belka/elna
1
0
Fork 0
Code Issues 34 Pull Requests Releases Activity

Implement .max and .min type properties

Browse Source
This commit is contained in:
Eugen Wissner
2025-02-16 07:53:31 +01:00
parent b358f8ba27
commit 031587e6f9
5 changed files with 259 additions and 262 deletions
Download Patch File Download Diff File Expand all files Collapse all files

12
boot/lexer.ll
View File

@@ -89,12 +89,6 @@ const {
var { var {
return yy::parser::make_VAR(this->location); return yy::parser::make_VAR(this->location);
} }
array {
return yy::parser::make_ARRAY(this->location);
}
of {
return yy::parser::make_OF(this->location);
}
type { type {
return yy::parser::make_TYPE(this->location); return yy::parser::make_TYPE(this->location);
} }
@@ -104,12 +98,6 @@ record {
union { union {
return yy::parser::make_UNION(this->location); return yy::parser::make_UNION(this->location);
} }
pointer {
return yy::parser::make_POINTER(this->location);
}
to {
return yy::parser::make_TO(this->location);
}
true { true {
return yy::parser::make_BOOLEAN(true, this->location); return yy::parser::make_BOOLEAN(true, this->location);
} }

25
boot/parser.yy
View File

@@ -82,7 +82,7 @@ along with GCC; see the file COPYING3. If not see
%token <std::string> STRING "string" %token <std::string> STRING "string"
%token <bool> BOOLEAN %token <bool> BOOLEAN
%token IF WHILE DO THEN ELSE ELSIF RETURN %token IF WHILE DO THEN ELSE ELSIF RETURN
%token CONST VAR PROCEDURE ARRAY OF TYPE RECORD POINTER TO UNION %token CONST VAR PROCEDURE TYPE RECORD UNION
%token BEGIN_BLOCK END_BLOCK EXTERN DEFER %token BEGIN_BLOCK END_BLOCK EXTERN DEFER
%token LEFT_PAREN RIGHT_PAREN LEFT_SQUARE RIGHT_SQUARE SEMICOLON DOT COMMA %token LEFT_PAREN RIGHT_PAREN LEFT_SQUARE RIGHT_SQUARE SEMICOLON DOT COMMA
%token AND OR NOT CAST SHIFT_LEFT SHIFT_RIGHT %token AND OR NOT CAST SHIFT_LEFT SHIFT_RIGHT
@@ -118,7 +118,7 @@ along with GCC; see the file COPYING3. If not see
%type <std::vector<elna::boot::type_definition *>> type_definitions type_part; %type <std::vector<elna::boot::type_definition *>> type_definitions type_part;
%type <elna::boot::block *> block; %type <elna::boot::block *> block;
%type <elna::boot::field_t> field_declaration; %type <elna::boot::field_t> field_declaration;
%type <std::vector<std::pair<std::string, std::shared_ptr<elna::boot::top_type>>>> field_list; %type <std::vector<std::pair<std::string, std::shared_ptr<elna::boot::top_type>>>> optional_fields fields;
%type <std::vector<elna::boot::conditional_statements *>> elsif_statement_list; %type <std::vector<elna::boot::conditional_statements *>> elsif_statement_list;
%type <elna::boot::cast_expression *> cast_expression; %type <elna::boot::cast_expression *> cast_expression;
%type <elna::boot::defer_statement *> defer_statement; %type <elna::boot::defer_statement *> defer_statement;
@@ -414,9 +414,9 @@ statement:
} }
| defer_statement { $$ = $1; } | defer_statement { $$ = $1; }
statements: statements:
statement SEMICOLON statements statement statements
{ {
std::swap($$, $3); std::swap($$, $2);
$$.emplace($$.cbegin(), $1); $$.emplace($$.cbegin(), $1);
} }
| statement { $$.push_back($1); } | statement { $$.push_back($1); }
@@ -425,27 +425,30 @@ optional_statements:
| /* no statements */ {} | /* no statements */ {}
field_declaration: field_declaration:
IDENTIFIER COLON type_expression { $$ = std::make_pair($1, $3); } IDENTIFIER COLON type_expression { $$ = std::make_pair($1, $3); }
field_list: fields:
field_declaration field_list field_declaration fields
{ {
std::swap($$, $2); std::swap($$, $2);
$$.emplace($$.cbegin(), $1); $$.emplace($$.cbegin(), $1);
} }
| field_declaration { $$.emplace_back($1); } | field_declaration { $$.emplace_back($1); }
optional_fields:
fields { std::swap($$, $1); }
| /* no fields */ {}
type_expression: type_expression:
ARRAY INTEGER OF type_expression LEFT_SQUARE INTEGER RIGHT_SQUARE type_expression
{ {
$$ = std::make_shared<elna::boot::array_type>(elna::boot::make_position(@1), $4, $2); $$ = std::make_shared<elna::boot::array_type>(elna::boot::make_position(@1), $4, $2);
} }
| POINTER TO type_expression | HAT type_expression
{ {
$$ = std::make_shared<elna::boot::pointer_type>(elna::boot::make_position(@1), $3); $$ = std::make_shared<elna::boot::pointer_type>(elna::boot::make_position(@1), $2);
} }
| RECORD field_list END_BLOCK | RECORD optional_fields END_BLOCK
{ {
$$ = std::make_shared<elna::boot::record_type>(elna::boot::make_position(@1), std::move($2)); $$ = std::make_shared<elna::boot::record_type>(elna::boot::make_position(@1), std::move($2));
} }
| UNION field_list END_BLOCK | UNION fields END_BLOCK
{ {
