private void AdvanceExponent(char test)
{
char c = Buffer.Peek();
if (char.ToLower(test) == char.ToLower(c))
{
Buffer.AdvanceColumn();
c = Buffer.Peek();
if (c == '+' || c == '-')
{
Buffer.AdvanceColumn();
}
Buffer.AdvanceColumnsWhile(SyntaxUtils.IsNumeric);
}
}
private bool LexUntilCodeEnd(CommandResult commandResult)
{
// Special case, we dont want to parse doxygen stuff inside a code section
// so we wait until a @endcode follows
bool isComplete = false;
while (!Buffer.IsEOF)
{
char c0 = Buffer.Peek();
char c1 = Buffer.Peek(1);
if ((c0 == '@' || c0 == '\\') && SyntaxUtils.IsIdentStart(c1))
{
Buffer.StartLexeme();
Buffer.AdvanceColumn();
Buffer.AdvanceColumnsWhile(SyntaxUtils.IsIdentPart);
string ident = Buffer.GetSourceText(Buffer.LexemeStart.Index + 1, Buffer.LexemeWidth - 1);
if ("endcode".Equals(ident))
{
PushToken(DoxygenTokenPool.Make(DoxygenTokenKind.CommandEnd, Buffer.LexemeRange, true));
isComplete = true;
break;
}
}
else if (SyntaxUtils.IsLineBreak(c0))
{
int lb = SyntaxUtils.GetLineBreakChars(c0, c1);
Buffer.AdvanceLine(lb);
}
else if ('\t'.Equals(c0))
{
Buffer.AdvanceTab();
}
else
{
Buffer.AdvanceColumn();
}
}
if (!isComplete)
{
AddError(commandResult.StartPos, $"Unterminated code-block, expect '@endcode' or '\\endcode'", "Code", commandResult.CommandName);
return(false);
}
return(true);
}
private CommandResult LexCommandTokens()
{
Debug.Assert(DoxygenSyntax.IsCommandBegin(Buffer.Peek()));
// Command
Buffer.StartLexeme();
Buffer.AdvanceColumn();
DoxygenTokenKind kind = DoxygenTokenKind.Command;
{
char first = Buffer.Peek();
switch (first)
{
case '{':
case '}':
kind = (first == '{') ? DoxygenTokenKind.GroupStart : DoxygenTokenKind.GroupEnd;
Buffer.AdvanceColumn();
break;
case '$':
case '@':
case '\\':
case '~':
case '<':
case '=':
case '>':
case '#':
case '"':
Buffer.AdvanceColumn();
break;
case ':':
case '|':
case '-':
Buffer.AdvanceColumnsWhile(d => d.Equals(first));
break;
default:
if (DoxygenSyntax.IsCommandIdentStart(first))
{
while (!Buffer.IsEOF)
{
if (!DoxygenSyntax.IsCommandIdentPart(Buffer.Peek()))
{
break;
}
Buffer.AdvanceColumn();
}
}
break;
}
}
TextPosition commandStart = Buffer.LexemeStart;
int commandLen = Buffer.LexemeWidth;
string commandName = Buffer.GetSourceText(Buffer.LexemeStart.Index + 1, commandLen - 1);
var rule = DoxygenSyntax.GetCommandRule(commandName);
if (rule != null)
{
if (rule.Kind == DoxygenSyntax.CommandKind.StartCommandBlock)
{
kind = DoxygenTokenKind.CommandStart;
}
else if (rule.Kind == DoxygenSyntax.CommandKind.EndCommandBlock)
{
kind = DoxygenTokenKind.CommandEnd;
}
}
else
{
// @NOTE(final): Group start/end are not a "known" command
if (kind != DoxygenTokenKind.GroupStart && kind != DoxygenTokenKind.GroupEnd)
{
kind = DoxygenTokenKind.InvalidCommand;
}
}
DoxygenToken commandToken = DoxygenTokenPool.Make(kind, Buffer.LexemeRange, true);
PushToken(commandToken);
CommandResult result = new CommandResult(commandStart, rule, commandName);
string typeName = "Command";
if (rule != null)
{
int argNumber = 0;
int argCount = rule.Args.Count();
bool noMoreArgs = false;
foreach (var arg in rule.Args)
{
// @TODO(final): Handle rule repeat type for arguments on same type
char first = Buffer.Peek();
if (!arg.Flags.HasFlag(DoxygenSyntax.ArgumentFlags.DirectlyAfterCommand))
