From 8ccb63530238f90e1d6611a6e1132d12c40962fe Mon Sep 17 00:00:00 2001 From: Eugen Wissner Date: Thu, 25 Sep 2025 10:38:19 +0200 Subject: [PATCH] Make begin optional in procedures with only return --- boot/stage13.elna | 43 ++-- boot/stage14.elna | 575 ++++++++++++++++++---------------------------- 2 files changed, 254 insertions(+), 364 deletions(-) diff --git a/boot/stage13.elna b/boot/stage13.elna index 925a1cd..f4fa817 100644 --- a/boot/stage13.elna +++ b/boot/stage13.elna @@ -6,6 +6,7 @@ (* - Multiline comments. *) (* - elsif conditions. *) +(* - Optional "begin" if the procedure body is a single return statement. *) const symbol_builtin_name_int := "Int"; @@ -71,7 +72,7 @@ begin if _load_byte(string) <> '"' then counter := counter + 1; - goto .string_length_loop; + goto .string_length_loop end; return counter @@ -100,9 +101,9 @@ begin contents := contents + 1; if current_byte <> '\\' then - compiler_strings_length := compiler_strings_length + 1; + compiler_strings_length := compiler_strings_length + 1 end; - goto .add_string_loop; + goto .add_string_loop end; return result @@ -145,10 +146,10 @@ begin local_buffer := @result + 11; if number >= 0 then - is_negative := 0; + is_negative := 0 else number = -number; - is_negative := 1; + is_negative := 1 end; .print_i_digit10; @@ -159,11 +160,11 @@ begin local_buffer := local_buffer + -1; if number <> 0 then - goto .print_i_digit10; + goto .print_i_digit10 end; if is_negative = 1 then _store_byte('-', local_buffer); - local_buffer := local_buffer + -1; + local_buffer := local_buffer + -1 end; result := @result + 11; result := result + -local_buffer; @@ -210,7 +211,7 @@ begin _write_c(next_byte); (* Advance the input string by one byte. *) - _write_z(string + 1); + _write_z(string + 1) end; end; @@ -314,8 +315,8 @@ begin count := count + -1; if result = 0 then - goto .memcmp_loop; - end; + goto .memcmp_loop + end end; return result @@ -341,7 +342,7 @@ begin destination := destination + 1; source := source + 1; count := count + -1; - goto .memcpy_loop; + goto .memcpy_loop end; return destination @@ -383,7 +384,7 @@ begin character := _load_byte(source_code_position); if character = '\\' then _write_c('\\'); - source_code_position := source_code_position + 1; + source_code_position := source_code_position + 1 end; _write_s(source_code_position, 1); _write_s("'\n", 2); @@ -981,9 +982,14 @@ end; proc _skip_spaces(); var current_byte: Word; + lhs: Word; + rhs: Word; begin current_byte := _load_byte(source_code_position); - if current_byte = '\t' then + lhs := current_byte = '\t'; + rhs := current_byte = ' '; + + if lhs or rhs then source_code_position := source_code_position + 1; _skip_spaces(); end; @@ -1194,9 +1200,14 @@ begin _read_procedure_temporaries(); (* Skip semicolon, "begin" and newline. *) - source_code_position := source_code_position + 6; - - _compile_procedure_body(); + _lexer_read_token(@token_kind); + if token_kind = _lexer_token_kind_begin() then + _lexer_skip_token(); + _compile_procedure_body(); + end; + if token_kind = _lexer_token_kind_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"); diff --git a/boot/stage14.elna b/boot/stage14.elna index fd3d2d3..e3091dc 100644 --- a/boot/stage14.elna +++ b/boot/stage14.elna @@ -18,7 +18,6 @@ const PRIMITIVE_TYPE = 1 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); @@ -32,7 +31,6 @@ const INFO_TEMPORARY = 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); @@ -74,7 +72,7 @@ begin if _load_byte(string) <> '"' then counter := counter + 1; - goto .string_length_loop; + goto .string_length_loop end; return counter @@ -105,9 +103,9 @@ begin contents := contents + 1; if current_byte <> '\\' then - compiler_strings_length := compiler_strings_length + 1; + compiler_strings_length := compiler_strings_length + 1 end; - goto .add_string_loop; + goto .add_string_loop end; return result @@ -123,7 +121,6 @@ end; * Returns the amount of bytes written in a0. *) proc _read_file(buffer: Word, size: Word); -begin return _syscall(0, buffer, size, 0, 0, 0, 63) end; @@ -136,7 +133,7 @@ end; *) proc _write_s(buffer: Word, size: Word); begin - _syscall(1, buffer, size, 0, 0, 0, 64); + _syscall(1, buffer, size, 0, 0, 0, 64) end; (** @@ -158,10 +155,10 @@ begin local_buffer := @result + 11; if number >= 0 then - is_negative := 0; + is_negative := 0 else number = -number; - is_negative := 1; + is_negative := 1 end; .print_i_digit10; @@ -172,10 +169,11 @@ begin local_buffer := local_buffer + -1; if number <> 0 then - goto .print_i_digit10; - elsif is_negative = 1 then + goto .print_i_digit10 + end; + if is_negative = 1 then _store_byte('-', local_buffer); - local_buffer := local_buffer + -1; + local_buffer := local_buffer + -1 end; result := @result + 11; result := result + -local_buffer; @@ -196,7 +194,7 @@ var length: Word; begin length := _print_i(number, @local_buffer); - _write_s(@local_buffer, length); + _write_s(@local_buffer, length) end; (** @@ -207,7 +205,7 @@ end; *) proc _write_c(character: Word); begin - _write_s(@character, 1); + _write_s(@character, 1) end; (** @@ -228,11 +226,13 @@ begin _write_c(next_byte); (* Advance the input string by one byte. *) - _write_z(string + 1); - end; + _write_z(string + 1) + end end; -(* Detects if a0 is an uppercase character. Sets a0 to 1 if so, otherwise to 0. *) +(** + * Detects if a0 is an uppercase character. Sets a0 to 1 if so, otherwise to 0. + *) proc _is_upper(character: Word); var lhs: Word; @@ -245,7 +245,9 @@ begin end; -(* Detects if a0 is an lowercase character. Sets a0 to 1 if so, otherwise to 0. *) +(** + * Detects if a0 is an lowercase character. Sets a0 to 1 if so, otherwise to 0. + *) proc _is_lower(character: Word); var lhs: Word; @@ -281,8 +283,7 @@ begin end; (** - * Detects whether the passed character is a digit - * (a value between 0 and 9). + * Detects whether the passed character is a digit (a value between 0 and 9). * * Parameters: * character - Exemined value. @@ -338,8 +339,8 @@ begin count := count + -1; if result = 0 then - goto .memcmp_loop; - end; + goto .memcmp_loop + end end; return result @@ -367,7 +368,7 @@ begin destination := destination + 1; source := source + 1; count := count + -1; - goto .memcpy_loop; + goto .memcpy_loop end; return destination @@ -398,7 +399,7 @@ begin _write_token(integer_token); _lexer_skip_token(); - _write_c('\n'); + _write_c('\n') end; proc _compile_character_literal(); @@ -411,24 +412,24 @@ begin character := _load_byte(source_code_position); if character = '\\' then _write_c('\\'); - source_code_position := source_code_position + 1; + source_code_position := source_code_position + 1 end; _write_s(source_code_position, 1); _write_s("'\n", 2); - source_code_position := source_code_position + 2; + source_code_position := source_code_position + 2 end; proc _compile_variable_expression(); begin _compile_designator(); - _write_z("\tlw t0, (t0)\n\0"); + _write_z("\tlw t0, (t0)\n\0") end; proc _compile_address_expression(); begin (* Skip the "@" sign. *) source_code_position := source_code_position + 1; - _compile_designator(); + _compile_designator() end; proc _compile_negate_expression(); @@ -437,7 +438,7 @@ begin source_code_position := source_code_position + 1; _compile_term(); - _write_z("\tneg t0, t0\n\0"); + _write_z("\tneg t0, t0\n\0") end; proc _compile_not_expression(); @@ -446,7 +447,7 @@ begin source_code_position := source_code_position + 1; _compile_term(); - _write_z("\tnot t0, t0\n\0"); + _write_z("\tnot t0, t0\n\0") end; proc _compile_string_literal(); @@ -465,7 +466,7 @@ begin _write_i(offset); _write_c('\n'); - _write_z("\tadd t0, t0, t1\n\0"); + _write_z("\tadd t0, t0, t1\n\0") end; proc _compile_term(); @@ -475,23 +476,23 @@ begin current_character := _load_byte(source_code_position); if current_character = '\'' then - _compile_character_literal(); + _compile_character_literal() elsif current_character = '@' then - _compile_address_expression(); + _compile_address_expression() elsif current_character = '-' then - _compile_negate_expression(); + _compile_negate_expression() elsif current_character = '~' then - _compile_not_expression(); + _compile_not_expression() elsif current_character = '"' then - _compile_string_literal(); + _compile_string_literal() elsif current_character = '_' then _compile_call(); - _write_z("\nmv t0, a0\n\0"); + _write_z("\nmv t0, a0\n\0") elsif _is_digit(current_character) = 1 then - _compile_integer_literal(); + _compile_integer_literal() elsif _is_lower(current_character) = 1 then - _compile_variable_expression(); - end; + _compile_variable_expression() + end end; proc _compile_binary_rhs(); @@ -501,7 +502,7 @@ begin _compile_term(); (* Load the left expression from the stack; *) - _write_z("\tlw t1, 64(sp)\n\0"); + _write_z("\tlw t1, 64(sp)\n\0") end; proc _compile_expression(); @@ -512,7 +513,7 @@ begin current_character := _load_byte(source_code_position); if current_character <> ' ' then - goto .compile_expression_end; + goto .compile_expression_end end; (* It is a binary expression. *) @@ -528,49 +529,49 @@ begin _compile_binary_rhs(); (* Execute the operation. *) - _write_z("add t0, t0, t1\n\0"); + _write_z("add t0, t0, t1\n\0") elsif current_character = '*' then source_code_position := source_code_position + 1; _compile_binary_rhs(); (* Execute the operation. *) - _write_z("\tmul t0, t0, t1\n\0"); + _write_z("\tmul t0, t0, t1\n\0") elsif current_character = '&' then source_code_position := source_code_position + 1; _compile_binary_rhs(); (* Execute the operation. *) - _write_z("\tand t0, t0, t1\n\0"); + _write_z("\tand t0, t0, t1\n\0") elsif current_character = 'o' then source_code_position := source_code_position + 2; _compile_binary_rhs(); (* Execute the operation. *) - _write_z("or t0, t0, t1\n\0"); + _write_z("or t0, t0, t1\n\0") elsif current_character = 'x' then source_code_position := source_code_position + 3; _compile_binary_rhs(); (* Execute the operation. *) - _write_z("xor t0, t0, t1\n\0"); + _write_z("xor t0, t0, t1\n\0") elsif current_character = '=' then source_code_position := source_code_position + 1; _compile_binary_rhs(); (* Execute the operation. *) - _write_z("xor t0, t0, t1\nseqz t0, t0\n\0"); + _write_z("xor t0, t0, t1\nseqz