$$ = std::make_shared<elna::boot::union_type>(elna::boot::make_position(@1), std::move($2)); $$ = std::make_shared<elna::boot::union_type>(elna::boot::make_position(@1), std::move($2));
} }

2
gcc/elna-diagnostic.cc
View File

@@ -68,7 +68,7 @@ namespace gcc
} }
else if (is_pointer_type(type)) else if (is_pointer_type(type))
{ {
return std::string("pointer to " + print_type(TREE_TYPE(type))); return std::string("^" + print_type(TREE_TYPE(type)));
} }
else if (is_procedure_type(type)) else if (is_procedure_type(type))
{ {

8
gcc/elna-generic.cc
View File

@@ -952,6 +952,14 @@ namespace gcc
this->current_expression = build_int_cstu(elna_word_type_node, this->current_expression = build_int_cstu(elna_word_type_node,
TYPE_ALIGN_UNIT(this->current_expression)); TYPE_ALIGN_UNIT(this->current_expression));
} }
else if (expression->field() == "min" && is_integral_type(this->current_expression))
{
this->current_expression = TYPE_MIN_VALUE(this->current_expression);
}
else if (expression->field() == "max" && is_integral_type(this->current_expression))
{
this->current_expression = TYPE_MAX_VALUE(this->current_expression);
}
else else
{ {
error_at(expression_location, "type '%s' does not have property '%s'", error_at(expression_location, "type '%s' does not have property '%s'",

474
source.elna
View File

@@ -76,7 +76,7 @@ type
end end
TokenValue* = union TokenValue* = union
int_value: Int int_value: Int
string_value: pointer to Char string_value: ^Char
string: String string: String
boolean_value: Bool boolean_value: Bool
char_value: Char char_value: Char
@@ -86,11 +86,9 @@ type
value: TokenValue value: TokenValue
location: Location location: Location
end end
FILE* = record FILE* = record end
dummy: Int
end
CommandLine* = record CommandLine* = record
input: pointer to Char input: ^Char
tokenize: Bool tokenize: Bool
syntax_tree: Bool syntax_tree: Bool
end end
@@ -98,11 +96,11 @@ type
value: Int value: Int
end end
ConstantDefinition* = record ConstantDefinition* = record
name: pointer to Char name: ^Char
body: pointer to Literal body: ^Literal
end end
ConstantPart* = record ConstantPart* = record
elements: pointer to pointer to ConstantDefinition elements: ^^ConstantDefinition
count: Word count: Word
end end
Program* = record Program* = record
@@ -112,48 +110,48 @@ type
(* (*
External procedures. External procedures.
*) *)
proc fopen(pathname: pointer to Char, mode: pointer to Char) -> pointer to FILE; extern proc fopen(pathname: ^Char, mode: ^Char) -> ^FILE; extern
proc fclose(stream: pointer to FILE) -> Int; extern proc fclose(stream: ^FILE) -> Int; extern
proc fseek(stream: pointer to FILE, off: Int, whence: Int) -> Int; extern proc fseek(stream: ^FILE, off: Int, whence: Int) -> Int; extern
proc rewind(stream: pointer to FILE); extern proc rewind(stream: ^FILE); extern
proc ftell(stream: pointer to FILE) -> Int; extern proc ftell(stream: ^FILE) -> Int; extern
proc fread(ptr: pointer to Byte, size: Word, nmemb: Word, stream: pointer to FILE) -> Word; extern proc fread(ptr: ^Byte, size: Word, nmemb: Word, stream: ^FILE) -> Word; extern
proc write(fd: Int, buf: pointer to Byte, Word: Int) -> Int; extern proc write(fd: Int, buf: ^Byte, Word: Int) -> Int; extern
proc malloc(size: Word) -> pointer to Byte; extern proc malloc(size: Word) -> ^Byte; extern
proc free(ptr: pointer to Byte); extern proc free(ptr: ^Byte); extern
proc calloc(nmemb: Word, size: Word) -> pointer to Byte; extern proc calloc(nmemb: Word, size: Word) -> ^Byte; extern
proc realloc(ptr: pointer to Byte, size: Word) -> pointer to Byte; extern proc realloc(ptr: ^Byte, size: Word) -> ^Byte; extern
proc memset(ptr: pointer to Char, c: Int, n: Int) -> pointer to Char; extern proc memset(ptr: ^Char, c: Int, n: Int) -> ^Char; extern
proc strcmp(s1: pointer to Char, s2: pointer to Char) -> Int; extern proc strcmp(s1: ^Char, s2: ^Char) -> Int; extern
proc strncmp(s1: pointer to Char, s2: pointer to Char, n: Word) -> Int; extern proc strncmp(s1: ^Char, s2: ^Char, n: Word) -> Int; extern
proc strncpy(dst: pointer to Char, src: pointer to Char, dsize: Word) -> pointer to Char; extern proc strncpy(dst: ^Char, src: ^Char, dsize: Word) -> ^Char; extern
proc strcpy(dst: pointer to Char, src: pointer to Char) -> pointer to Char; extern proc strcpy(dst: ^Char, src: ^Char) -> ^Char; extern
proc strlen(ptr: pointer to Char) -> Word; extern proc strlen(ptr: ^Char) -> Word; extern
proc strtol(nptr: pointer to Char, endptr: pointer to pointer to Char, base: Int) -> Int; extern proc strtol(nptr: ^Char, endptr: ^^Char, base: Int) -> Int; extern
proc perror(s: pointer to Char); extern proc perror(s: ^Char); extern
proc exit(code: Int); extern proc exit(code: Int); extern
(* (*
Standard procedures. Standard procedures.
