name with assignment sign and surrounding whitespaces. *)
|
|
_lexer_skip_token();
|
|
_lexer_read_token(@token_kind);
|
|
_lexer_skip_token();
|
|
_compile_global_initializer();
|
|
|
|
(* Skip semicolon and newline. *)
|
|
_lexer_read_token(@token_kind);
|
|
_lexer_skip_token();
|
|
_write_c('\n')
|
|
end;
|
|
|
|
proc _compile_type_declaration();
|
|
var
|
|
token_kind: Word;
|
|
type_name: Word;
|
|
name_length: Word;
|
|
type_info: Word;
|
|
begin
|
|
type_name := _lexer_global_get_start();
|
|
name_length := _lexer_global_get_end() - type_name;
|
|
|
|
_lexer_skip_token();
|
|
_lexer_read_token(@token_kind);
|
|
_lexer_skip_token();
|
|
type_info := _read_type_expression();
|
|
|
|
_symbol_table_enter(@symbol_table_global, type_name, name_length, type_info);
|
|
|
|
_lexer_read_token(@token_kind);
|
|
_lexer_skip_token()
|
|
end;
|
|
|
|
proc _compile_type_part();
|
|
var
|
|
token_kind: Word;
|
|
begin
|
|
_skip_empty_lines();
|
|
_lexer_read_token(@token_kind);
|
|
|
|
if token_kind <> LexerTokenKind._type then
|
|
goto compile_type_part_end
|
|
end;
|
|
_lexer_skip_token();
|
|
|
|
.compile_type_part_loop;
|
|
_skip_empty_lines();
|
|
|
|
_lexer_read_token(@token_kind);
|
|
if token_kind = LexerTokenKind.identifier then
|
|
_compile_type_declaration();
|
|
goto compile_type_part_loop
|
|
end;
|
|
|
|
.compile_type_part_end
|
|
end;
|
|
|
|
proc _compile_const_part();
|
|
var
|
|
token_kind: Word;
|
|
begin
|
|
_skip_empty_lines();
|
|
_lexer_read_token(@token_kind);
|
|
|
|
if token_kind <> LexerTokenKind._const then
|
|
goto compile_const_part_end
|
|
end;
|
|
(* Skip "const" with the newline after it. *)
|
|
_lexer_skip_token();
|
|
_write_z(".section .rodata # Compiled from const section.\n\n\0");
|
|
|
|
.compile_const_part_loop;
|
|
_skip_empty_lines();
|
|
|
|
(* If the character at the line beginning is not indentation,
|
|
it is probably the next code section. *)
|
|
_lexer_read_token(@token_kind);
|
|
if token_kind = LexerTokenKind.identifier then
|
|
_compile_constant_declaration();
|
|
goto compile_const_part_loop
|
|
end;
|
|
|
|
.compile_const_part_end
|
|
end;
|
|
|
|
proc _compile_variable_declaration();
|
|
var
|
|
name: Word;
|
|
name_length: Word;
|
|
token_kind: Word;
|
|
begin
|
|
_lexer_read_token(@token_kind);
|
|
name := _lexer_global_get_start();
|
|
name_length := _lexer_global_get_end() - name;
|
|
|
|
_write_z(".type \0");
|
|
_write_s(name, name_length);
|
|
_write_z(", @object\n\0");
|
|
|
|
_write_s(name, name_length);
|
|
_write_c(':');
|
|
|
|
(* Skip the variable name and colon with space before the type. *)
|
|
_lexer_skip_token();
|
|
_lexer_read_token(@token_kind);
|
|
_lexer_skip_token();
|
|
_read_type_expression();
|
|
|
|
_lexer_read_token(@token_kind);
|
|
|
|
if token_kind <> LexerTokenKind.assignment then
|
|
(* Else we assume this is a zeroed 81920 bytes big array. *)
|
|
_write_z(" .zero 81920\0")
|
|
else
|
|
(* Skip the assignment sign with surrounding whitespaces. *)
|
|
_lexer_skip_token();
|
|
_compile_global_initializer();
|
|
_lexer_read_token(@token_kind)
|
|
end;
|
|
|
|
(* Skip semicolon and newline. *)
|
|
_lexer_read_token(@token_kind);
|
|
_lexer_skip_token();
|
|
_write_c('\n')
|
|
end;
|
|
|
|
proc _compile_var_part();
|
|
var
|
|
token_kind: Word;
|
|
begin
|
|
_lexer_read_token(@token_kind);
|
|
|
|
if token_kind <> LexerTokenKind._var then
|
|
goto compile_var_part_end
|
|
end;
|
|
(* Skip "var" and newline. *)
|
|
_lexer_skip_token();
|
|
_write_z(".section .data\n\0");
|
|
|
|
.compile_var_part_loop;
|
|
_skip_empty_lines();
|
|
_lexer_read_token(@token_kind);
|
|
|
|
if token_kind = LexerTokenKind.identifier then
|
|
_compile_variable_declaration();
|
|
goto compile_var_part_loop
|
|
end;
|
|
|
|
.compile_var_part_end
|
|
end;
|
|
|
|
(**
|
|
* Process the source code and print the generated code.
|
|
*)
|
|
proc _compile_module();
|
|
var
|
|
token_kind: Word;
|
|
begin
|
|
_compile_type_part();
|
|
_compile_const_part();
|
|
_skip_empty_lines();
|
|
_compile_var_part();
|
|
|
|
_write_z(".section .text\n\n\0");
|
|
_write_z(".type _syscall, @function\n_syscall:\n\tmv a7, a6\n\tecall\n\tret\n\n\0");
|
|
_write_z(".type _load_byte, @function\n_load_byte:\n\tlb a0, (a0)\nret\n\n\0");
|
|
_write_z(".type _load_word, @function\n_load_word:\n\tlw a0, (a0)\nret\n\n\0");
|
|
_write_z(".type _store_byte, @function\n_store_byte:\n\tsb a0, (a1)\nret\n\n\0");
|
|
_write_z(".type _store_word, @function\n_store_word:\n\tsw a0, (a1)\nret\n\n\0");
|
|
|
|
.compile_module_loop;
|
|
_skip_empty_lines();
|
|
_lexer_read_token(@token_kind);
|
|
|
|
if token_kind = LexerTokenKind._proc then
|
|
_compile_procedure();
|
|
goto compile_module_loop
|
|
end
|
|
end;
|
|
|
|
proc _compile();
|
|
var
|
|
compiler_strings_copy: Word;
|
|
compiler_strings_end: Word;
|
|
current_byte: Word;
|
|
begin
|
|
_write_z(".globl _start\n\n\0");
|
|
_compile_module();
|
|
|
|
_write_z(".section .rodata\n.type strings, @object\nstrings: .ascii \0");
|
|
_write_c('"');
|
|
|
|
compiler_strings_copy := @compiler_strings;
|
|
compiler_strings_end := compiler_strings_position;
|
|
|
|
.compile_loop;
|
|
if compiler_strings_copy < compiler_strings_end then
|
|
current_byte := _load_byte(compiler_strings_copy);
|
|
compiler_strings_copy := compiler_strings_copy + 1;
|
|
_write_c(current_byte);
|
|
|
|
goto compile_loop
|
|
end;
|
|
_write_c('"');
|
|
_write_c('\n')
|
|
end;
|
|
|
|
(**
|
|
* Terminates the program. a0 contains the return code.
