gencpp/project/components/lexer.cpp

1272 lines
24 KiB
C++
Raw Normal View History

#ifdef GEN_INTELLISENSE_DIRECTIVES
#pragma once
#include "interface.upfront.cpp"
#include "gen/etoktype.cpp"
#endif
namespace parser {
enum TokFlags : u32
{
TF_Operator = bit(0),
TF_Assign = bit(1),
TF_Preprocess = bit(2),
TF_Preprocess_Cond = bit(3),
TF_Attribute = bit(6),
TF_AccessOperator = bit( 7 ),
TF_AccessSpecifier = bit( 8 ),
TF_Specifier = bit( 9 ),
TF_EndDefinition = bit( 10 ), // Either ; or }
TF_Formatting = bit( 11 ),
TF_Literal = bit( 12 ),
TF_Null = 0,
};
struct Token
{
char const* Text;
sptr Length;
TokType Type;
s32 Line;
s32 Column;
u32 Flags;
operator bool()
{
return Text && Length && Type != TokType::Invalid;
}
operator StrC()
{
return { Length, Text };
}
bool is_access_operator()
{
return bitfield_is_equal( u32, Flags, TF_AccessOperator );
}
bool is_access_specifier()
{
return bitfield_is_equal( u32, Flags, TF_AccessSpecifier );
}
bool is_attribute()
{
return bitfield_is_equal( u32, Flags, TF_Attribute );
}
bool is_operator()
{
return bitfield_is_equal( u32, Flags, TF_Operator );
}
bool is_preprocessor()
{
return bitfield_is_equal( u32, Flags, TF_Preprocess );
}
bool is_preprocess_cond()
{
return bitfield_is_equal( u32, Flags, TF_Preprocess_Cond );
}
bool is_specifier()
{
return bitfield_is_equal( u32, Flags, TF_Specifier );
}
bool is_end_definition()
{
return bitfield_is_equal( u32, Flags, TF_EndDefinition );
}
AccessSpec to_access_specifier()
{
return scast(AccessSpec, Type);
}
String to_string()
{
String result = String::make_reserve( GlobalAllocator, kilobytes(4) );
StrC type_str = ETokType::to_str( Type );
result.append_fmt( "Line: %d Column: %d, Type: %.*s Content: %.*s"
, Line, Column
, type_str.Len, type_str.Ptr
, Length, Text
);
return result;
}
};
constexpr Token NullToken { nullptr, 0, TokType::Invalid, false, 0, TF_Null };
struct TokArray
{
Array<Token> Arr;
s32 Idx;
bool __eat( TokType type );
Token& current( bool skip_formatting = true )
{
if ( skip_formatting )
{
while ( Arr[Idx].Type == TokType::NewLine || Arr[Idx].Type == TokType::Comment )
Idx++;
}
return Arr[Idx];
}
Token& previous( bool skip_formatting = false )
{
s32 idx = this->Idx;
if ( skip_formatting )
{
while ( Arr[idx].Type == TokType::NewLine )
idx--;
return Arr[idx];
}
return Arr[idx - 1];
}
Token& next( bool skip_formatting = false )
{
s32 idx = this->Idx;
if ( skip_formatting )
{
while ( Arr[idx].Type == TokType::NewLine )
idx++;
return Arr[idx + 1];
}
return Arr[idx + 1];
}
Token& operator []( s32 idx )
{
return Arr[idx];
}
};
global Arena_256KB defines_map_arena;
global HashTable<StrC> defines;
global Array<Token> Tokens;
#define current ( * scanner )
#define move_forward() \
{ \
if ( current == '\n' ) \
{ \
line++; \
column = 1; \
} \
else \
{ \
column++; \
} \
left--; \
scanner++; \
}
#define SkipWhitespace() \
while ( left && char_is_space( current ) ) \
{ \
move_forward(); \
}
#define end_line() \
do \
{ \
while ( left && current == ' ' ) \
{ \
move_forward(); \
} \
if ( left && current == '\r' ) \
{ \
move_forward(); \
move_forward(); \
} \
else if ( left && current == '\n' ) \
{ \
move_forward(); \
} \
} \
while (0)
enum
{
Lex_Continue,
Lex_ReturnNull,
};
forceinline
s32 lex_preprocessor_directive(
StrC& content
, s32& left
, char const*& scanner