*) *)
proc reallocarray(ptr: pointer to Byte, n: Word, size: Word) -> pointer to Byte; proc reallocarray(ptr: ^Byte, n: Word, size: Word) -> ^Byte;
begin begin
return realloc(ptr, n * size) return realloc(ptr, n * size)
end end
proc write_s(value: String); proc write_s(value: String);
begin begin
write(0, cast(value.ptr: pointer to Byte), cast(value.length: Int)) write(0, cast(value.ptr: ^Byte), cast(value.length: Int))
end end
proc write_z(value: pointer to Char); proc write_z(value: ^Char);
begin begin
write(0, cast(value: pointer to Byte), cast(strlen(value): Int)) write(0, cast(value: ^Byte), cast(strlen(value): Int))
end end
proc write_b(value: Bool); proc write_b(value: Bool);
@@ -167,29 +165,29 @@ end
proc write_c(value: Char); proc write_c(value: Char);
begin begin
write(0, cast(@value: pointer to Byte), 1) write(0, cast(@value: ^Byte), 1)
end end
proc write_i(value: Int); proc write_i(value: Int);
var var
digit: Int digit: Int
n: Word n: Word
buffer: array 10 of Char buffer: [10]Char
begin begin
n := 10u; n := 10u
if value = 0 then if value = 0 then
write_c('0') write_c('0')
end; end
while value <> 0 do while value <> 0 do
digit := value % 10; digit := value % 10
value := value / 10; value := value / 10
buffer[n] := cast(cast('0': Int) + digit: Char); buffer[n] := cast(cast('0': Int) + digit: Char)
n := n - 1u n := n - 1u
end; end
while n < 10u do while n < 10u do
n := n + 1u; n := n + 1u
write_c(buffer[n]) write_c(buffer[n])
end end
end end
@@ -231,10 +229,10 @@ end
proc string_dup(origin: String) -> String; proc string_dup(origin: String) -> String;
var var
copy: pointer to Char copy: ^Char
begin begin
copy := cast(malloc(origin.length): pointer to Char); copy := cast(malloc(origin.length): ^Char)
strncpy(copy, origin.ptr, origin.length); strncpy(copy, origin.ptr, origin.length)
return String(copy, origin.length) return String(copy, origin.length)
end end
@@ -248,75 +246,75 @@ begin
return Position(1u, 1u) return Position(1u, 1u)
end end
proc read_source(filename: pointer to Char, result: pointer to String) -> Bool; proc read_source(filename: ^Char, result: ^String) -> Bool;
var var
input_file: pointer to FILE input_file: ^FILE
source_size: Int source_size: Int
input: pointer to Byte input: ^Byte
begin begin
input_file := fopen(filename, "rb\0".ptr); input_file := fopen(filename, "rb\0".ptr)
if input_file = nil then if input_file = nil then
return false return false
end; end
defer defer
fclose(input_file) fclose(input_file)
end; end
if fseek(input_file, 0, SEEK_END) <> 0 then if fseek(input_file, 0, SEEK_END) <> 0 then
return false return false
end; end
source_size := ftell(input_file); source_size := ftell(input_file)
if source_size < 0 then if source_size < 0 then
return false return false
end; end
rewind(input_file); rewind(input_file)
input := malloc(cast(source_size: Word)); input := malloc(cast(source_size: Word))
if fread(input, cast(source_size: Word), 1u, input_file) <> 1u then if fread(input, cast(source_size: Word), 1u, input_file) <> 1u then
return false return false
end; end
result^ := String(cast(input: pointer to Char), cast(source_size: Word)); result^ := String(cast(input: ^Char), cast(source_size: Word))
return true return true
end end
proc escape_char(escape: Char, result: pointer to Char) -> Bool; proc escape_char(escape: Char, result: ^Char) -> Bool;
begin begin
if escape = 'n' then if escape = 'n' then
result^ := '\n'; result^ := '\n'
return true return true
elsif escape = 'a' then elsif escape = 'a' then
result^ := '\a'; result^ := '\a'
return true return true
elsif escape = 'b' then elsif escape = 'b' then
result^ := '\b'; result^ := '\b'
return true return true
elsif escape = 't' then elsif escape = 't' then
result^ := '\t'; result^ := '\t'
return true return true
elsif escape = 'f' then elsif escape = 'f' then
result^ := '\f'; result^ := '\f'
return true return true
elsif escape = 'r' then elsif escape = 'r' then
result^ := '\r'; result^ := '\r'
return true return true
elsif escape = 'v' then elsif escape = 'v' then
result^ := '\v'; result^ := '\v'
return true return true
elsif escape = '\\' then elsif escape = '\\' then
result^ := '\\'; result^ := '\\'
return true return true
elsif escape = '\'' then elsif escape = '\'' then
result^ := '\''; result^ := '\''
return true return true
elsif escape = '"' then elsif escape = '"' then
result^ := '"'; result^ := '"'
return true return true
elsif escape = '?' then elsif escape = '?' then
result^ := '\?'; result^ := '\?'