{
if (SyntaxUtils.IsSpacing(first) || first == '\t')
{
Buffer.SkipSpacings(TextStream.SkipType.All);
}
else
{
// No more arguments are following
noMoreArgs = true;
}
}
Buffer.StartLexeme();
// Prefix
string prefix = arg.Prefix;
string postfix = arg.Postfix;
bool hadPrefix = false;
if (prefix != null && !noMoreArgs)
{
if (!string.IsNullOrEmpty(prefix))
{
if (Buffer.CompareText(0, prefix) == 0)
{
Buffer.AdvanceColumns(prefix.Length);
hadPrefix = true;
}
}
else if ((prefix.Length == 0) && (!string.IsNullOrEmpty(postfix)))
{
hadPrefix = true;
}
}
switch (arg.Kind)
{
case DoxygenSyntax.ArgumentKind.PrefixToPostfix:
{
if (hadPrefix && !noMoreArgs)
{
Debug.Assert(!string.IsNullOrEmpty(postfix));
bool foundPrefixToPostfix = false;
while (!Buffer.IsEOF)
{
if (Buffer.CompareText(0, postfix) == 0)
{
Buffer.AdvanceColumns(postfix.Length);
foundPrefixToPostfix = true;
break;
}
else if (SyntaxUtils.IsLineBreak(Buffer.Peek()))
{
break;
}
else
{
Buffer.AdvanceColumn();
}
}
if (arg.IsOptional || foundPrefixToPostfix)
{
DoxygenToken argToken = DoxygenTokenPool.Make(DoxygenTokenKind.ArgumentCaption, Buffer.LexemeRange, foundPrefixToPostfix);
PushToken(argToken);
}
else if (arg.IsRequired)
{
AddError(Buffer.TextPosition, $"Expected postfix '{postfix}' for argument ({argNumber}:{arg}) in command '{commandName}'", typeName, commandName);
return(result);
}
}
else if (arg.IsOptional)
{
DoxygenToken argToken = DoxygenTokenPool.Make(DoxygenTokenKind.ArgumentCaption, Buffer.LexemeRange, false);
PushToken(argToken);
}
else if (arg.IsRequired)
{
AddError(Buffer.TextPosition, $"Expected prefix '{prefix}' for argument ({argNumber}:{arg}) in command '{commandName}'", typeName, commandName);
return(result);
}
}
break;
case DoxygenSyntax.ArgumentKind.MultipleObjectReference:
case DoxygenSyntax.ArgumentKind.SingleObjectReference:
{
// @TODO(final): ReferencedObject is not always a identifier
// Here are some examples of valid referenced objects:
// simple_identifier
// a_function()
// my::awesome::namespace::object
// my::awesome::namespace::function()
// my#awesome#namespace#function()
// method1,method2(),class#field
bool foundRef = false;
if (!noMoreArgs)
{
bool allowMultiple = arg.Kind == DoxygenSyntax.ArgumentKind.MultipleObjectReference;
bool requireIdent = true;
int referenceCount = 0;
while (!Buffer.IsEOF)
{
int oldPos = Buffer.StreamPosition;
char c0 = Buffer.Peek();
char c1 = Buffer.Peek(1);
if (!requireIdent)
{
if (c0 == ':' && c1 == ':')
{
Buffer.AdvanceColumns(2);
requireIdent = true;
continue;
}
else if (c0 == '#')
{
Buffer.AdvanceColumn();
requireIdent = true;
continue;
}
else if (c0 == ',' && referenceCount > 0 && allowMultiple)
{
Buffer.AdvanceColumn();
requireIdent = true;
continue;
}
else
{
// Correct termination of object-reference
foundRef = true;
break;
}
}
else
{
if (SyntaxUtils.IsIdentStart(c0))
{
requireIdent = false;
while (!Buffer.IsEOF)
{
if (!SyntaxUtils.IsIdentPart(Buffer.Peek()))
{
break;
}
Buffer.AdvanceColumn();
}
if (Buffer.Peek() == '(')
{
// Parse until right parent
Buffer.AdvanceColumn();
bool terminatedFunc = false;
while (!Buffer.IsEOF)
{
if (Buffer.Peek() == ')')
{
Buffer.AdvanceColumn();
terminatedFunc = true;
break;
}
Buffer.AdvanceAuto();
}
if (!terminatedFunc)
{