, s32& line
, s32& column
, HashTable<StrC>& defines
, Token& token )
{
char const* hash = scanner;
Tokens.append( { hash, 1, TokType::Preprocess_Hash, line, column, TF_Preprocess } );
move_forward();
SkipWhitespace();
token.Text = scanner;
while (left && ! char_is_space(current) )
{
move_forward();
token.Length++;
}
token.Type = ETokType::to_type( token );
bool is_preprocessor = token.Type >= TokType::Preprocess_Define && token.Type <= TokType::Preprocess_Pragma;
if ( ! is_preprocessor )
{
token.Type = TokType::Preprocess_Unsupported;
// Its an unsupported directive, skip it
s32 within_string = false;
s32 within_char = false;
while ( left )
{
if ( current == '"' && ! within_char )
within_string ^= true;
if ( current == '\'' && ! within_string )
within_char ^= true;
if ( current == '\\' && ! within_string && ! within_char )
{
move_forward();
token.Length++;
if ( current == '\r' )
{
move_forward();
token.Length++;
}
if ( current == '\n' )
{
move_forward();
token.Length++;
continue;
}
else
{
log_failure( "gen::Parser::lex: Invalid escape sequence '\\%c' (%d, %d)"
" in preprocessor directive (%d, %d)\n%.100s"
, current, line, column
, token.Line, token.Column, token.Text );
break;
}
}
if ( current == '\r' )
{
move_forward();
token.Length++;
}
if ( current == '\n' )
{
move_forward();
token.Length++;
break;
}
move_forward();
token.Length++;
}
token.Length = token.Length + token.Text - hash;
token.Text = hash;
Tokens.append( token );
return Lex_Continue; // Skip found token, its all handled here.
}
if ( token.Type == TokType::Preprocess_Else || token.Type == TokType::Preprocess_EndIf )
{
token.Flags |= TF_Preprocess_Cond;
Tokens.append( token );
end_line();
return Lex_Continue;
}
else if ( token.Type >= TokType::Preprocess_If && token.Type <= TokType::Preprocess_ElIf )
{
token.Flags |= TF_Preprocess_Cond;
}
Tokens.append( token );
SkipWhitespace();
if ( token.Type == TokType::Preprocess_Define )
{
Token name = { scanner, 0, TokType::Identifier, line, column, TF_Preprocess };
name.Text = scanner;
name.Length = 1;
move_forward();
while ( left && ( char_is_alphanumeric(current) || current == '_' ) )
{
move_forward();
name.Length++;
}
if ( left && current == '(' )
{
move_forward();
name.Length++;
}
Tokens.append( name );
u64 key = crc32( name.Text, name.Length );
defines.set( key, name );
}
Token preprocess_content = { scanner, 0, TokType::Preprocess_Content, line, column, TF_Preprocess };
if ( token.Type == TokType::Preprocess_Include )
{
preprocess_content.Type = TokType::String;
if ( current != '"' && current != '<' )
{
String directive_str = String::fmt_buf( GlobalAllocator, "%.*s", min( 80, left + preprocess_content.Length ), token.Text );
log_failure( "gen::Parser::lex: Expected '\"' or '<' after #include, not '%c' (%d, %d)\n%s"
, current
, preprocess_content.Line
, preprocess_content.Column
, directive_str.Data
);
return Lex_ReturnNull;
}
move_forward();
preprocess_content.Length++;
while ( left && current != '"' && current != '>' )
{
move_forward();
preprocess_content.Length++;
}
move_forward();
preprocess_content.Length++;
if ( current == '\r' && scanner[1] == '\n' )
{
move_forward();
move_forward();
}
else if ( current == '\n' )
{
move_forward();
}
Tokens.append( preprocess_content );
return Lex_Continue; // Skip found token, its all handled here.