|
|
*
|
|
* Parameters:
|
|
* a0 - Status code.
|
|
*)
|
|
proc _exit();
|
|
begin
|
|
_syscall(0, 0, 0, 0, 0, 0, 93)
|
|
end;
|
|
|
|
(**
|
|
* Looks for a symbol in the given symbol table.
|
|
*
|
|
* Parameters:
|
|
* symbol_table - Symbol table.
|
|
* symbol_name - Symbol name pointer.
|
|
* name_length - Symbol name length.
|
|
*
|
|
* Returns the symbol pointer or 0 in a0.
|
|
*)
|
|
proc _symbol_table_lookup(symbol_table: Word, symbol_name: Word, name_length: Word);
|
|
var
|
|
result: Word;
|
|
symbol_table_length: Word;
|
|
current_name: Word;
|
|
current_length: Word;
|
|
begin
|
|
result := 0;
|
|
|
|
(* The first word in the symbol table is its length, get it. *)
|
|
symbol_table_length := symbol_table^;
|
|
|
|
(* Go to the first symbol position. *)
|
|
symbol_table := symbol_table + 4;
|
|
|
|
.symbol_table_lookup_loop;
|
|
if symbol_table_length = 0 then
|
|
goto symbol_table_lookup_end
|
|
end;
|
|
|
|
(* Symbol name pointer and length. *)
|
|
current_name := symbol_table^;
|
|
current_length := _load_word(symbol_table + 4);
|
|
|
|
(* If lengths don't match, exit and return nil. *)
|
|
if name_length <> current_length then
|
|
goto symbol_table_lookup_repeat
|
|
end;
|
|
(* If names don't match, exit and return nil. *)
|
|
if _memcmp(symbol_name, current_name, name_length) <> 0 then
|
|
goto symbol_table_lookup_repeat
|
|
end;
|
|
(* Otherwise, the symbol is found. *)
|
|
result := _load_word(symbol_table + 8);
|
|
goto symbol_table_lookup_end;
|
|
|
|
.symbol_table_lookup_repeat;
|
|
symbol_table := symbol_table + 12;
|
|
symbol_table_length := symbol_table_length - 1;
|
|
goto symbol_table_lookup_loop;
|
|
|
|
.symbol_table_lookup_end;
|
|
return result
|
|
end;
|
|
|
|
(**
|
|
* Inserts a symbol into the table.
|
|
*
|
|
* Parameters:
|
|
* symbol_table - Symbol table.
|
|
* symbol_name - Symbol name pointer.
|
|
* name_length - Symbol name length.
|
|
* symbol - Symbol pointer.
|
|
*)
|
|
proc _symbol_table_enter(symbol_table: Word, symbol_name: Word, name_length: Word, symbol: Word);
|
|
var
|
|
table_length: Word;
|
|
symbol_pointer: Word;
|
|
begin
|
|
(* The first word in the symbol table is its length, get it. *)
|
|
table_length := symbol_table^;
|
|
|
|
(* Calculate the offset for the new symbol. *)
|
|
symbol_pointer := table_length * 12;
|
|
symbol_pointer := symbol_pointer + 4;
|
|
symbol_pointer := symbol_table + symbol_pointer;
|
|
|
|
_store_word(symbol_name, symbol_pointer);
|
|
symbol_pointer := symbol_pointer + 4;
|
|
_store_word(name_length, symbol_pointer);
|
|
symbol_pointer := symbol_pointer + 4;
|
|
_store_word(symbol, symbol_pointer);
|
|
|
|
(* Increment the symbol table length. *)
|
|
table_length := table_length + 1;
|
|
_store_word(table_length, symbol_table)
|
|
end;
|
|
|
|
proc _symbol_table_build();
|
|
begin
|
|
(* Set the table length to 0. *)
|
|
_store_word(0, @symbol_table_global);
|
|
|
|
(* Enter built-in symbols. *)
|
|
_symbol_table_enter(@symbol_table_global, symbol_builtin_name_int, 3, @symbol_type_info_int);
|
|
_symbol_table_enter(@symbol_table_global, symbol_builtin_name_word, 4, @symbol_type_info_word);
|
|
_symbol_table_enter(@symbol_table_global, symbol_builtin_name_pointer, 7, @symbol_type_info_pointer);
|
|
_symbol_table_enter(@symbol_table_global, symbol_builtin_name_char, 4, @symbol_type_info_char);
|
|
_symbol_table_enter(@symbol_table_global, symbol_builtin_name_array, 5, @symbol_type_info_array)
|
|
end;
|
|
|
|
(**
|
|
* Assigns some value to at array index.
|
|
*
|
|
* Parameters:
|
|
* array - Array pointer.
|
|
* index - Index (word offset into the array).
|
|
* data - Data to assign.
|
|
*)
|
|
proc _assign_at(array: Word, index: Word, data: Word);
|
|
var
|
|
target: Word;
|
|
begin
|
|
target := index - 1;
|
|
target := target * 4;
|
|
target := array + target;
|
|
|
|
_store_word(data, target)
|
|
end;
|
|
|
|
proc _get_at(array: Word, index: Word);
|
|
var
|
|
target: Word;
|
|
begin
|
|
target := index - 1;
|
|
target := target * 4;
|
|
target := array + target;
|
|
|
|
return target^
|
|
end;
|
|
|
|
(**
|
|
* Initializes the array with character classes.