return true return true
elsif escape = '0' then elsif escape = '0' then
result^ := '\0'; result^ := '\0'
return true return true
else else
return false return false
@@ -325,8 +323,8 @@ end
proc advance_source(source_code: SourceCode, length: Word) -> SourceCode; proc advance_source(source_code: SourceCode, length: Word) -> SourceCode;
begin begin
source_code.text := open_substring(source_code.text, length); source_code.text := open_substring(source_code.text, length)
source_code.position.column := source_code.position.column + length; source_code.position.column := source_code.position.column + length
return source_code return source_code
end end
@@ -335,112 +333,112 @@ proc skip_spaces(source_code: SourceCode) -> SourceCode;
begin begin
while source_code.text.length > 0u and is_space(source_code.text[1u]) do while source_code.text.length > 0u and is_space(source_code.text[1u]) do
if source_code.text[1u] = '\n' then if source_code.text[1u] = '\n' then
source_code.position.line := source_code.position.line + 1u; source_code.position.line := source_code.position.line + 1u
source_code.position.column := 1u source_code.position.column := 1u
else else
source_code.position.column := source_code.position.column + 1u source_code.position.column := source_code.position.column + 1u
end; end
source_code.text := open_substring(source_code.text, 1u) source_code.text := open_substring(source_code.text, 1u)
end; end
return source_code return source_code
end end
proc lex_identifier(source_code: pointer to SourceCode, token_content: pointer to String); proc lex_identifier(source_code: ^SourceCode, token_content: ^String);
var var
content_length: Word content_length: Word
begin begin
content_length := 0u; content_length := 0u
token_content^ := source_code^.text; token_content^ := source_code^.text
while is_alnum(source_code^.text[1u]) or source_code^.text[1u] = '_' do while is_alnum(source_code^.text[1u]) or source_code^.text[1u] = '_' do
content_length := content_length + 1u; content_length := content_length + 1u
source_code^ := advance_source(source_code^, 1u) source_code^ := advance_source(source_code^, 1u)
end; end
token_content^ := substring(token_content^, 0u, content_length) token_content^ := substring(token_content^, 0u, content_length)
end end
proc lex_comment(source_code: pointer to SourceCode, token_content: pointer to String) -> Bool; proc lex_comment(source_code: ^SourceCode, token_content: ^String) -> Bool;
var var
content_length: Word content_length: Word
begin begin
content_length := 0u; content_length := 0u
token_content^ := source_code^.text; token_content^ := source_code^.text
while source_code^.text.length > 1u do while source_code^.text.length > 1u do
if source_code^.text[1u] = '*' and source_code^.text[2u] = ')' then if source_code^.text[1u] = '*' and source_code^.text[2u] = ')' then
source_code^ := advance_source(source_code^, 2u); source_code^ := advance_source(source_code^, 2u)
token_content^ := substring(token_content^, 0u, content_length); token_content^ := substring(token_content^, 0u, content_length)
return true return true
end; end
content_length := content_length + 1u; content_length := content_length + 1u
source_code^ := advance_source(source_code^, 1u) source_code^ := advance_source(source_code^, 1u)
end; end
return false return false
end end
proc lex_character(input: pointer to Char, current_token: pointer to Token) -> pointer to Char; proc lex_character(input: ^Char, current_token: ^Token) -> ^Char;
begin begin
if input^ = '\\' then if input^ = '\\' then
input := input + 1; input := input + 1
if escape_char(input^, @current_token^.value.char_value) then if escape_char(input^, @current_token^.value.char_value) then
input := input + 1 input := input + 1
end end
elsif input^ <> '\0' then elsif input^ <> '\0' then
current_token^.value.char_value := input^; current_token^.value.char_value := input^
input := input + 1 input := input + 1
end; end
return input return input
end end
proc lex_string(input: pointer to Char, current_token: pointer to Token) -> pointer to Char; proc lex_string(input: ^Char, current_token: ^Token) -> ^Char;
var var
token_end, constructed_string: pointer to Char token_end, constructed_string: ^Char
token_length: Word token_length: Word
is_valid: Bool is_valid: Bool
begin begin
token_end := input; token_end := input
while token_end^ <> '\0' and not ((token_end - 1)^ <> '\\' and token_end^ = '"') do while token_end^ <> '\0' and not ((token_end - 1)^ <> '\\' and token_end^ = '"') do
token_end := token_end + 1 token_end := token_end + 1
end; end
if token_end^ <> '\"' then if token_end^ <> '\"' then
return input return input
end; end
token_length := cast(token_end - input: Word); token_length := cast(token_end - input: Word)
current_token^.value.string_value := cast(calloc(token_length, 1u): pointer to Char); current_token^.value.string_value := cast(calloc(token_length, 1u): ^Char)
is_valid := true; is_valid := true
constructed_string := current_token^.value.string_value; constructed_string := current_token^.value.string_value
while input < token_end and is_valid do while input < token_end and is_valid do
if input^ = '\\' then if input^ = '\\' then
input := input + 1; input := input + 1
if escape_char(input^, constructed_string) then if escape_char(input^, constructed_string) then