}
s32 within_string = false;
s32 within_char = false;
// SkipWhitespace();
while ( left )
{
if ( current == '"' && ! within_char )
within_string ^= true;
if ( current == '\'' && ! within_string )
within_char ^= true;
if ( current == '\\' && ! within_string && ! within_char )
{
move_forward();
preprocess_content.Length++;
if ( current == '\r' )
{
move_forward();
preprocess_content.Length++;
}
if ( current == '\n' )
{
move_forward();
preprocess_content.Length++;
continue;
}
else
{
String directive_str = String::make_length( GlobalAllocator, token.Text, token.Length );
String content_str = String::fmt_buf( GlobalAllocator, "%.*s", min( 400, left + preprocess_content.Length ), preprocess_content.Text );
log_failure( "gen::Parser::lex: Invalid escape sequence '\\%c' (%d, %d)"
" in preprocessor directive '%s' (%d, %d)\n%s"
, current, line, column
, directive_str, preprocess_content.Line, preprocess_content.Column
, content_str );
break;
}
}
if ( current == '\r' )
{
move_forward();
}
if ( current == '\n' )
{
move_forward();
break;
}
move_forward();
preprocess_content.Length++;
}
Tokens.append( preprocess_content );
return Lex_Continue; // Skip found token, its all handled here.
}
forceinline
void lex_found_token( StrC& content
, s32& left
, char const*& scanner
, s32& line
, s32& column
, HashTable<StrC>& defines
, Token& token )
{
if ( token.Type != TokType::Invalid )
{
Tokens.append( token );
return;
}
TokType type = ETokType::to_type( token );
if (type <= TokType::Access_Public && type >= TokType::Access_Private )
{
token.Flags |= TF_AccessSpecifier;
}
if ( type > TokType::__Attributes_Start )
{
token.Flags |= TF_Attribute;
}
if ( type == ETokType::Decl_Extern_Linkage )
{
SkipWhitespace();
if ( current != '"' )
{
type = ETokType::Spec_Extern;
token.Flags |= TF_Specifier;
}
token.Type = type;
Tokens.append( token );
return;
}
if ( ( type <= TokType::Star && type >= TokType::Spec_Alignas)
|| type == TokType::Ampersand
|| type == TokType::Ampersand_DBL )
{
token.Type = type;
token.Flags |= TF_Specifier;
Tokens.append( token );
return;
}
if ( type != TokType::Invalid )
{
token.Type = type;
Tokens.append( token );
return;
}
u64 key = 0;
if ( current == '(')
key = crc32( token.Text, token.Length + 1 );
else
key = crc32( token.Text, token.Length );
StrC* define = defines.get( key );
if ( define )
{
token.Type = TokType::Preprocess_Macro;
// Want to ignore any arguments the define may have as they can be execution expressions.
if ( left && current == '(' )
{
move_forward();
token.Length++;
s32 level = 0;
while ( left && (current != ')' || level > 0) )
{
if ( current == '(' )
level++;
else if ( current == ')' && level > 0 )
level--;
move_forward();
token.Length++;
}
move_forward();
token.Length++;
}
if ( current == '\r' && scanner[1] == '\n' )
{
move_forward();
}
else if ( current == '\n' )
{
move_forward();
}
}
else
{
token.Type = TokType::Identifier;
}
Tokens.append( token );
}
neverinline
// TokArray lex( Array<Token> tokens, StrC content )
TokArray lex( StrC content )
{
s32 left = content.Len;
char const* scanner = content.Ptr;