|
|
*)
|
|
proc _lexer_classifications();
|
|
var
|
|
code: Word;
|
|
begin
|
|
_assign_at(@classification, 1, LexerClass.eof);
|
|
_assign_at(@classification, 2, LexerClass.invalid);
|
|
_assign_at(@classification, 3, LexerClass.invalid);
|
|
_assign_at(@classification, 4, LexerClass.invalid);
|
|
_assign_at(@classification, 5, LexerClass.invalid);
|
|
_assign_at(@classification, 6, LexerClass.invalid);
|
|
_assign_at(@classification, 7, LexerClass.invalid);
|
|
_assign_at(@classification, 8, LexerClass.invalid);
|
|
_assign_at(@classification, 9, LexerClass.invalid);
|
|
_assign_at(@classification, 10, LexerClass.space);
|
|
_assign_at(@classification, 11, LexerClass.space);
|
|
_assign_at(@classification, 12, LexerClass.invalid);
|
|
_assign_at(@classification, 13, LexerClass.invalid);
|
|
_assign_at(@classification, 14, LexerClass.space);
|
|
_assign_at(@classification, 15, LexerClass.invalid);
|
|
_assign_at(@classification, 16, LexerClass.invalid);
|
|
_assign_at(@classification, 17, LexerClass.invalid);
|
|
_assign_at(@classification, 18, LexerClass.invalid);
|
|
_assign_at(@classification, 19, LexerClass.invalid);
|
|
_assign_at(@classification, 20, LexerClass.invalid);
|
|
_assign_at(@classification, 21, LexerClass.invalid);
|
|
_assign_at(@classification, 22, LexerClass.invalid);
|
|
_assign_at(@classification, 23, LexerClass.invalid);
|
|
_assign_at(@classification, 24, LexerClass.invalid);
|
|
_assign_at(@classification, 25, LexerClass.invalid);
|
|
_assign_at(@classification, 26, LexerClass.invalid);
|
|
_assign_at(@classification, 27, LexerClass.invalid);
|
|
_assign_at(@classification, 28, LexerClass.invalid);
|
|
_assign_at(@classification, 29, LexerClass.invalid);
|
|
_assign_at(@classification, 30, LexerClass.invalid);
|
|
_assign_at(@classification, 31, LexerClass.invalid);
|
|
_assign_at(@classification, 32, LexerClass.invalid);
|
|
_assign_at(@classification, 33, LexerClass.space);
|
|
_assign_at(@classification, 34, LexerClass.single);
|
|
_assign_at(@classification, 35, LexerClass.double_quote);
|
|
_assign_at(@classification, 36, LexerClass.other);
|
|
_assign_at(@classification, 37, LexerClass.other);
|
|
_assign_at(@classification, 38, LexerClass.single);
|
|
_assign_at(@classification, 39, LexerClass.single);
|
|
_assign_at(@classification, 40, LexerClass.single_quote);
|
|
_assign_at(@classification, 41, LexerClass.left_paren);
|
|
_assign_at(@classification, 42, LexerClass.right_paren);
|
|
_assign_at(@classification, 43, LexerClass.asterisk);
|
|
_assign_at(@classification, 44, LexerClass.single);
|
|
_assign_at(@classification, 45, LexerClass.single);
|
|
_assign_at(@classification, 46, LexerClass.minus);
|
|
_assign_at(@classification, 47, LexerClass.dot);
|
|
_assign_at(@classification, 48, LexerClass.single);
|
|
_assign_at(@classification, 49, LexerClass.zero);
|
|
_assign_at(@classification, 50, LexerClass.digit);
|
|
_assign_at(@classification, 51, LexerClass.digit);
|
|
_assign_at(@classification, 52, LexerClass.digit);
|
|
_assign_at(@classification, 53, LexerClass.digit);
|
|
_assign_at(@classification, 54, LexerClass.digit);
|
|
_assign_at(@classification, 55, LexerClass.digit);
|
|
_assign_at(@classification, 56, LexerClass.digit);
|
|
_assign_at(@classification, 57, LexerClass.digit);
|
|
_assign_at(@classification, 58, LexerClass.digit);
|
|
_assign_at(@classification, 59, LexerClass.colon);
|
|
_assign_at(@classification, 60, LexerClass.single);
|
|
_assign_at(@classification, 61, LexerClass.less);
|
|
_assign_at(@classification, 62, LexerClass.equals);
|
|
_assign_at(@classification, 63, LexerClass.greater);
|
|
_assign_at(@classification, 64, LexerClass.other);
|
|
_assign_at(@classification, 65, LexerClass.single);
|
|
_assign_at(@classification, 66, LexerClass.alpha);
|
|
_assign_at(@classification, 67, LexerClass.alpha);
|
|
_assign_at(@classification, 68, LexerClass.alpha);
|
|
_assign_at(@classification, 69, LexerClass.alpha);
|
|
_assign_at(@classification, 70, LexerClass.alpha);
|
|
_assign_at(@classification, 71, LexerClass.alpha);
|
|
_assign_at(@classification, 72, LexerClass.alpha);
|
|
_assign_at(@classification, 73, LexerClass.alpha);
|
|
_assign_at(@classification, 74, LexerClass.alpha);
|
|
_assign_at(@classification, 75, LexerClass.alpha);
|
|
_assign_at(@classification, 76, LexerClass.alpha);
|
|
_assign_at(@classification, 77, LexerClass.alpha);
|
|
_assign_at(@classification, 78, LexerClass.alpha);
|
|
_assign_at(@classification, 79, LexerClass.alpha);
|
|
_assign_at(@classification, 80, LexerClass.alpha);
|
|
_assign_at(@classification, 81, LexerClass.alpha);
|
|
_assign_at(@classification, 82, LexerClass.alpha);
|
|
_assign_at(@classification, 83, LexerClass.alpha);
|
|
_assign_at(@classification, 84, LexerClass.alpha);
|
|
_assign_at(@classification, 85, LexerClass.alpha);
|
|
_assign_at(@classification, 86, LexerClass.alpha);
|
|
_assign_at(@classification, 87, LexerClass.alpha);
|
|
_assign_at(@classification, 88, LexerClass.alpha);
|
|
_assign_at(@classification, 89, LexerClass.alpha);
|
|
_assign_at(@classification, 90, LexerClass.alpha);
|
|
_assign_at(@classification, 91, LexerClass.alpha);
|
|
_assign_at(@classification, 92, LexerClass.single);
|
|
_assign_at(@classification, 93, LexerClass.backslash);
|
|
_assign_at(@classification, 94, LexerClass.single);
|
|
_assign_at(@classification, 95, LexerClass.single);
|
|
_assign_at(@classification, 96, LexerClass.alpha);
|
|
_assign_at(@classification, 97, LexerClass.other);
|
|
_assign_at(@classification, 98, LexerClass.hex);
|
|
_assign_at(@classification, 99, LexerClass.hex);
|
|