input := input + 1 input := input + 1
else else
is_valid := false is_valid := false
end end
elsif input^ <> '\0' then elsif input^ <> '\0' then
constructed_string^ := input^; constructed_string^ := input^
input := input + 1 input := input + 1
end; end
constructed_string := constructed_string + 1 constructed_string := constructed_string + 1
end; end
return token_end return token_end
end end
proc print_tokens(tokens: pointer to Token, tokens_size: Word); proc print_tokens(tokens: ^Token, tokens_size: Word);
var var
current_token: pointer to Token current_token: ^Token
i: Word i: Word
begin begin
i := 0u; i := 0u
while i < tokens_size do while i < tokens_size do
current_token := tokens + i; current_token := tokens + i
if current_token^.kind = TOKEN_IF then if current_token^.kind = TOKEN_IF then
write_s("IF") write_s("IF")
@@ -481,8 +479,8 @@ begin
elsif current_token^.kind = TOKEN_TO then elsif current_token^.kind = TOKEN_TO then
write_s("TO") write_s("TO")
elsif current_token^.kind = TOKEN_BOOLEAN then elsif current_token^.kind = TOKEN_BOOLEAN then
write_s("BOOLEAN<"); write_s("BOOLEAN<")
write_b(current_token^.value.boolean_value); write_b(current_token^.value.boolean_value)
write_c('>') write_c('>')
elsif current_token^.kind = TOKEN_NIL then elsif current_token^.kind = TOKEN_NIL then
write_s("NIL") write_s("NIL")
@@ -501,8 +499,8 @@ begin
elsif current_token^.kind = TOKEN_SHIFT_RIGHT then elsif current_token^.kind = TOKEN_SHIFT_RIGHT then
write_s(">>") write_s(">>")
elsif current_token^.kind = TOKEN_IDENTIFIER then elsif current_token^.kind = TOKEN_IDENTIFIER then
write_c('<'); write_c('<')
write_s(current_token^.value.string); write_s(current_token^.value.string)
write_c('>') write_c('>')
elsif current_token^.kind = TOKEN_LEFT_PAREN then elsif current_token^.kind = TOKEN_LEFT_PAREN then
write_s("(") write_s("(")
@@ -551,30 +549,30 @@ begin
elsif current_token^.kind = TOKEN_COMMENT then elsif current_token^.kind = TOKEN_COMMENT then
write_s("(* COMMENT *)") write_s("(* COMMENT *)")
elsif current_token^.kind = TOKEN_INTEGER then elsif current_token^.kind = TOKEN_INTEGER then
write_c('<'); write_c('<')
write_i(current_token^.value.int_value); write_i(current_token^.value.int_value)
write_c('>') write_c('>')
elsif current_token^.kind = TOKEN_WORD then elsif current_token^.kind = TOKEN_WORD then
write_c('<'); write_c('<')
write_i(current_token^.value.int_value); write_i(current_token^.value.int_value)
write_s("u>") write_s("u>")
elsif current_token^.kind = TOKEN_CHARACTER then elsif current_token^.kind = TOKEN_CHARACTER then
write_c('<'); write_c('<')
write_i(cast(current_token^.value.char_value: Int)); write_i(cast(current_token^.value.char_value: Int))
write_s("c>") write_s("c>")
elsif current_token^.kind = TOKEN_STRING then elsif current_token^.kind = TOKEN_STRING then
write_s("\"...\"") write_s("\"...\"")
elsif current_token^.kind = TOKEN_DEFER then elsif current_token^.kind = TOKEN_DEFER then
write_s("DEFER") write_s("DEFER")
else else
write_s("UNKNOWN<"); write_s("UNKNOWN<")
write_i(current_token^.kind); write_i(current_token^.kind)
write_c('>') write_c('>')
end; end
write_c(' '); write_c(' ')
i := i + 1u i := i + 1u
end; end
write_c('\n') write_c('\n')
end end
@@ -621,10 +619,10 @@ begin
elsif "to" = token_content then elsif "to" = token_content then
current_token.kind := TOKEN_TO current_token.kind := TOKEN_TO
elsif "true" = token_content then elsif "true" = token_content then
current_token.kind := TOKEN_BOOLEAN; current_token.kind := TOKEN_BOOLEAN
current_token.value.boolean_value := true current_token.value.boolean_value := true
elsif "false" = token_content then elsif "false" = token_content then
current_token.kind := TOKEN_BOOLEAN; current_token.kind := TOKEN_BOOLEAN
current_token.value.boolean_value := false current_token.value.boolean_value := false
elsif "nil" = token_content then elsif "nil" = token_content then
current_token.kind := TOKEN_NIL current_token.kind := TOKEN_NIL
@@ -641,55 +639,55 @@ begin
elsif "defer" = token_content then elsif "defer" = token_content then
current_token.kind := TOKEN_DEFER current_token.kind := TOKEN_DEFER
else else
current_token.kind := TOKEN_IDENTIFIER; current_token.kind := TOKEN_IDENTIFIER
current_token.value.string := string_dup(token_content) current_token.value.string := string_dup(token_content)
end; end
return current_token return current_token
end end
proc tokenize(source_code: SourceCode, tokens_size: pointer to Word) -> pointer to Token; proc tokenize(source_code: SourceCode, tokens_size: ^Word) -> ^Token;
var var
token_end: pointer to Char token_end: ^Char
tokens, current_token: pointer to Token tokens, current_token: ^Token
token_length: Word token_length: Word
first_char: Char first_char: Char
token_content: String token_content: String
begin begin
tokens_size^ := 0u; tokens_size^ := 0u
tokens := nil; tokens := nil
source_code := skip_spaces(source_code); source_code := skip_spaces(source_code)
while source_code.text.length <> 0u do while source_code.text.length <> 0u do
tokens := cast(reallocarray(cast(tokens: pointer to Byte), tokens_size^ + 1u, Token.size): pointer to Token); tokens := cast(reallocarray(cast(tokens: ^Byte), tokens_size^ + 1u, Token.size): ^Token)
current_token := tokens + tokens_size^; current_token := tokens + tokens_size^