char const* word = scanner;
s32 word_length = 0;
s32 line = 1;
s32 column = 1;
SkipWhitespace();
if ( left <= 0 )
{
log_failure( "gen::lex: no tokens found (only whitespace provided)" );
return { { nullptr }, 0 };
}
for ( StringCached entry : PreprocessorDefines )
{
s32 length = 0;
char const* scanner = entry.Data;
while ( entry.length() > length && (char_is_alphanumeric( *scanner ) || *scanner == '_') )
{
scanner++;
length ++;
}
if ( scanner[0] == '(' )
{
length++;
}
u64 key = crc32( entry.Data, length );
defines.set( key, entry );
}
Tokens.clear();
while (left )
{
#if 0
if (Tokens.num())
{
log_fmt("\nLastTok: %S", Tokens.back().to_string());
}
#endif
Token token = { scanner, 0, TokType::Invalid, line, column, TF_Null };
bool is_define = false;
if ( column == 1 )
{
if ( current == '\r')
{
move_forward();
token.Length = 1;
}
if ( current == '\n' )
{
move_forward();
token.Type = TokType::NewLine;
token.Length++;
Tokens.append( token );
continue;
}
}
token.Length = 0;
SkipWhitespace();
if ( left <= 0 )
break;
switch ( current )
{
case '#':
{
s32 result = lex_preprocessor_directive( content, left, scanner, line, column, defines, token );
switch ( result )
{
case Lex_Continue:
continue;
case Lex_ReturnNull:
return { { nullptr }, 0 };
}
}
case '.':
{
token.Text = scanner;
token.Length = 1;
token.Type = TokType::Access_MemberSymbol;
token.Flags = TF_AccessOperator;
if (left) {
move_forward();
}
if ( current == '.' )
{
move_forward();
if( current == '.' )
{
token.Length = 3;
token.Type = TokType::Varadic_Argument;
token.Flags = TF_Null;
move_forward();
}
else
{
String context_str = String::fmt_buf( GlobalAllocator, "%s", scanner, min( 100, left ) );
log_failure( "gen::lex: invalid varadic argument, expected '...' got '..%c' (%d, %d)\n%s", current, line, column, context_str );
}
}
goto FoundToken;
}
case '&' :
{
token.Text = scanner;
token.Length = 1;
token.Type = TokType::Ampersand;
token.Flags |= TF_Operator;
token.Flags |= TF_Specifier;
if (left)
move_forward();
if ( current == '&' ) // &&
{
token.Length = 2;
token.Type = TokType::Ampersand_DBL;
if (left)
move_forward();
}
goto FoundToken;
}
case ':':
{
token.Text = scanner;
token.Length = 1;
token.Type = TokType::Assign_Classifer;
// Can be either a classifier (ParentType, Bitfield width), or ternary else
// token.Type = TokType::Colon;
if (left)
move_forward();
if ( current == ':' )
{
move_forward();
token.Type = TokType::Access_StaticSymbol;
token.Length++;
}
goto FoundToken;
}
case '{':
{
token.Text = scanner;
token.Length = 1;
token.Type = TokType::BraceCurly_Open;
if (left)
move_forward();
goto FoundToken;
}
case '}':
{
token.Text = scanner;
token.Length = 1;
token.Type = TokType::BraceCurly_Close;
token.Flags = TF_EndDefinition;
if (left)
move_forward();
end_line();
goto FoundToken;
}
case '[':
{
token.Text = scanner;
token.Length = 1;
token.Type = TokType::BraceSquare_Open;
if ( left )
{
move_forward();
if ( current == ']' )
{
token.Length = 2;
token.Type = TokType::Operator;
move_forward();
}
}
goto FoundToken;
}
case ']':
{
token.Text = scanner;
token.Length = 1;