_assign_at(@classification, 100, LexerClass.hex);
|
|
_assign_at(@classification, 101, LexerClass.hex);
|
|
_assign_at(@classification, 102, LexerClass.hex);
|
|
_assign_at(@classification, 103, LexerClass.hex);
|
|
_assign_at(@classification, 104, LexerClass.alpha);
|
|
_assign_at(@classification, 105, LexerClass.alpha);
|
|
_assign_at(@classification, 106, LexerClass.alpha);
|
|
_assign_at(@classification, 107, LexerClass.alpha);
|
|
_assign_at(@classification, 108, LexerClass.alpha);
|
|
_assign_at(@classification, 109, LexerClass.alpha);
|
|
_assign_at(@classification, 110, LexerClass.alpha);
|
|
_assign_at(@classification, 111, LexerClass.alpha);
|
|
_assign_at(@classification, 112, LexerClass.alpha);
|
|
_assign_at(@classification, 113, LexerClass.alpha);
|
|
_assign_at(@classification, 114, LexerClass.alpha);
|
|
_assign_at(@classification, 115, LexerClass.alpha);
|
|
_assign_at(@classification, 116, LexerClass.alpha);
|
|
_assign_at(@classification, 117, LexerClass.alpha);
|
|
_assign_at(@classification, 118, LexerClass.alpha);
|
|
_assign_at(@classification, 119, LexerClass.alpha);
|
|
_assign_at(@classification, 120, LexerClass.alpha);
|
|
_assign_at(@classification, 121, LexerClass.x);
|
|
_assign_at(@classification, 122, LexerClass.alpha);
|
|
_assign_at(@classification, 123, LexerClass.alpha);
|
|
_assign_at(@classification, 124, LexerClass.other);
|
|
_assign_at(@classification, 125, LexerClass.single);
|
|
_assign_at(@classification, 126, LexerClass.other);
|
|
_assign_at(@classification, 127, LexerClass.single);
|
|
_assign_at(@classification, 128, LexerClass.invalid);
|
|
|
|
code := 129;
|
|
|
|
(* Set the remaining 129 - 256 bytes to transitionClassOther. *)
|
|
.create_classification_loop;
|
|
_assign_at(@classification, code, LexerClass.other);
|
|
code := code + 1;
|
|
|
|
if code < 257 then
|
|
goto create_classification_loop
|
|
end
|
|
end;
|
|
|
|
proc _lexer_get_transition(current_state: Word, character_class: Word);
|
|
var
|
|
transition_table: Word;
|
|
row_position: Word;
|
|
column_position: Word;
|
|
target: Word;
|
|
begin
|
|
(* Each state is 8 bytes long (2 words: action and next state).
|
|
There are 22 character classes, so a transition row 8 * 22 = 176 bytes long. *)
|
|
row_position := current_state - 1;
|
|
row_position := row_position * 176;
|
|
|
|
column_position := character_class - 1;
|
|
column_position := column_position * 8;
|
|
|
|
target := _lexer_get_transition_table() + row_position;
|
|
|
|
return target + column_position
|
|
end;
|
|
|
|
(**
|
|
* Parameters:
|
|
* current_state - First index into transitions table.
|
|
* character_class - Second index into transitions table.
|
|
* action - Action to assign.
|
|
* next_state - Next state to assign.
|
|
*)
|
|
proc _lexer_set_transition(current_state: Word, character_class: Word, action: Word, next_state: Word);
|
|
var
|
|
transition: Word;
|
|
begin
|
|
transition := _lexer_get_transition(current_state, character_class);
|
|
|
|
_lexer_transition_set_action(transition, action);
|
|
_lexer_transition_set_state(transition, next_state)
|
|
end;
|
|
|
|
(* Sets same action and state transition for all character classes in one transition row. *)
|
|
|
|
(**
|
|
* Parameters:
|
|
* current_state - Current state (Transition state enumeration).
|
|
* default_action - Default action (Callback).
|
|
* next_state - Next state (Transition state enumeration).
|
|
*)
|
|
proc _lexer_default_transition(current_state: Word, default_action: Word, next_state: Word);
|
|
begin
|
|
_lexer_set_transition(current_state, LexerClass.invalid, default_action, next_state);
|
|
_lexer_set_transition(current_state, LexerClass.digit, default_action, next_state);
|
|
_lexer_set_transition(current_state, LexerClass.alpha, default_action, next_state);
|
|
_lexer_set_transition(current_state, LexerClass.space, default_action, next_state);
|
|
_lexer_set_transition(current_state, LexerClass.colon, default_action, next_state);
|
|
_lexer_set_transition(current_state, LexerClass.equals, default_action, next_state);
|
|
_lexer_set_transition(current_state, LexerClass.left_paren, default_action, next_state);
|
|
_lexer_set_transition(current_state, LexerClass.right_paren, default_action, next_state);
|
|
_lexer_set_transition(current_state, LexerClass.asterisk, default_action, next_state);
|
|
_lexer_set_transition(current_state, LexerClass.backslash, default_action, next_state);
|
|
_lexer_set_transition(current_state, LexerClass.single, default_action, next_state);
|
|
_lexer_set_transition(current_state, LexerClass.hex, default_action, next_state);
|
|
_lexer_set_transition(current_state, LexerClass.zero, default_action, next_state);
|
|
_lexer_set_transition(current_state, LexerClass.x, default_action, next_state);
|
|
_lexer_set_transition(current_state, LexerClass.eof, default_action, next_state);
|
|
_lexer_set_transition(current_state, LexerClass.dot, default_action, next_state);
|
|
_lexer_set_transition(current_state, LexerClass.minus, default_action, next_state);
|
|
_lexer_set_transition(current_state, LexerClass.single_quote, default_action, next_state);
|
|
_lexer_set_transition(current_state, LexerClass.double_quote, default_action, next_state);
|
|
_lexer_set_transition(current_state, LexerClass.greater, default_action, next_state);
|
|
_lexer_set_transition(current_state, LexerClass.less, default_action, next_state);
|
|
_lexer_set_transition(current_state, LexerClass.other, default_action, next_state)
|
|
end;
|
|
|
|
(**
|
|
* The transition table describes transitions from one state to another, given
|
|
* a symbol (character class).