first_char := source_code.text[1u]; first_char := source_code.text[1u]
if is_alpha(first_char) or first_char = '_' then if is_alpha(first_char) or first_char = '_' then
lex_identifier(@source_code, @token_content); lex_identifier(@source_code, @token_content)
current_token^ := categorize_identifier(token_content) current_token^ := categorize_identifier(token_content)
elsif is_digit(first_char) then elsif is_digit(first_char) then
token_end := nil; token_end := nil
current_token^.value.int_value := strtol(source_code.text.ptr, @token_end, 10); current_token^.value.int_value := strtol(source_code.text.ptr, @token_end, 10)
token_length := cast(token_end - source_code.text.ptr: Word); token_length := cast(token_end - source_code.text.ptr: Word)
if token_end^ = 'u' then if token_end^ = 'u' then
current_token^.kind := TOKEN_WORD; current_token^.kind := TOKEN_WORD
source_code := advance_source(source_code, token_length + 1u) source_code := advance_source(source_code, token_length + 1u)
else else
current_token^.kind := TOKEN_INTEGER; current_token^.kind := TOKEN_INTEGER
source_code := advance_source(source_code, token_length) source_code := advance_source(source_code, token_length)
end end
elsif first_char = '(' then elsif first_char = '(' then
source_code := advance_source(source_code, 1u); source_code := advance_source(source_code, 1u)
if source_code.text.length = 0u then if source_code.text.length = 0u then
current_token^.kind := TOKEN_LEFT_PAREN current_token^.kind := TOKEN_LEFT_PAREN
elsif source_code.text[1u] = '*' then elsif source_code.text[1u] = '*' then
source_code := advance_source(source_code, 1u); source_code := advance_source(source_code, 1u)
if lex_comment(@source_code, @token_content) then if lex_comment(@source_code, @token_content) then
current_token^.value.string := string_dup(token_content); current_token^.value.string := string_dup(token_content)
current_token^.kind := TOKEN_COMMENT current_token^.kind := TOKEN_COMMENT
else else
current_token^.kind := 0 current_token^.kind := 0
@@ -698,180 +696,180 @@ begin
current_token^.kind := TOKEN_LEFT_PAREN current_token^.kind := TOKEN_LEFT_PAREN
end end
elsif first_char = ')' then elsif first_char = ')' then
current_token^.kind := TOKEN_RIGHT_PAREN; current_token^.kind := TOKEN_RIGHT_PAREN
source_code := advance_source(source_code, 1u) source_code := advance_source(source_code, 1u)
elsif first_char = '\'' then elsif first_char = '\'' then
token_end := lex_character(source_code.text.ptr + 1, current_token); token_end := lex_character(source_code.text.ptr + 1, current_token)
token_length := cast(token_end - source_code.text.ptr: Word); token_length := cast(token_end - source_code.text.ptr: Word)
if token_end^ = '\'' then if token_end^ = '\'' then
current_token^.kind := TOKEN_CHARACTER; current_token^.kind := TOKEN_CHARACTER
source_code := advance_source(source_code, token_length + 1u) source_code := advance_source(source_code, token_length + 1u)
else else
source_code := advance_source(source_code, 1u) source_code := advance_source(source_code, 1u)
end end
elsif first_char = '"' then elsif first_char = '"' then
token_end := lex_string(source_code.text.ptr + 1, current_token); token_end := lex_string(source_code.text.ptr + 1, current_token)
if token_end^ = '"' then if token_end^ = '"' then
current_token^.kind := TOKEN_STRING; current_token^.kind := TOKEN_STRING
token_length := cast(token_end - source_code.text.ptr: Word); token_length := cast(token_end - source_code.text.ptr: Word)
source_code := advance_source(source_code, token_length + 1u) source_code := advance_source(source_code, token_length + 1u)
end end
elsif first_char = '[' then elsif first_char = '[' then
current_token^.kind := TOKEN_LEFT_SQUARE; current_token^.kind := TOKEN_LEFT_SQUARE
source_code := advance_source(source_code, 1u) source_code := advance_source(source_code, 1u)
elsif first_char = ']' then elsif first_char = ']' then
current_token^.kind := TOKEN_RIGHT_SQUARE; current_token^.kind := TOKEN_RIGHT_SQUARE
source_code := advance_source(source_code, 1u) source_code := advance_source(source_code, 1u)
elsif first_char = '>' then elsif first_char = '>' then
source_code := advance_source(source_code, 1u); source_code := advance_source(source_code, 1u)
if source_code.text.length = 0u then if source_code.text.length = 0u then
current_token^.kind := TOKEN_GREATER_THAN current_token^.kind := TOKEN_GREATER_THAN
elsif source_code.text[1u] = '=' then elsif source_code.text[1u] = '=' then
current_token^.kind := TOKEN_GREATER_EQUAL; current_token^.kind := TOKEN_GREATER_EQUAL
source_code := advance_source(source_code, 1u) source_code := advance_source(source_code, 1u)
elsif source_code.text[1u] = '>' then elsif source_code.text[1u] = '>' then
current_token^.kind := TOKEN_SHIFT_RIGHT; current_token^.kind := TOKEN_SHIFT_RIGHT
source_code := advance_source(source_code, 1u) source_code := advance_source(source_code, 1u)
else else
current_token^.kind := TOKEN_GREATER_THAN current_token^.kind := TOKEN_GREATER_THAN
end end
elsif first_char = '<' then elsif first_char = '<' then
source_code := advance_source(source_code, 1u); source_code := advance_source(source_code, 1u)
if source_code.text.length = 0u then if source_code.text.length = 0u then
current_token^.kind := TOKEN_LESS_THAN current_token^.kind := TOKEN_LESS_THAN