token.Type = TokType::BraceSquare_Close;
if (left)
move_forward();
goto FoundToken;
}
case '(':
{
token.Text = scanner;
token.Length = 1;
token.Type = TokType::Capture_Start;
if (left)
move_forward();
goto FoundToken;
}
case ')':
{
token.Text = scanner;
token.Length = 1;
token.Type = TokType::Capture_End;
if (left)
move_forward();
goto FoundToken;
}
case '\'':
{
token.Text = scanner;
token.Length = 1;
token.Type = TokType::Char;
token.Flags = TF_Literal;
move_forward();
if ( left && current == '\\' )
{
move_forward();
token.Length++;
if ( current == '\'' )
{
move_forward();
token.Length++;
}
}
while ( left && current != '\'' )
{
move_forward();
token.Length++;
}
if ( left )
{
move_forward();
token.Length++;
}
goto FoundToken;
}
case ',':
{
token.Text = scanner;
token.Length = 1;
token.Type = TokType::Comma;
token.Flags = TF_Operator;
if (left)
move_forward();
goto FoundToken;
}
case '*':
{
token.Text = scanner;
token.Length = 1;
token.Type = TokType::Star;
token.Flags |= TF_Specifier;
token.Flags |= TF_Operator;
if (left)
move_forward();
if ( current == '=' )
{
token.Length++;
token.Flags |= TF_Assign;
// token.Type = TokType::Assign_Multiply;
if ( left )
move_forward();
}
goto FoundToken;
}
case ';':
{
token.Text = scanner;
token.Length = 1;
token.Type = TokType::Statement_End;
token.Flags = TF_EndDefinition;
if (left)
move_forward();
end_line();
goto FoundToken;
}
case '"':
{
token.Text = scanner;
token.Length = 1;
token.Type = TokType::String;
token.Flags |= TF_Literal;
move_forward();
while ( left )
{
if ( current == '"' )
{
move_forward();
break;
}
if ( current == '\\' )
{
move_forward();
token.Length++;
if ( left )
{
move_forward();
token.Length++;
}
continue;
}
move_forward();
token.Length++;
}
goto FoundToken;
}
case '?':
{
token.Text = scanner;
token.Length = 1;
token.Type = TokType::Operator;
// token.Type = TokType::Ternary;
token.Flags = TF_Operator;
if (left)
move_forward();
goto FoundToken;
}
case '=':
{
token.Text = scanner;
token.Length = 1;
token.Type = TokType::Operator;
// token.Type = TokType::Assign;
token.Flags = TF_Operator;
token.Flags |= TF_Assign;
if (left)
move_forward();
if ( current == '=' )
{
token.Length++;
token.Flags = TF_Operator;
if (left)
move_forward();
}
goto FoundToken;
}
case '+':
{
// token.Type = TokType::Add
}
case '%':
{
// token.Type = TokType::Modulo;
}
case '^':
{
// token.Type = TokType::B_XOr;
}
case '~':
{
// token.Type = TokType::Unary_Not;
}
case '!':
{
// token.Type = TokType::L_Not;
}
case '<':
{
// token.Type = TokType::Lesser;
}
case '>':
{
// token.Type = TokType::Greater;
}
case '|':
{
token.Text = scanner;
token.Length = 1;
token.Type = TokType::Operator;
token.Flags = TF_Operator;
// token.Type = TokType::L_Or;
if (left)
move_forward();
if ( current == '=' )
{
token.Length++;
token.Flags |= TF_Assign;
// token.Flags |= TokFlags::Assignment;
// token.Type = TokType::Assign_L_Or;
if (left)
move_forward();
}
else while ( left && current == *(scanner - 1) && token.Length < 3 )
{
token.Length++;
if (left)
move_forward();
}
goto FoundToken;
}
// Dash is unfortunatlly a bit more complicated...