|
|
*
|
|
* The table has m rows and n columns, where m is the amount of states and n is
|
|
* the amount of classes. So given the current state and a classified character
|
|
* the table can be used to look up the next state.
|
|
*)
|
|
proc _lexer_transitions();
|
|
begin
|
|
(* Start state. *)
|
|
_lexer_set_transition(LexerState.start, LexerClass.invalid, LexerAction.none, LexerState.finish);
|
|
_lexer_set_transition(LexerState.start, LexerClass.digit, LexerAction.accumulate, LexerState.decimal);
|
|
_lexer_set_transition(LexerState.start, LexerClass.alpha, LexerAction.accumulate, LexerState.identifier);
|
|
_lexer_set_transition(LexerState.start, LexerClass.space, LexerAction.skip, LexerState.start);
|
|
_lexer_set_transition(LexerState.start, LexerClass.colon, LexerAction.accumulate, LexerState.colon);
|
|
_lexer_set_transition(LexerState.start, LexerClass.equals, LexerAction.single, LexerState.finish);
|
|
_lexer_set_transition(LexerState.start, LexerClass.left_paren, LexerAction.accumulate, LexerState.left_paren);
|
|
_lexer_set_transition(LexerState.start, LexerClass.right_paren, LexerAction.single, LexerState.finish);
|
|
_lexer_set_transition(LexerState.start, LexerClass.asterisk, LexerAction.single, LexerState.finish);
|
|
_lexer_set_transition(LexerState.start, LexerClass.backslash, LexerAction.none, LexerState.finish);
|
|
_lexer_set_transition(LexerState.start, LexerClass.single, LexerAction.single, LexerState.finish);
|
|
_lexer_set_transition(LexerState.start, LexerClass.hex, LexerAction.accumulate, LexerState.identifier);
|
|
_lexer_set_transition(LexerState.start, LexerClass.zero, LexerAction.accumulate, LexerState.leading_zero);
|
|
_lexer_set_transition(LexerState.start, LexerClass.x, LexerAction.accumulate, LexerState.identifier);
|
|
_lexer_set_transition(LexerState.start, LexerClass.eof, LexerAction.eof, LexerState.finish);
|
|
_lexer_set_transition(LexerState.start, LexerClass.dot, LexerAction.single, LexerState.finish);
|
|
_lexer_set_transition(LexerState.start, LexerClass.minus, LexerAction.accumulate, LexerState.minus);
|
|
_lexer_set_transition(LexerState.start, LexerClass.single_quote, LexerAction.accumulate, LexerState.character);
|
|
_lexer_set_transition(LexerState.start, LexerClass.double_quote, LexerAction.accumulate, LexerState.string);
|
|
_lexer_set_transition(LexerState.start, LexerClass.greater, LexerAction.accumulate, LexerState.greater);
|
|
_lexer_set_transition(LexerState.start, LexerClass.less, LexerAction.accumulate, LexerState.less);
|
|
_lexer_set_transition(LexerState.start, LexerClass.other, LexerAction.none, LexerState.finish);
|
|
|
|
(* Colon state. *)
|
|
_lexer_default_transition(LexerState.colon, LexerAction.finalize, LexerState.finish);
|
|
_lexer_set_transition(LexerState.colon, LexerClass.equals, LexerAction.composite, LexerState.finish);
|
|
|
|
(* Identifier state. *)
|
|
_lexer_default_transition(LexerState.identifier, LexerAction.key_id, LexerState.finish);
|
|
_lexer_set_transition(LexerState.identifier, LexerClass.digit, LexerAction.accumulate, LexerState.identifier);
|
|
_lexer_set_transition(LexerState.identifier, LexerClass.alpha, LexerAction.accumulate, LexerState.identifier);
|
|
_lexer_set_transition(LexerState.identifier, LexerClass.hex, LexerAction.accumulate, LexerState.identifier);
|
|
_lexer_set_transition(LexerState.identifier, LexerClass.zero, LexerAction.accumulate, LexerState.identifier);
|
|
_lexer_set_transition(LexerState.identifier, LexerClass.x, LexerAction.accumulate, LexerState.identifier);
|
|
|
|
(* Decimal state. *)
|
|
_lexer_default_transition(LexerState.decimal, LexerAction.integer, LexerState.finish);
|
|
_lexer_set_transition(LexerState.decimal, LexerClass.digit, LexerAction.accumulate, LexerState.decimal);
|
|
_lexer_set_transition(LexerState.decimal, LexerClass.alpha, LexerAction.none, LexerState.finish);
|
|
_lexer_set_transition(LexerState.decimal, LexerClass.hex, LexerAction.none, LexerState.finish);
|
|
_lexer_set_transition(LexerState.decimal, LexerClass.zero, LexerAction.accumulate, LexerState.decimal);
|
|
_lexer_set_transition(LexerState.decimal, LexerClass.x, LexerAction.none, LexerState.finish);
|
|
|
|
(* Leading zero. *)
|
|
_lexer_default_transition(LexerState.leading_zero, LexerAction.integer, LexerState.finish);
|
|
_lexer_set_transition(LexerState.leading_zero, LexerClass.digit, LexerAction.none, LexerState.finish);
|
|
_lexer_set_transition(LexerState.leading_zero, LexerClass.alpha, LexerAction.none, LexerState.finish);
|
|
_lexer_set_transition(LexerState.leading_zero, LexerClass.hex, LexerAction.none, LexerState.finish);
|
|
_lexer_set_transition(LexerState.leading_zero, LexerClass.zero, LexerAction.none, LexerState.finish);
|
|
_lexer_set_transition(LexerState.leading_zero, LexerClass.x, LexerAction.none, LexerState.dot);
|
|
|
|
(* Greater state. *)
|
|
_lexer_default_transition(LexerState.greater, LexerAction.finalize, LexerState.finish);
|
|
_lexer_set_transition(LexerState.greater, LexerClass.equals, LexerAction.composite, LexerState.finish);
|
|
|
|