elsif source_code.text[1u] = '=' then elsif source_code.text[1u] = '=' then
current_token^.kind := TOKEN_LESS_EQUAL; current_token^.kind := TOKEN_LESS_EQUAL
source_code := advance_source(source_code, 1u) source_code := advance_source(source_code, 1u)
elsif source_code.text[1u] = '<' then elsif source_code.text[1u] = '<' then
current_token^.kind := TOKEN_SHIFT_LEFT; current_token^.kind := TOKEN_SHIFT_LEFT
source_code := advance_source(source_code, 1u) source_code := advance_source(source_code, 1u)
elsif source_code.text[1u] = '>' then elsif source_code.text[1u] = '>' then
current_token^.kind := TOKEN_NOT_EQUAL; current_token^.kind := TOKEN_NOT_EQUAL
source_code := advance_source(source_code, 1u) source_code := advance_source(source_code, 1u)
else else
current_token^.kind := TOKEN_LESS_THAN current_token^.kind := TOKEN_LESS_THAN
end end
elsif first_char = '=' then elsif first_char = '=' then
current_token^.kind := TOKEN_EQUAL; current_token^.kind := TOKEN_EQUAL
source_code := advance_source(source_code, 1u) source_code := advance_source(source_code, 1u)
elsif first_char = ';' then elsif first_char = ';' then
current_token^.kind := TOKEN_SEMICOLON; current_token^.kind := TOKEN_SEMICOLON
source_code := advance_source(source_code, 1u) source_code := advance_source(source_code, 1u)
elsif first_char = '.' then elsif first_char = '.' then
current_token^.kind := TOKEN_DOT; current_token^.kind := TOKEN_DOT
source_code := advance_source(source_code, 1u) source_code := advance_source(source_code, 1u)
elsif first_char = ',' then elsif first_char = ',' then
current_token^.kind := TOKEN_COMMA; current_token^.kind := TOKEN_COMMA
source_code := advance_source(source_code, 1u) source_code := advance_source(source_code, 1u)
elsif first_char = '+' then elsif first_char = '+' then
current_token^.kind := TOKEN_PLUS; current_token^.kind := TOKEN_PLUS
source_code := advance_source(source_code, 1u) source_code := advance_source(source_code, 1u)
elsif first_char = '-' then elsif first_char = '-' then
current_token^.kind := TOKEN_MINUS; current_token^.kind := TOKEN_MINUS
source_code := advance_source(source_code, 1u) source_code := advance_source(source_code, 1u)
elsif first_char = '*' then elsif first_char = '*' then
current_token^.kind := TOKEN_MULTIPLICATION; current_token^.kind := TOKEN_MULTIPLICATION
source_code := advance_source(source_code, 1u) source_code := advance_source(source_code, 1u)
elsif first_char = '/' then elsif first_char = '/' then
current_token^.kind := TOKEN_DIVISION; current_token^.kind := TOKEN_DIVISION
source_code := advance_source(source_code, 1u) source_code := advance_source(source_code, 1u)
elsif first_char = '%' then elsif first_char = '%' then
current_token^.kind := TOKEN_REMAINDER; current_token^.kind := TOKEN_REMAINDER
source_code := advance_source(source_code, 1u) source_code := advance_source(source_code, 1u)
elsif first_char = ':' then elsif first_char = ':' then
source_code := advance_source(source_code, 1u); source_code := advance_source(source_code, 1u)
if source_code.text.length = 0u then if source_code.text.length = 0u then
current_token^.kind := TOKEN_COLON current_token^.kind := TOKEN_COLON
elsif source_code.text[1u] = '=' then elsif source_code.text[1u] = '=' then
current_token^.kind := TOKEN_ASSIGNMENT; current_token^.kind := TOKEN_ASSIGNMENT
source_code := advance_source(source_code, 1u) source_code := advance_source(source_code, 1u)
else else
current_token^.kind := TOKEN_COLON current_token^.kind := TOKEN_COLON
end end
elsif first_char = '^' then elsif first_char = '^' then
current_token^.kind := TOKEN_HAT; current_token^.kind := TOKEN_HAT
source_code := advance_source(source_code, 1u) source_code := advance_source(source_code, 1u)
elsif first_char = '@' then elsif first_char = '@' then
current_token^.kind := TOKEN_AT; current_token^.kind := TOKEN_AT
source_code := advance_source(source_code, 1u) source_code := advance_source(source_code, 1u)
else else
current_token^.kind := 0; current_token^.kind := 0
source_code := advance_source(source_code, 1u) source_code := advance_source(source_code, 1u)
end; end
if current_token^.kind <> 0 then if current_token^.kind <> 0 then
tokens_size^ := tokens_size^ + 1u; tokens_size^ := tokens_size^ + 1u
source_code := skip_spaces(source_code) source_code := skip_spaces(source_code)
else else
write_s("Lexical analysis error on \""); write_s("Lexical analysis error on \"")
write_c(first_char); write_c(first_char)
write_s("\".\n") write_s("\".\n")
end end
end; end
return tokens return tokens
end end
proc parse_literal(tokens: pointer to pointer to Token, tokens_size: pointer to Word) -> pointer to Literal; proc parse_literal(tokens: ^^Token, tokens_size: ^Word) -> ^Literal;
begin begin
return cast(calloc(1u, Literal.size): pointer to Literal) return cast(calloc(1u, Literal.size): ^Literal)
end end
proc parse_constant_definition(tokens: pointer to pointer to Token, proc parse_constant_definition(tokens: ^^Token,
tokens_size: pointer to Word) -> pointer to ConstantDefinition; tokens_size: ^Word) -> ^ConstantDefinition;
var var
result: pointer to ConstantDefinition result: ^ConstantDefinition
begin begin
result := cast(calloc(1u, ConstantDefinition.size): pointer to ConstantDefinition); result := cast(calloc(1u, ConstantDefinition.size): ^ConstantDefinition)