t0, t0\n\0") elsif current_character = '%' then source_code_position := source_code_position + 1; _compile_binary_rhs(); (* Execute the operation. *) - _write_z("rem t0, t1, t0\n\0"); + _write_z("rem t0, t1, t0\n\0") elsif current_character = '/' then source_code_position := source_code_position + 1; _compile_binary_rhs(); (* Execute the operation. *) - _write_z("div t0, t1, t0\n\0"); + _write_z("div t0, t1, t0\n\0") elsif current_character = '<' then source_code_position := source_code_position + 1; current_character := _load_byte(source_code_position); @@ -580,19 +581,19 @@ begin _compile_binary_rhs(); (* Execute the operation. *) - _write_z("\txor t0, t0, t1\nsnez t0, t0\n\0"); + _write_z("\txor t0, t0, t1\nsnez t0, t0\n\0") elsif current_character = '=' then source_code_position := source_code_position + 1; _compile_binary_rhs(); (* Execute the operation. *) - _write_z("\tslt t0, t0, t1\nxori t0, t0, 1\n\0"); + _write_z("\tslt t0, t0, t1\nxori t0, t0, 1\n\0") else _compile_binary_rhs(); (* Execute the operation. *) - _write_z("slt t0, t1, t0\n\0"); - end; + _write_z("slt t0, t1, t0\n\0") + end elsif current_character = '>' then source_code_position := source_code_position + 1; current_character := _load_byte(source_code_position); @@ -601,13 +602,13 @@ begin _compile_binary_rhs(); (* Execute the operation. *) - _write_z("\tslt t0, t1, t0\nxori t0, t0, 1\n\0"); + _write_z("\tslt t0, t1, t0\nxori t0, t0, 1\n\0") else _compile_binary_rhs(); (* Execute the operation. *) - _write_z("\tslt t0, t0, t1\n\0"); - end; + _write_z("\tslt t0, t0, t1\n\0") + end end; .compile_expression_end; @@ -633,7 +634,7 @@ begin source_code_position := source_code_position + 1; if _load_byte(source_code_position) = ')' then - goto .compile_call_finalize; + goto .compile_call_finalize end; .compile_call_loop; _compile_expression(); @@ -651,7 +652,7 @@ begin argument_count := argument_count + 1; if _load_byte(source_code_position) <> ',' then - goto .compile_call_finalize; + goto .compile_call_finalize end; source_code_position := source_code_position + 2; goto .compile_call_loop; @@ -673,7 +674,7 @@ begin _write_z("(sp)\n\0"); - goto .compile_call_finalize; + goto .compile_call_finalize end; .compile_call_end; @@ -681,7 +682,7 @@ begin _write_s(name, name_length); (* Skip the right paren. *) - source_code_position := source_code_position + 1; + source_code_position := source_code_position + 1 end; proc _compile_goto(); @@ -698,7 +699,7 @@ begin _write_z("\tj .\0"); _write_token(next_token); - _lexer_skip_token(); + _lexer_skip_token() end; proc _compile_local_designator(symbol: Word); @@ -709,7 +710,7 @@ begin variable_offset := _parameter_info_get_offset(symbol); _write_i(variable_offset); _write_c('\n'); - _lexer_skip_token(); + _lexer_skip_token() end; proc _compile_global_designator(); @@ -723,7 +724,7 @@ begin _write_token(name); _lexer_skip_token(); - _write_c('\n'); + _write_c('\n') end; proc _compile_designator(); @@ -741,12 +742,10 @@ begin lookup_result := _symbol_table_lookup(@symbol_table_local, name, name_token); if lookup_result <> 0 then - _compile_local_designator(lookup_result); - goto .compile_designator_end; - end; - _compile_global_designator(); - - .compile_designator_end; + _compile_local_designator(lookup_result) + else + _compile_global_designator() + end end; proc _compile_assignment(); @@ -762,7 +761,7 @@ begin (* Compile the assignment. *) _compile_expression(); - _write_z("\tlw t1, 60(sp)\nsw t0, (t1)\n\0"); + _write_z("\tlw t1, 60(sp)\nsw t0, (t1)\n\0") end; proc _compile_return_statement(); @@ -775,7 +774,7 @@ begin source_code_position := source_code_position + 1; _compile_expression(); - _write_z("\tmv a0, t0\n\0"); + _write_z("\tmv a0, t0\n\0") end; (** @@ -787,18 +786,14 @@ end; proc _write_label(counter: Word); begin _write_z(".L\0"); - _write_i(counter); + _write_i(counter) end; -proc _compile_if(); +proc _compile_condition(after_end_label: Word); var - after_end_label: Word; condition_label: Word; token_kind: Word; begin - (* Skip "if ". *) - _lexer_read_token(@token_kind); - _lexer_skip_token(); source_code_position := source_code_position + 1; (* Compile condition. *) @@ -807,9 +802,6 @@ begin _lexer_read_token(@token_kind); _lexer_skip_token(); - after_end_label := label_counter; - label_counter := label_counter + 1; - (* condition_label is the label in front of the next elsif condition or end. *) condition_label := label_counter; label_counter := label_counter + 1; @@ -825,47 +817,39 @@ begin _write_c('\n'); _write_label(condition_label); - _write_z(":\n\0"); + _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 = _lexer_token_kind_else() then _lexer_skip_token(); - _compile_procedure_body(); + _compile_procedure_body() elsif token_kind = _lexer_token_kind_elsif() then _lexer_skip_token(); - source_code_position := source_code_position + 1; + _compile_condition(after_end_label); - (* 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_procedure_body(); - - _write_z("\tj \0"); - _write_label(after_end_label); - _write_c('\n'); - - _write_label(condition_label); - _write_z(":\n\0"); - - goto .compile_if_loop; + goto .compile_if_loop end; _lexer_skip_token(); _write_label(after_end_label); - _write_z(":\n\0"); + _write_z(":\n\0") end; proc _compile_label_declaration(); @@ -883,7 +867,7 @@ begin _write_c('.'); _write_s(name, label_token); _write_z(":\n\0"); - _lexer_skip_token(); + _lexer_skip_token() end; proc _compile_statement(); @@ -894,13 +878,13 @@ begin _lexer_read_token(@token_kind); if token_kind = _lexer_token_kind_goto() then - _compile_goto(); + _compile_goto() elsif token_kind = _lexer_token_kind_if() then - _compile_if(); + _compile_if() elsif token_kind = _lexer_token_kind_return() then - _compile_return_statement(); + _compile_return_statement() elsif token_kind = _lexer_token_kind_dot() then - _compile_label_declaration(); + _compile_label_declaration() elsif token_kind = _lexer_token_kind_identifier() then current_byte := _lexer_global_start(); current_byte := _load_word(current_byte); @@ -908,12 +892,12 @@ begin (* This is a call if the statement starts with an underscore. *) if current_byte = '_' then - _compile_call(); + _compile_call() else - _compile_assignment(); - end; + _compile_assignment() + end end; - _write_c('\n'); + _write_c('\n') end; proc _compile_procedure_body(); @@ -926,10 +910,9 @@ begin if token_kind = _lexer_token_kind_semicolon() then _lexer_skip_token(); - _compile_procedure_body(); - else - _skip_empty_lines(); + _compile_procedure_body() end; + _skip_empty_lines() end; (** @@ -943,18 +926,23 @@ proc _write_register(register_character: Word, register_number: Word); begin _write_c(register_character); register_number := register_number + '0'; - _write_c(register_number); + _write_c(register_number) end; proc _skip_spaces(); var current_byte: Word; + lhs: Word; + rhs: Word; begin current_byte := _load_byte(source_code_position); - if current_byte = '\t' then + lhs := current_byte = '\t'; + rhs := current_byte = ' '; + + if lhs or rhs then source_code_position := source_code_position + 1; - _skip_spaces(); - end; + _skip_spaces() + end end; proc _read_type_expression(); @@ -963,12 +951,11 @@ var token_kind: Word; begin type_name := _lexer_read_token(@token_kind); - _lexer_skip_token(); + _lexer_skip_token() end; (** * Parameters: - * * parameter_index - Parameter index. *) proc _parameter_info_create(parameter_index: Word); @@ -1001,7 +988,6 @@ end; (** * Parameters: - * * temporary_index - Parameter index. *) proc _temporary_info_create(temporary_index: Word); @@ -1033,7 +1019,6 @@ end; (** * Parameters: - * * parameter_index - Parameter index. *) proc _read_procedure_parameter(parameter_index: Word); @@ -1063,7 +1048,7 @@ begin info := _parameter_info_get_offset(info); _write_i(info); - _write_z("(sp)\n\0"); + _write_z("(sp)\n\0") end; proc _read_procedure_parameters(); @@ -1081,11 +1066,11 @@ begin if _load_byte(source_code_position) = ',' then source_code_position := source_code_position + 2; - goto .compile_procedure_prologue_skip; - end; + goto .compile_procedure_prologue_skip + end end; (* Skip close paren. *) - source_code_position := source_code_position + 1; + source_code_position := source_code_position + 1 end; (** @@ -1113,25 +1098,27 @@ begin _symbol_table_enter(@symbol_table_local, name_position, name_length, info); (* Skip semicolon and newline after the variable declaration *) - source_code_position := source_code_position + 2; + source_code_position := source_code_position + 2 end; proc _read_procedure_temporaries(); var temporary_counter: Word; begin - if _memcmp(source_code_position, "var", 3) = 0 then - source_code_position := source_code_position + 4; - temporary_counter := 0; - - .read_local_variables_loop; - if _memcmp(source_code_position, "begin", 5) <> 0 then - _read_procedure_temporary(temporary_counter); - - temporary_counter := temporary_counter + 1; - goto .read_local_variables_loop; - end; + if _memcmp(source_code_position, "var", 3) <> 0 then + goto .read_local_variables_end end; + source_code_position := source_code_position + 4; + temporary_counter := 0; + + .read_local_variables_loop; + if _memcmp(source_code_position, "begin", 5) <> 0 then + _read_procedure_temporary(temporary_counter); + + temporary_counter := temporary_counter + 1; + goto .read_local_variables_loop + end; + .read_local_variables_end end; proc _compile_procedure(); @@ -1166,15 +1153,20 @@ begin _read_procedure_temporaries(); (* Skip semicolon, "begin" and newline. *) - source_code_position := source_code_position + 6; - - _compile_procedure_body(); + _lexer_read_token(@token_kind); + if token_kind = _lexer_token_kind_begin() then + _lexer_skip_token(); + _compile_procedure_body() + end; + if token_kind = _lexer_token_kind_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. *) - source_code_position := source_code_position + 5; + source_code_position := source_code_position + 5 end; (** @@ -1186,7 +1178,7 @@ var begin _lexer_read_token(@token_kind); _lexer_skip_token(); - source_code_position := source_code_position + 1; + source_code_position := source_code_position + 1 end; (** @@ -1205,19 +1197,19 @@ begin if current_byte = '\n' then source_code_position := current_position + 1; - _skip_empty_lines(); + _skip_empty_lines() elsif current_byte = '\t' then current_position := current_position + 1; - goto .skip_empty_lines_loop; + goto .skip_empty_lines_loop elsif current_byte = '(' then current_byte := _load_byte(current_position + 1); if current_byte = '*' then source_code_position := current_position; _skip_comment(); - _skip_empty_lines(); + goto .skip_empty_lines_rerun end; - end; + end end; proc _compile_global_initializer(); @@ -1239,15 +1231,15 @@ begin source_code_position := source_code_position + length; source_code_position := source_code_position + 2; - goto .compile_global_initializer_end; + goto .compile_global_initializer_end elsif current_byte = 'S' then (* Skip "S(". *) source_code_position := source_code_position + 2; if _load_byte(source_code_position) = ')' then - goto .compile_global_initializer_closing; + goto .compile_global_initializer_closing end; - goto .compile_global_initializer_loop; + goto .compile_global_initializer_loop elsif current_byte = '@' then (* Skip @. *) source_code_position := source_code_position + 1; @@ -1256,15 +1248,16 @@ begin _write_token(current_byte); _lexer_skip_token(); - goto .compile_global_initializer_end; + goto .compile_global_initializer_end elsif _is_digit(current_byte) = 1 then _write_z("\n\t.word \0"); current_byte := _lexer_read_token(@token_kind); _write_token(current_byte); source_code_position := source_code_position + 1; - goto .compile_global_initializer_end; + goto .compile_global_initializer_end end; + .compile_global_initializer_loop; _compile_global_initializer(); @@ -1272,16 +1265,14 @@ begin (* Skip comma and whitespace after it. *) source_code_position := source_code_position + 2; - goto .compile_global_initializer_loop; + goto .compile_global_initializer_loop end; .compile_global_initializer_closing; (* Skip ")" *) source_code_position := source_code_position + 1; - goto .compile_global_initializer_end; - - .compile_global_initializer_end; + .compile_global_initializer_end end; proc _compile_constant_declaration(); @@ -1304,7 +1295,7 @@ begin _compile_global_initializer(); (* Skip semicolon and newline. *) source_code_position := source_code_position + 2; - _write_c('\n'); + _write_c('\n') end; proc _compile_const_part(); @@ -1315,7 +1306,7 @@ begin _lexer_read_token(@token_kind); if token_kind <> _lexer_token_kind_const() then - goto .compile_const_part_end; + goto .compile_const_part_end end; (* Skip "const" with the newline after it. *) _lexer_skip_token(); @@ -1329,10 +1320,10 @@ begin if _load_byte(source_code_position) = '\t' then source_code_position := source_code_position + 1; _compile_constant_declaration(); - goto .compile_const_part_loop; + goto .compile_const_part_loop end; - .compile_const_part_end; + .compile_const_part_end end; proc _compile_variable_declaration(); @@ -1357,17 +1348,17 @@ begin if _load_byte(source_code_position) <> ' ' then (* Else we assume this is a zeroed 81920 bytes big array. *) - _write_z(" .zero 81920\0"); + _write_z(" .zero 81920\0") else (* Skip the assignment sign with surrounding whitespaces. *) source_code_position := source_code_position + 4; - _compile_global_initializer(); + _compile_global_initializer() end; (* Skip semicolon and newline. *) _lexer_read_token(@token_kind); _lexer_skip_token(); - _write_c('\n'); + _write_c('\n') end; proc _compile_var_part(); @@ -1377,7 +1368,7 @@ begin _lexer_read_token(@token_kind); if token_kind <> _lexer_token_kind_var() then - goto .compile_var_part_end; + goto .compile_var_part_end end; (* Skip "var" and newline. *) _lexer_skip_token(); @@ -1389,13 +1380,15 @@ begin if token_kind = _lexer_token_kind_identifier() then _compile_variable_declaration(); - goto .compile_var_part_loop; + goto .compile_var_part_loop end; - .compile_var_part_end; + .compile_var_part_end end; -(* Process the source code and print the generated code. *) +(** + * Process the source code and print the generated code. + *) proc _compile_module(); begin _compile_const_part(); @@ -1416,10 +1409,10 @@ begin (* 5 is "proc " length. Space is needed to distinguish from "procedure". *) if _memcmp(source_code_position, "proc ", 5) = 0 then _compile_procedure(); - goto .compile_module_loop; - end; + goto .compile_module_loop + end end; - .compile_module_end; + .compile_module_end end; proc _compile(); @@ -1443,18 +1436,21 @@ begin compiler_strings_copy := compiler_strings_copy + 1; _write_c(current_byte); - goto .compile_loop; + goto .compile_loop end; _write_c('"'); - _write_c('\n'); + _write_c('\n') end; (** - * Terminates the program. + * Terminates the program. a0 contains the return code. + * + * Parameters: + * a0 - Status code. *) proc _exit(); begin - _syscall(0, 0, 0, 0, 0, 0, 93); + _syscall(0, 0, 0, 0, 0, 0, 93) end; (** @@ -1484,7 +1480,7 @@ begin .symbol_table_lookup_loop; if symbol_table_length = 0 then - goto .symbol_table_lookup_end; + goto .symbol_table_lookup_end end; (* Symbol name pointer and length. *) @@ -1493,11 +1489,11 @@ begin (* If lengths don't match, exit and return nil. *) if name_length <> current_length