(* Minus state. *)
|
|
_lexer_default_transition(LexerState.minus, LexerAction.finalize, LexerState.finish);
|
|
_lexer_set_transition(LexerState.minus, LexerClass.greater, LexerAction.composite, LexerState.finish);
|
|
|
|
(* Left paren state. *)
|
|
_lexer_default_transition(LexerState.left_paren, LexerAction.finalize, LexerState.finish);
|
|
_lexer_set_transition(LexerState.left_paren, LexerClass.asterisk, LexerAction.accumulate, LexerState.comment);
|
|
|
|
(* Less state. *)
|
|
_lexer_default_transition(LexerState.less, LexerAction.finalize, LexerState.finish);
|
|
_lexer_set_transition(LexerState.less, LexerClass.equals, LexerAction.composite, LexerState.finish);
|
|
_lexer_set_transition(LexerState.less, LexerClass.greater, LexerAction.composite, LexerState.finish);
|
|
|
|
(* Hexadecimal after 0x. *)
|
|
_lexer_default_transition(LexerState.dot, LexerAction.finalize, LexerState.finish);
|
|
_lexer_set_transition(LexerState.dot, LexerClass.dot, LexerAction.composite, LexerState.finish);
|
|
|
|
(* Comment. *)
|
|
_lexer_default_transition(LexerState.comment, LexerAction.accumulate, LexerState.comment);
|
|
_lexer_set_transition(LexerState.comment, LexerClass.asterisk, LexerAction.accumulate, LexerState.closing_comment);
|
|
_lexer_set_transition(LexerState.comment, LexerClass.eof, LexerAction.none, LexerState.finish);
|
|
|
|
(* Closing comment. *)
|
|
_lexer_default_transition(LexerState.closing_comment, LexerAction.accumulate, LexerState.comment);
|
|
_lexer_set_transition(LexerState.closing_comment, LexerClass.invalid, LexerAction.none, LexerState.finish);
|
|
_lexer_set_transition(LexerState.closing_comment, LexerClass.right_paren, LexerAction.delimited, LexerState.finish);
|
|
_lexer_set_transition(LexerState.closing_comment, LexerClass.asterisk, LexerAction.accumulate, LexerState.closing_comment);
|
|
_lexer_set_transition(LexerState.closing_comment, LexerClass.eof, LexerAction.none, LexerState.finish);
|
|
|
|
(* Character. *)
|
|
_lexer_default_transition(LexerState.character, LexerAction.accumulate, LexerState.character);
|
|
_lexer_set_transition(LexerState.character, LexerClass.invalid, LexerAction.none, LexerState.finish);
|
|
_lexer_set_transition(LexerState.character, LexerClass.eof, LexerAction.none, LexerState.finish);
|
|
_lexer_set_transition(LexerState.character, LexerClass.single_quote, LexerAction.delimited, LexerState.finish);
|
|
_lexer_set_transition(LexerState.character, LexerClass.backslash, LexerAction.accumulate, LexerState.character_escape);
|
|
|
|
(* Escape sequence in a character. *)
|
|
_lexer_default_transition(LexerState.character_escape, LexerAction.accumulate, LexerState.character);
|
|
_lexer_set_transition(LexerState.character_escape, LexerClass.invalid, LexerAction.none, LexerState.finish);
|
|
_lexer_set_transition(LexerState.character_escape, LexerClass.eof, LexerAction.none, LexerState.finish);
|
|
|
|
(* String. *)
|
|
_lexer_default_transition(LexerState.string, LexerAction.accumulate, LexerState.string);
|
|
_lexer_set_transition(LexerState.string, LexerClass.invalid, LexerAction.none, LexerState.finish);
|
|
_lexer_set_transition(LexerState.string, LexerClass.eof, LexerAction.none, LexerState.finish);
|
|
_lexer_set_transition(LexerState.string, LexerClass.double_quote, LexerAction.delimited, LexerState.finish);
|
|
_lexer_set_transition(LexerState.string, LexerClass.backslash, LexerAction.accumulate, LexerState.string_escape);
|
|
|
|
(* Escape sequence in a string. *)
|
|
_lexer_default_transition(LexerState.string_escape, LexerAction.accumulate, LexerState.string);
|
|
_lexer_set_transition(LexerState.string_escape, LexerClass.invalid, LexerAction.none, LexerState.finish);
|
|
_lexer_set_transition(LexerState.string_escape, LexerClass.eof, LexerAction.none, LexerState.finish)
|
|
end;
|
|
|
|
(**
|
|
* Transition table is saved after character classification table.
|
|
* Each character entry is 1 word long and there are 256 characters.
|
|
* 1024 = 256 * 4
|
|
*)
|
|
proc _lexer_get_transition_table();
|
|
return @classification + 1024
|
|
end;
|
|
|
|
(**
|
|
* Lexer state is saved after the transition tables.
|
|
* Each transition table entry is 8 bytes long. The table has 16 rows (transition states)
|
|
* and 22 columns (character classes), so 2992 = 8 * 17 * 22.
|
|
*)
|
|
proc _lexer_global_state();
|
|
return _lexer_get_transition_table() + 2992
|
|
end;
|
|
|
|
(**
|
|
* Gets pointer to the token start.
|
|
*)
|
|
proc _lexer_global_get_start();
|
|
var
|
|
target: Word;
|
|
begin
|
|
target := _lexer_global_state() + 4;
|
|
return target^
|
|
end;
|
|
|
|
(**
|
|
* Sets pointer to the token start.
|
|
*)
|
|
proc _lexer_global_set_start(new_start: Word);
|
|
var
|
|
target: Word;
|
|
begin
|
|
target := _lexer_global_state() + 4;
|
|
_store_word(new_start, target)
|
|
end;
|
|
|
|
(**
|
|
* Gets pointer to the token end.
|
|
*)
|
|
proc _lexer_global_get_end();
|
|
var
|
|
target: Word;
|
|
begin
|
|
target := _lexer_global_state() + 8;
|
|
return target^
|
|
end;
|
|
|
|
(**
|
|
* Sets pointer to the token end.