case '-':
{
token.Text = scanner;
token.Length = 1;
token.Type = TokType::Operator;
// token.Type = TokType::Subtract;
token.Flags = TF_Operator;
if ( left )
{
move_forward();
if ( current == '>' )
{
token.Length++;
// token.Type = TokType::Access_PointerToMemberSymbol;
token.Flags |= TF_AccessOperator;
move_forward();
if ( current == '*' )
{
// token.Type = TokType::Access_PointerToMemberOfPointerSymbol;
token.Length++;
move_forward();
}
}
else if ( current == '=' )
{
token.Length++;
// token.Type = TokType::Assign_Subtract;
token.Flags |= TF_Assign;
if (left)
move_forward();
}
else while ( left && current == *(scanner - 1) && token.Length < 3 )
{
token.Length++;
if (left)
move_forward();
}
}
goto FoundToken;
}
case '/':
{
token.Text = scanner;
token.Length = 1;
token.Type = TokType::Operator;
// token.Type = TokType::Divide;
token.Flags = TF_Operator;
move_forward();
if ( left )
{
if ( current == '=' )
{
// token.Type = TokeType::Assign_Divide;
move_forward();
token.Length++;
token.Flags = TF_Assign;
}
else if ( current == '/' )
{
token.Type = TokType::Comment;
token.Length = 2;
token.Flags = TF_Null;
move_forward();
while ( left && current != '\n' && current != '\r' )
{
move_forward();
token.Length++;
}
if ( current == '\r' )
{
move_forward();
token.Length++;
}
if ( current == '\n' )
{
move_forward();
token.Length++;
}
Tokens.append( token );
continue;
}
else if ( current == '*' )
{
token.Type = TokType::Comment;
token.Length = 2;
token.Flags = TF_Null;
move_forward();
bool star = current == '*';
bool slash = scanner[1] == '/';
bool at_end = star && slash;
while ( left && ! at_end )
{
move_forward();
token.Length++;
star = current == '*';
slash = scanner[1] == '/';
at_end = star && slash;
}
token.Length += 2;
move_forward();
move_forward();
if ( current == '\r' )
{
move_forward();
token.Length++;
}
if ( current == '\n' )
{
move_forward();
token.Length++;
}
Tokens.append( token );
// end_line();
continue;
}
}
goto FoundToken;
}
}
if ( char_is_alpha( current ) || current == '_' )
{
token.Text = scanner;
token.Length = 1;
move_forward();
while ( left && ( char_is_alphanumeric(current) || current == '_' ) )
{
move_forward();
token.Length++;
}
goto FoundToken;
}
else if ( char_is_digit(current) )
{
// This is a very brute force lex, no checks are done for validity of literal.
token.Text = scanner;
token.Length = 1;
token.Type = TokType::Number;
token.Flags = TF_Literal;
move_forward();
if (left
&& ( current == 'x' || current == 'X'
|| current == 'b' || current == 'B'
|| current == 'o' || current == 'O' )
)
{
move_forward();
token.Length++;
while ( left && char_is_hex_digit(current) )
{
move_forward();
token.Length++;
}
goto FoundToken;
}
while ( left && char_is_digit(current) )
{
move_forward();
token.Length++;
}
if ( left && current == '.' )
{
move_forward();
token.Length++;
while ( left && char_is_digit(current) )
{
move_forward();
token.Length++;
}
// Handle number literal suffixes in a botched way
if (left && (
current == 'l' || current == 'L' || // long/long long
current == 'u' || current == 'U' || // unsigned
current == 'f' || current == 'F' || // float
current == 'i' || current == 'I' || // imaginary
current == 'z' || current == 'Z')) // complex
{
char prev = current;
move_forward();
token.Length++;
// Handle 'll'/'LL' as a special case when we just processed an 'l'/'L'
if (left && (prev == 'l' || prev == 'L') && (current == 'l' || current == 'L'))
{
move_forward();
token.Length++;
}
}
}
goto FoundToken;
}
else
{
s32 start = max( 0, Tokens.num() - 100 );
log_fmt("\n%d\n", start);
for ( s32 idx = start; idx < Tokens.num(); idx++ )
{
log_fmt( "Token %d Type: %s : %.*s\n"
, idx
, ETokType::to_str( Tokens[ idx ].Type ).Ptr
, Tokens[ idx ].Length, Tokens[ idx ].Text
);
}
String context_str = String::fmt_buf( GlobalAllocator, "%.*s", min( 100, left ), scanner );
log_failure( "Failed to lex token '%c' (%d, %d)\n%s", current, line, column, context_str );
// Skip to next whitespace since we can't know if anything else is valid until then.
while ( left && ! char_is_space( current ) )
{
move_forward();
}
}
FoundToken:
lex_found_token( content, left, scanner, line, column, defines, token );
}
if ( Tokens.num() == 0 )
{
log_failure( "Failed to lex any tokens" );
return { { nullptr }, 0 };
}
defines.clear();
// defines_map_arena.free();
return { Tokens, 0 };
}
#undef current
#undef move_forward
#undef SkipWhitespace
// namespace parser
}