then - goto .symbol_table_lookup_repeat; + 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; + goto .symbol_table_lookup_repeat end; (* Otherwise, the symbol is found. *) result := _load_word(symbol_table + 8); @@ -1542,7 +1538,7 @@ begin (* Increment the symbol table length. *) table_length := table_length + 1; - _store_word(table_length, symbol_table); + _store_word(table_length, symbol_table) end; proc _symbol_table_build(); @@ -1555,7 +1551,7 @@ begin _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_bool, 4, @symbol_type_info_bool); + _symbol_table_enter(@symbol_table_global, symbol_builtin_name_bool, 4, @symbol_type_info_bool) end; @@ -1564,247 +1560,199 @@ end; * characters of the same group a handled equivalently. * * Transition = record - * action: TransitionAction; - * next_state: TransitionState + * action: TransitionAction; + * next_state: TransitionState * end; *) proc _lexer_class_invalid(); -begin return 1 end; proc _lexer_class_digit(); -begin return 2 end; proc _lexer_class_alpha(); -begin return 3 end; proc _lexer_class_space(); -begin return 4 end; proc _lexer_class_colon(); -begin return 5 end; proc _lexer_class_equals(); -begin return 6 end; proc _lexer_class_left_paren(); -begin return 7 end; proc _lexer_class_right_paren(); -begin return 8 end; proc _lexer_class_asterisk(); -begin return 9 end; proc _lexer_class_underscore(); -begin return 10 end; proc _lexer_class_single(); -begin return 11 end; proc _lexer_class_hex(); -begin return 12 end; proc _lexer_class_zero(); -begin return 13 end; proc _lexer_class_x(); -begin return 14 end; proc _lexer_class_eof(); -begin return 15 end; proc _lexer_class_dot(); -begin return 16 end; proc _lexer_class_minus(); -begin return 17 end; proc _lexer_class_single_quote(); -begin return 18 end; proc _lexer_class_double_quote(); -begin return 19 end; proc _lexer_class_greater(); -begin return 20 end; proc _lexer_class_less(); -begin return 21 end; proc _lexer_class_other(); -begin return 22 end; proc _lexer_state_start(); -begin return 1 end; proc _lexer_state_colon(); -begin return 2 end; proc _lexer_state_identifier(); -begin return 3 end; proc _lexer_state_decimal(); -begin return 4 end; proc _lexer_state_greater(); -begin return 5 end; proc _lexer_state_minus(); -begin return 6 end; proc _lexer_state_left_paren(); -begin return 7 end; proc _lexer_state_less(); -begin return 8 end; proc _lexer_state_dot(); -begin return 9 end; proc _lexer_state_comment(); -begin return 10 end; proc _lexer_state_closing_comment(); -begin return 11 end; proc _lexer_state_character(); -begin return 12 end; proc _lexer_state_string(); -begin return 13 end; proc _lexer_state_leading_zero(); -begin return 14 end; proc _lexer_state_decimal_suffix(); -begin return 15 end; proc _lexer_state_end(); -begin return 16 end; proc _lexer_action_none(); -begin return 1 end; proc _lexer_action_accumulate(); -begin return 2 end; proc _lexer_action_skip(); -begin return 3 end; proc _lexer_action_single(); -begin return 4 end; proc _lexer_action_eof(); -begin return 5 end; proc _lexer_action_finalize(); -begin return 6 end; proc _lexer_action_composite(); -begin return 7 end; proc _lexer_action_key_id(); -begin return 8 end; proc _lexer_action_integer(); -begin return 9 end; proc _lexer_action_delimited(); -begin return 10 end; @@ -1824,7 +1772,7 @@ begin target := target * 4; target := array + target; - _store_word(data, target); + _store_word(data, target) end; proc _get_at(array: Word, index: Word); @@ -1838,7 +1786,9 @@ begin return _load_word(target) end; -(* Initializes the array with character classes. *) +(** + * Initializes the array with character classes. + *) proc _lexer_classifications(); var code: Word; @@ -1980,8 +1930,8 @@ begin code := code + 1; if code < 257 then - goto .create_classification_loop; - end; + goto .create_classification_loop + end end; proc _lexer_get_transition(current_state: Word, character_class: Word); @@ -2018,12 +1968,12 @@ begin transition := _lexer_get_transition(current_state, character_class); _lexer_transition_set_action(transition, action); - _lexer_transition_set_state(transition, next_state); + _lexer_transition_set_state(transition, next_state) end; +(* Sets same action and state transition for all character classes in one transition row. *) + (** - * 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). @@ -2052,10 +2002,9 @@ begin _lexer_set_transition(current_state, _lexer_class_double_quote(), default_action, next_state); _lexer_set_transition(current_state, _lexer_class_greater(), default_action, next_state); _lexer_set_transition(current_state, _lexer_class_less(), default_action, next_state); - _lexer_set_transition(current_state, _lexer_class_other(), default_action, next_state); + _lexer_set_transition(current_state, _lexer_class_other(), default_action, next_state) end; - (** * The transition table describes transitions from one state to another, given * a symbol (character class). @@ -2172,7 +2121,7 @@ begin _lexer_set_transition(_lexer_state_decimal_suffix(), _lexer_class_alpha(), _lexer_action_none(), _lexer_state_end()); _lexer_set_transition(_lexer_state_decimal_suffix(), _lexer_class_hex(), _lexer_action_none(), _lexer_state_end()); _lexer_set_transition(_lexer_state_decimal_suffix(), _lexer_class_zero(), _lexer_action_none(), _lexer_state_end()); - _lexer_set_transition(_lexer_state_decimal_suffix(), _lexer_class_x(), _lexer_action_none(), _lexer_state_end()); + _lexer_set_transition(_lexer_state_decimal_suffix(), _lexer_class_x(), _lexer_action_none(), _lexer_state_end()) end; (** @@ -2181,7 +2130,6 @@ end; * 1024 = 256 * 4 *) proc _lexer_get_transition_table(); -begin return @classification + 1024 end; @@ -2191,7 +2139,6 @@ end; * and 22 columns (character classes), so 2816 = 8 * 16 * 22. *) proc _lexer_global_state(); -begin return _lexer_get_transition_table() + 2816 end; @@ -2199,7 +2146,6 @@ end; * Gets pointer to the token start. *) proc _lexer_global_start(); -begin return _lexer_global_state() + 4 end; @@ -2207,28 +2153,25 @@ end; * Gets pointer to the token end. *) proc _lexer_global_end(); -begin return _lexer_global_start() + 4 end; proc _lexer_transition_get_action(transition: Word); -begin return _load_word(transition) end; proc _lexer_transition_set_action(transition: Word, action: Word); begin - _store_word(action, transition); + _store_word(action, transition) end; proc _lexer_transition_get_state(transition: Word); -begin return _load_word(transition + 4) end; proc _lexer_transition_set_state(transition: Word, state: Word); begin - _store_word(state, transition + 4); + _store_word(state, transition + 4) end; (** @@ -2247,7 +2190,7 @@ begin _store_word(source_code_position, current); current := _lexer_global_end(); - _store_word(source_code_position, current); + _store_word(source_code_position, current) end; (** @@ -2256,7 +2199,7 @@ end; proc _lexer_initialize(); begin _lexer_classifications(); - _lexer_transitions(); + _lexer_transitions() end; proc _lexer_next_transition(); @@ -2278,322 +2221,258 @@ begin end; proc _lexer_token_kind_identifier(); -begin return 1 end; proc _lexer_token_kind_const(); -begin return 2 end; proc _lexer_token_kind_var(); -begin return 3 end; proc _lexer_token_kind_proc(); -begin return 4 end; proc _lexer_token_kind_type(); -begin return 5 end; proc _lexer_token_kind_begin(); -begin return 6 end; proc _lexer_token_kind_end(); -begin return 7 end; proc _lexer_token_kind_if(); -begin return 8 end; proc _lexer_token_kind_then(); -begin return 9 end; proc _lexer_token_kind_else(); -begin return 10 end; proc _lexer_token_kind_elsif(); -begin return 11 end; proc _lexer_token_kind_while(); -begin return 12 end; proc _lexer_token_kind_do(); -begin return 13 end; proc _lexer_token_kind_extern(); -begin return 14 end; proc _lexer_token_kind_record(); -begin return 15 end; proc _lexer_token_kind_union(); -begin return 16 end; proc _lexer_token_kind_true(); -begin return 17 end; proc _lexer_token_kind_false(); -begin return 18 end; proc _lexer_token_kind_nil(); -begin return 19 end; proc _lexer_token_kind_and(); -begin return 20 end; proc _lexer_token_kind_or(); -begin return 21 end; proc _lexer_token_kind_xor(); -begin return 22 end; proc _lexer_token_kind_pipe(); -begin return 23 end; proc _lexer_token_kind_not(); -begin return 24 end; proc _lexer_token_kind_return(); -begin return 24 end; proc _lexer_token_kind_module(); -begin return 25 end; proc _lexer_token_kind_program(); -begin return 26 end; proc _lexer_token_kind_import(); -begin return 27 end; proc _lexer_token_kind_cast(); -begin return 28 end; proc _lexer_token_kind_defer(); -begin return 29 end; proc _lexer_token_kind_case(); -begin return 30 end; proc _lexer_token_kind_of(); -begin return 31 end; proc _lexer_token_kind_trait(); -begin return 32 end; proc _lexer_token_kind_left_paren(); -begin return 33 end; proc _lexer_token_kind_right_paren(); -begin return 34 end; proc _lexer_token_kind_left_square(); -begin return 35 end; proc _lexer_token_kind_right_square(); -begin return 36 end; proc _lexer_token_kind_shift_left(); -begin return 37 end; proc _lexer_token_kind_shift_right(); -begin return 38 end; proc _lexer_token_kind_greater_equal(); -begin return 39 end; proc _lexer_token_kind_less_equal(); -begin return 40 end; proc _lexer_token_kind_greater_than(); -begin return 41 end; proc _lexer_token_kind_less_than(); -begin return 42 end; proc _lexer_token_kind_not_equal(); -begin return 43 end; proc _lexer_token_kind_equals(); -begin return 44 end; proc _lexer_token_kind_semicolon(); -begin return 45 end; proc _lexer_token_kind_dot(); -begin return 46 end; proc _lexer_token_kind_comma(); -begin return 47 end; proc _lexer_token_kind_plus(); -begin return 48 end; proc _lexer_token_kind_arrow(); -begin return 49 end; proc _lexer_token_kind_minus(); -begin return 50 end; proc _lexer_token_kind_multiplication(); -begin return 51 end; proc _lexer_token_kind_division(); -begin return 52 end; proc _lexer_token_kind_remainder(); -begin return 53 end; proc _lexer_token_kind_assignment(); -begin return 54 end; proc _lexer_token_kind_colon(); -begin return 55 end; proc _lexer_token_kind_hat(); -begin return 56 end; proc _lexer_token_kind_at(); -begin return 57 end; proc _lexer_token_kind_exclamation(); -begin return 58 end; proc _lexer_token_kind_string(); -begin return 59 end; proc _lexer_token_kind_character(); -begin return 60 end; proc _lexer_token_kind_integer(); -begin return 61 end; proc _lexer_token_kind_word(); -begin return 62 end; proc _lexer_token_kind_goto(); -begin return 63 end; @@ -2604,7 +2483,7 @@ begin result := 0; if lhs_length = rhs_length then - result := _memcmp(lhs_pointer, rhs_pointer, lhs_length) = 0; + result := _memcmp(lhs_pointer, rhs_pointer, lhs_length) = 0 end; return result end; @@ -2618,31 +2497,31 @@ begin token_length := position_end + -position_start; if _lexer_compare_keyword(position_start, token_length, "const", 5) = 1 then - result := _lexer_token_kind_const(); + result := _lexer_token_kind_const() elsif _lexer_compare_keyword(position_start, token_length, "var", 3) = 1 then - result := _lexer_token_kind_var(); + result := _lexer_token_kind_var() elsif _lexer_compare_keyword(position_start, token_length, "proc", 4) = 1 then - result := _lexer_token_kind_proc(); + result := _lexer_token_kind_proc() elsif _lexer_compare_keyword(position_start, token_length, "type", 4) = 1 then - result := _lexer_token_kind_type(); + result := _lexer_token_kind_type() elsif _lexer_compare_keyword(position_start, token_length, "begin", 5) = 1 then - result := _lexer_token_kind_begin(); + result := _lexer_token_kind_begin() elsif _lexer_compare_keyword(position_start, token_length, "end", 3) = 1 then - result := _lexer_token_kind_end(); + result := _lexer_token_kind_end() elsif _lexer_compare_keyword(position_start, token_length, "return", 6) = 1 then - result := _lexer_token_kind_return(); + result := _lexer_token_kind_return() elsif _lexer_compare_keyword(position_start, token_length, "goto", 4) = 1 then - result := _lexer_token_kind_goto(); + result := _lexer_token_kind_goto() elsif _lexer_compare_keyword(position_start, token_length, "if", 2) = 1 then - result := _lexer_token_kind_if(); + result := _lexer_token_kind_if() elsif _lexer_compare_keyword(position_start, token_length, "while", 5) = 1 then - result := _lexer_token_kind_while(); + result := _lexer_token_kind_while() elsif _lexer_compare_keyword(position_start, token_length, "then", 4) = 1 then - result := _lexer_token_kind_then(); + result := _lexer_token_kind_then() elsif _lexer_compare_keyword(position_start, token_length, "else", 4) = 1 then - result := _lexer_token_kind_else(); + result := _lexer_token_kind_else() elsif _lexer_compare_keyword(position_start, token_length, "elsif", 5) = 1 then - result := _lexer_token_kind_elsif(); + result := _lexer_token_kind_elsif() end; return result end; @@ -2656,9 +2535,9 @@ begin character := _load_byte(start_position); if character = ':' then - result := _lexer_token_kind_colon(); + result := _lexer_token_kind_colon() elsif character = '.' then - result := _lexer_token_kind_dot(); + result := _lexer_token_kind_dot() end; return result end; @@ -2672,7 +2551,7 @@ begin character := _load_byte(start_position); if character = ';' then - result := _lexer_token_kind_semicolon(); + result := _lexer_token_kind_semicolon() end; return result end; @@ -2692,23 +2571,23 @@ begin if action_to_perform = _lexer_action_none() then elsif action_to_perform = _lexer_action_accumulate() then - _store_word(position_end + 1, pointer_end); + _store_word(position_end + 1, pointer_end) elsif action_to_perform = _lexer_action_skip() then _store_word(position_start + 1, pointer_start); - _store_word(position_end + 1, pointer_end); + _store_word(position_end + 1, pointer_end) elsif action_to_perform = _lexer_action_single() then _store_word(position_end + 1, pointer_end); intermediate := _lexer_classify_single(position_start); - _store_word(intermediate, kind); + _store_word(intermediate, kind) elsif action_to_perform = _lexer_action_eof() then elsif action_to_perform = _lexer_action_finalize() then intermediate := _lexer_classify_finalize(position_start); - _store_word(intermediate, kind); + _store_word(intermediate, kind) elsif action_to_perform = _lexer_action_composite() then elsif action_to_perform = _lexer_action_key_id() then intermediate := _lexer_classify_keyword(position_start, position_end); - _store_word(intermediate, kind); + _store_word(intermediate, kind) elsif action_to_perform = _lexer_action_integer() then elsif action_to_perform = _lexer_action_delimited() then end; @@ -2736,8 +2615,8 @@ end; proc _lexer_advance_token(kind: Word); begin if _lexer_execute_transition(kind) <> _lexer_state_end() then - _lexer_advance_token(kind); - end; + _lexer_advance_token(kind) + end end; (** @@ -2764,10 +2643,10 @@ var new_position: Word; begin new_position := _lexer_global_end(); - source_code_position := _load_word(new_position); + source_code_position := _load_word(new_position) end; -(** +(* * Entry point. *) proc _start(); @@ -2786,9 +2665,9 @@ begin last_read := _read_file(offset, 81920); if last_read > 0 then offset := offset + last_read; - goto .start_read; + goto .start_read end; _compile(); - _exit(0); + _exit(0) end;