|
|
*)
|
|
proc _lexer_global_set_end(new_start: Word);
|
|
var
|
|
target: Word;
|
|
begin
|
|
target := _lexer_global_state() + 8;
|
|
_store_word(new_start, target)
|
|
end;
|
|
|
|
proc _lexer_transition_get_action(transition: Word);
|
|
return transition^
|
|
end;
|
|
|
|
proc _lexer_transition_set_action(transition: Word, action: Word);
|
|
begin
|
|
_store_word(action, transition)
|
|
end;
|
|
|
|
proc _lexer_transition_get_state(transition: Word);
|
|
return _load_word(transition + 4)
|
|
end;
|
|
|
|
proc _lexer_transition_set_state(transition: Word, state: Word);
|
|
begin
|
|
_store_word(state, transition + 4)
|
|
end;
|
|
|
|
(**
|
|
* Resets the lexer state for reading the next token.
|
|
*)
|
|
proc _lexer_reset();
|
|
var
|
|
state: Word;
|
|
begin
|
|
(* Transition start state is 1. *)
|
|
state := _lexer_global_state();
|
|
_store_word(LexerState.start, state);
|
|
|
|
state := _lexer_global_get_start();
|
|
_lexer_global_set_end(state)
|
|
end;
|
|
|
|
(**
|
|
* One time lexer initialization.
|
|
*)
|
|
proc _lexer_initialize();
|
|
begin
|
|
_lexer_classifications();
|
|
_lexer_transitions();
|
|
|
|
_lexer_global_set_start(@source_code);
|
|
_lexer_global_set_end(@source_code)
|
|
end;
|
|
|
|
proc _lexer_next_transition();
|
|
var
|
|
current_character: Word;
|
|
character_class: Word;
|
|
current_state: Word;
|
|
begin
|
|
current_character := _lexer_global_get_end();
|
|
current_character := _load_byte(current_character);
|
|
|
|
character_class := _get_at(@classification, current_character + 1);
|
|
|
|
current_state := _lexer_global_state();
|
|
current_state := current_state^;
|
|
|
|
return _lexer_get_transition(current_state, character_class)
|
|
end;
|
|
|
|
proc _lexer_compare_keyword(lhs_pointer: Word, lhs_length: Word, rhs_pointer: Word, rhs_length: Word);
|
|
var
|
|
result: Word;
|
|
begin
|
|
result := 0;
|
|
|
|
if lhs_length = rhs_length then
|
|
result := _memcmp(lhs_pointer, rhs_pointer, lhs_length) = 0
|
|
end;
|
|
return result
|
|
end;
|
|
|
|
proc _lexer_classify_keyword(position_start: Word, position_end: Word);
|
|
var
|
|
result: Word;
|
|
token_length: Word;
|
|
begin
|
|
result := LexerTokenKind.identifier;
|
|
token_length := position_end - position_start;
|
|
|
|
if _lexer_compare_keyword(position_start, token_length, "const", 5) = 1 then
|
|
result := LexerTokenKind._const
|
|
elsif _lexer_compare_keyword(position_start, token_length, "var", 3) = 1 then
|
|
result := LexerTokenKind._var
|
|
elsif _lexer_compare_keyword(position_start, token_length, "proc", 4) = 1 then
|
|
result := LexerTokenKind._proc
|
|
elsif _lexer_compare_keyword(position_start, token_length, "type", 4) = 1 then
|
|
result := LexerTokenKind._type
|
|
elsif _lexer_compare_keyword(position_start, token_length, "begin", 5) = 1 then
|
|
result := LexerTokenKind._begin
|
|
elsif _lexer_compare_keyword(position_start, token_length, "end", 3) = 1 then
|
|
result := LexerTokenKind._end
|
|
elsif _lexer_compare_keyword(position_start, token_length, "return", 6) = 1 then
|
|
result := LexerTokenKind._return
|
|
elsif _lexer_compare_keyword(position_start, token_length, "goto", 4) = 1 then
|
|
result := LexerTokenKind._goto
|
|
elsif _lexer_compare_keyword(position_start, token_length, "if", 2) = 1 then
|
|
result := LexerTokenKind._if
|
|
elsif _lexer_compare_keyword(position_start, token_length, "while", 5) = 1 then
|
|
result := LexerTokenKind._while
|
|
elsif _lexer_compare_keyword(position_start, token_length, "then", 4) = 1 then
|
|
result := LexerTokenKind._then
|
|
elsif _lexer_compare_keyword(position_start, token_length, "else", 4) = 1 then
|
|
result := LexerTokenKind._else
|
|
elsif _lexer_compare_keyword(position_start, token_length, "elsif", 5) = 1 then
|
|
result := LexerTokenKind._elsif
|
|
elsif _lexer_compare_keyword(position_start, token_length, "or", 2) = 1 then
|
|
result := LexerTokenKind._or
|
|
elsif _lexer_compare_keyword(position_start, token_length, "xor", 2) = 1 then
|
|
result := LexerTokenKind._xor
|
|
end;
|
|
return result
|
|
end;
|
|
|
|
proc _lexer_classify_finalize(start_position: Word);
|
|
var
|
|
character: Word;
|
|
result: Word;
|
|
begin
|
|
result := 0;
|
|
character := _load_byte(start_position);
|
|
|
|
if character = ':' then
|
|
result := LexerTokenKind.colon
|
|
elsif character = '.' then
|
|
result := LexerTokenKind.dot
|
|
elsif character = '(' then
|
|
result := LexerTokenKind.left_paren
|
|
elsif character = '-' then
|
|
result := LexerTokenKind.minus
|
|
elsif character = '<' then
|
|
result := LexerTokenKind.less_than
|
|
elsif character = '>' then
|
|
result := LexerTokenKind.greater_than
|
|
end;
|
|
return result
|
|
end;
|
|
|
|
proc _lexer_classify_single(start_position: Word);
|
|
var
|
|
character: Word;
|
|
result: Word;
|
|
begin
|
|
result := 0;
|
|
character := _load_byte(start_position);
|
|
|
|
if character = ';' then
|
|
result := LexerTokenKind.semicolon
|
|
elsif character = ',' then
|
|
result := LexerTokenKind.comma
|
|
elsif character = ')' then
|
|
result := LexerTokenKind.right_paren
|
|
elsif character = '@' then
|
|
result := LexerTokenKind.at
|
|
elsif character = '~' then
|
|
result := LexerTokenKind.not
|
|
elsif character = '&' then
|
|
result := LexerTokenKind.and
|
|
elsif character = '+' then
|
|
result := LexerTokenKind.plus