result^.name := cast(malloc(strlen(tokens^^.value.string_value)): pointer to Char); result^.name := cast(malloc(strlen(tokens^^.value.string_value)): ^Char)
strcpy(result^.name, tokens^^.value.string_value); strcpy(result^.name, tokens^^.value.string_value)
tokens^ := tokens^ + 2u; tokens^ := tokens^ + 2u
tokens_size := tokens_size - 2u; tokens_size := tokens_size - 2u
write_z(result^.name); write_z(result^.name)
write_c('\n'); write_c('\n')
result^.body := parse_literal(tokens, tokens_size); result^.body := parse_literal(tokens, tokens_size)
tokens^ := tokens^ + 2u; tokens^ := tokens^ + 2u
tokens_size := tokens_size - 2u; tokens_size := tokens_size - 2u
return result return result
end end
proc parse_program(tokens: pointer to pointer to Token, tokens_size: pointer to Word) -> pointer to Program; proc parse_program(tokens: ^^Token, tokens_size: ^Word) -> ^Program;
var var
result: pointer to Program result: ^Program
current_constant: pointer to pointer to ConstantDefinition current_constant: ^^ConstantDefinition
begin begin
result := cast(calloc(1u, Program.size): pointer to Program); result := cast(calloc(1u, Program.size): ^Program)
result^.constants.elements := nil; result^.constants.elements := nil
result^.constants.count := 0u; result^.constants.count := 0u
if tokens^^.kind = TOKEN_CONST then if tokens^^.kind = TOKEN_CONST then
tokens^ := tokens^ + 1; tokens^ := tokens^ + 1
tokens_size^ := tokens_size^ - 1u; tokens_size^ := tokens_size^ - 1u
while tokens_size^ > 0u and tokens^^.kind = TOKEN_IDENTIFIER do while tokens_size^ > 0u and tokens^^.kind = TOKEN_IDENTIFIER do
result^.constants.elements := cast( result^.constants.elements := cast(
reallocarray( reallocarray(
cast(result^.constants.elements: pointer to Byte), cast(result^.constants.elements: ^Byte),
result^.constants.count + 1u, result^.constants.count + 1u,
(pointer to ConstantDefinition).size (^ConstantDefinition).size
) : pointer to pointer to ConstantDefinition); ) : ^^ConstantDefinition)
current_constant := result^.constants.elements + result^.constants.count; current_constant := result^.constants.elements + result^.constants.count
result^.constants.count := result^.constants.count + 1u; result^.constants.count := result^.constants.count + 1u
current_constant^ := parse_constant_definition(tokens, tokens_size); current_constant^ := parse_constant_definition(tokens, tokens_size)
if current_constant^ = nil then if current_constant^ = nil then
return nil return nil
end end
@@ -879,20 +877,20 @@ begin
end end
end end
proc parse_command_line*(argc: Int, argv: pointer to pointer to Char) -> pointer to CommandLine; proc parse_command_line*(argc: Int, argv: ^^Char) -> ^CommandLine;
var var
parameter: pointer to pointer to Char parameter: ^^Char
i: Int i: Int
result: pointer to CommandLine result: ^CommandLine
begin begin
i := 1; i := 1
result := cast(malloc(CommandLine.size): pointer to CommandLine); result := cast(malloc(CommandLine.size): ^CommandLine)
result^.tokenize := false; result^.tokenize := false
result^.syntax_tree := false; result^.syntax_tree := false
result^.input := nil; result^.input := nil
while i < argc do while i < argc do
parameter := argv + i; parameter := argv + i
if strcmp(parameter^, "--tokenize\0".ptr) = 0 then if strcmp(parameter^, "--tokenize\0".ptr) = 0 then
result^.tokenize := true result^.tokenize := true
@@ -901,53 +899,53 @@ begin
elsif parameter^^ <> '-' then elsif parameter^^ <> '-' then
result^.input := parameter^ result^.input := parameter^
else else
write_s("Fatal error: Unknown command line options:"); write_s("Fatal error: Unknown command line options:")
write_c(' '); write_c(' ')
write_z(parameter^); write_z(parameter^)
write_s(".\n"); write_s(".\n")
return nil return nil
end; end
i := i + 1 i := i + 1
end; end
if result^.input = nil then if result^.input = nil then
write_s("Fatal error: no input files.\n"); write_s("Fatal error: no input files.\n")
return nil return nil
end; end
return result return result
end end
proc process(argc: Int, argv: pointer to pointer to Char) -> Int; proc process(argc: Int, argv: ^^Char) -> Int;
var var
tokens: pointer to Token tokens: ^Token
tokens_size: Word tokens_size: Word
source_code: SourceCode source_code: SourceCode
command_line: pointer to CommandLine command_line: ^CommandLine
begin begin
command_line := parse_command_line(argc, argv); command_line := parse_command_line(argc, argv)
if command_line = nil then if command_line = nil then
return 2 return 2
end; end
source_code.position := make_position(); source_code.position := make_position()
if not read_source(command_line^.input, @source_code.text) then if not read_source(command_line^.input, @source_code.text) then
perror(command_line^.input); perror(command_line^.input)
return 3 return 3
end; end
tokens := tokenize(source_code, @tokens_size); tokens := tokenize(source_code, @tokens_size)
if command_line^.tokenize then if command_line^.tokenize then
print_tokens(tokens, tokens_size) print_tokens(tokens, tokens_size)
end; end
if command_line^.syntax_tree then if command_line^.syntax_tree then
parse_program(@tokens, @tokens_size) parse_program(@tokens, @tokens_size)
end; end
return 0 return 0
end end
begin begin
exit(process(cast(count: Int), cast(parameters: pointer to pointer to Char))) exit(process(cast(count: Int), cast(parameters: ^^Char)))
end. end.