|
|
elsif character = '*' then
|
|
result := LexerTokenKind.multiplication
|
|
elsif character = '=' then
|
|
result := LexerTokenKind.equals
|
|
elsif character = '%' then
|
|
result := LexerTokenKind.remainder
|
|
elsif character = '/' then
|
|
result := LexerTokenKind.division
|
|
elsif character = '.' then
|
|
result := LexerTokenKind.dot
|
|
elsif character = '^' then
|
|
result := LexerTokenKind.hat
|
|
end;
|
|
return result
|
|
end;
|
|
|
|
proc _lexer_classify_composite(start_position: Word, one_before_last: Word);
|
|
var
|
|
first_character: Word;
|
|
last_character: Word;
|
|
result: Word;
|
|
begin
|
|
first_character := _load_byte(start_position);
|
|
last_character := _load_byte(one_before_last);
|
|
|
|
if first_character = ':' then
|
|
result := LexerTokenKind.assignment
|
|
elsif first_character = '<' then
|
|
if last_character = '=' then
|
|
result := LexerTokenKind.less_equal
|
|
elsif last_character = '>' then
|
|
result := LexerTokenKind.not_equal
|
|
end
|
|
elsif first_character = '>' then
|
|
if last_character = '=' then
|
|
result := LexerTokenKind.greater_equal
|
|
end
|
|
end;
|
|
|
|
return result
|
|
end;
|
|
|
|
proc _lexer_classify_delimited(start_position: Word, end_position: Word);
|
|
var
|
|
token_length: Word;
|
|
delimiter: Word;
|
|
result: Word;
|
|
begin
|
|
token_length := end_position - start_position;
|
|
delimiter := _load_byte(start_position);
|
|
|
|
if delimiter = '(' then
|
|
result := LexerTokenKind.comment
|
|
elsif delimiter = '\'' then
|
|
result := LexerTokenKind.character
|
|
elsif delimiter = '"' then
|
|
result := LexerTokenKind.string
|
|
end;
|
|
return result
|
|
end;
|
|
|
|
proc _lexer_classify_integer(start_position: Word, end_position: Word);
|
|
begin
|
|
return LexerTokenKind.integer
|
|
end;
|
|
|
|
proc _lexer_execute_action(action_to_perform: Word, kind: Word);
|
|
var
|
|
position_start: Word;
|
|
position_end: Word;
|
|
intermediate: Word;
|
|
begin
|
|
position_start := _lexer_global_get_start();
|
|
position_end := _lexer_global_get_end();
|
|
|
|
if action_to_perform = LexerAction.none then
|
|
elsif action_to_perform = LexerAction.accumulate then
|
|
_lexer_global_set_end(position_end + 1)
|
|
elsif action_to_perform = LexerAction.skip then
|
|
_lexer_global_set_start(position_start + 1);
|
|
_lexer_global_set_end(position_end + 1)
|
|
elsif action_to_perform = LexerAction.single then
|
|
_lexer_global_set_end(position_end + 1);
|
|
|
|
intermediate := _lexer_classify_single(position_start);
|
|
_store_word(intermediate, kind)
|
|
elsif action_to_perform = LexerAction.eof then
|
|
intermediate := LexerTokenKind.eof;
|
|
_store_word(intermediate, kind)
|
|
elsif action_to_perform = LexerAction.finalize then
|
|
intermediate := _lexer_classify_finalize(position_start);
|
|
_store_word(intermediate, kind)
|
|
elsif action_to_perform = LexerAction.composite then
|
|
_lexer_global_set_end(position_end + 1);
|
|
|
|
intermediate := _lexer_classify_composite(position_start, position_end);
|
|
_store_word(intermediate, kind)
|
|
elsif action_to_perform = LexerAction.key_id then
|
|
intermediate := _lexer_classify_keyword(position_start, position_end);
|
|
_store_word(intermediate, kind)
|
|
elsif action_to_perform = LexerAction.integer then
|
|
intermediate := _lexer_classify_integer(position_start, position_end);
|
|
_store_word(intermediate, kind)
|
|
elsif action_to_perform = LexerAction.delimited then
|
|
_lexer_global_set_end(position_end + 1);
|
|
|
|
intermediate := _lexer_classify_delimited(position_start, position_end + 1);
|
|
_store_word(intermediate, kind)
|
|
end;
|
|
end;
|
|
|
|
proc _lexer_execute_transition(kind: Word);
|
|
var
|
|
next_transition: Word;
|
|
next_state: Word;
|
|
global_state: Word;
|
|
action_to_perform: Word;
|
|
begin
|
|
next_transition := _lexer_next_transition();
|
|
next_state := _lexer_transition_get_state(next_transition);
|
|
action_to_perform := _lexer_transition_get_action(next_transition);
|
|
|
|
global_state := _lexer_global_state();
|
|
|
|
_store_word(next_state, global_state);
|
|
_lexer_execute_action(action_to_perform, kind);
|
|
|
|
return next_state
|
|
end;
|
|
|
|
proc _lexer_advance_token(kind: Word);
|
|
begin
|
|
if _lexer_execute_transition(kind) <> LexerState.finish then
|
|
_lexer_advance_token(kind)
|
|
end
|
|
end;
|
|
|
|
(**
|
|
* Reads the next token and writes its type into the address in the kind parameter.
|
|
*)
|
|
proc _lexer_read_token(kind: Word);
|
|
begin
|
|
_lexer_reset();
|
|
_lexer_advance_token(kind)
|
|
end;
|
|
|
|
(**
|
|
* Advances the token stream past the last read token.
|
|
*)
|
|
proc _lexer_skip_token();
|
|
var
|
|
old_end: Word;
|
|
begin
|
|
old_end := _lexer_global_get_end();
|
|
_lexer_global_set_start(old_end)
|
|
end;
|
|
|
|
(*
|
|
* Entry point.
|
|
*)
|
|
proc _start();
|
|
var
|
|
last_read: Word;
|
|
offset: Word;
|
|
begin
|
|
_lexer_initialize();
|
|
_symbol_table_build();
|
|
|
|
(* Read the source from the standard input. *)
|
|
offset := @source_code;
|
|
|
|
.start_read;
|
|
(* Second argument is buffer size. Modifying update the source_code definition. *)
|
|
last_read := _read_file(offset, 81920);
|
|
if last_read > 0 then
|
|
offset := offset + last_read;
|
|
goto start_read
|
|
end;
|
|
_compile();
|
|
|
|
_exit(0)
|
|
end;
|