AddError(Buffer.TextPosition, $"Unterminated function reference for argument ({argNumber}:{arg}) in command '{commandName}'", typeName, commandName);
return(result);
}
}
++referenceCount;
continue;
}
else
{
AddError(Buffer.TextPosition, $"Requires identifier, but found '{Buffer.Peek()}' for argument ({argNumber}:{arg}) in command '{commandName}'", typeName, commandName);
return(result);
}
}
}
if (Buffer.IsEOF)
{
// Correct termination of object-reference when stream ends (Single-line)
foundRef = true;
}
}
if (arg.IsOptional || foundRef)
{
DoxygenToken argToken = DoxygenTokenPool.Make(DoxygenTokenKind.ArgumentIdent, Buffer.LexemeRange, foundRef);
PushToken(argToken);
}
else if (arg.IsRequired)
{
AddError(Buffer.TextPosition, $"Unexpected character '{Buffer.Peek()}' for argument ({argNumber}:{arg}) in command '{commandName}'", typeName, commandName);
return(result);
}
}
break;
case DoxygenSyntax.ArgumentKind.Identifier:
{
bool foundIdent = false;
// Special handling for @param command and ... parameter
if (!noMoreArgs && "param".Equals(commandName) && (arg.Kind == DoxygenSyntax.ArgumentKind.Identifier))
{
if (Buffer.Peek() == '.')
{
char c1 = Buffer.Peek(1);
char c2 = Buffer.Peek(2);
if (c1 == '.' && c2 == '.')
{
Buffer.AdvanceColumns(3);
foundIdent = true;
}
}
}
// We dont allow parsing a ident, when any special handling was matched
if (!noMoreArgs && !foundIdent && SyntaxUtils.IsIdentStart(Buffer.Peek()))
{
foundIdent = true;
while (!Buffer.IsEOF)
{
if (!SyntaxUtils.IsIdentPart(Buffer.Peek()))
{
break;
}
Buffer.AdvanceColumn();
}
}
if (arg.IsOptional || foundIdent)
{
DoxygenToken argToken = DoxygenTokenPool.Make(DoxygenTokenKind.ArgumentIdent, Buffer.LexemeRange, foundIdent);
PushToken(argToken);
}
else if (arg.IsRequired)
{
AddError(Buffer.TextPosition, $"Unexpected character '{Buffer.Peek()}' for argument ({argNumber}:{arg}) in command '{commandName}'", typeName, commandName);
return(result);
}
}
break;
case DoxygenSyntax.ArgumentKind.HeaderFile:
case DoxygenSyntax.ArgumentKind.HeaderName:
{
bool foundFilename = false;
if (!noMoreArgs)
{
bool requiredQuotes = arg.Kind == DoxygenSyntax.ArgumentKind.HeaderName;
char curChar = Buffer.Peek();
if (curChar == '<' || curChar == '\"')
{
char quoteChar = curChar == '<' ? '>' : '\"';
Buffer.AdvanceColumn();
while (!Buffer.IsEOF)
{
curChar = Buffer.Peek();
if (curChar == quoteChar)
{
Buffer.AdvanceColumn();
foundFilename = true;
break;
}
else if (SyntaxUtils.IsLineBreak(curChar))
{
break;
}
Buffer.AdvanceColumn();
}
if (!foundFilename)
{
AddError(Buffer.TextPosition, $"Unterminated filename, expect quote char '{quoteChar}' but got '{Buffer.Peek()}' for argument ({argNumber}:{arg}) in command '{commandName}'", typeName, commandName);
return(result);
}
}
else if (!requiredQuotes)
{
if (SyntaxUtils.IsFilename(Buffer.Peek()))
{
foundFilename = true;
while (!Buffer.IsEOF)
{
if (!SyntaxUtils.IsFilename(Buffer.Peek()))
{
break;
}
Buffer.AdvanceColumn();
}
}
}
}
if (arg.IsOptional || foundFilename)
{
DoxygenToken argToken = DoxygenTokenPool.Make(DoxygenTokenKind.ArgumentFile, Buffer.LexemeRange, foundFilename);
PushToken(argToken);
}
else if (arg.IsRequired)
{
AddError(Buffer.TextPosition, $"Unexpected character '{Buffer.Peek()}' for argument ({argNumber}:{arg}) in command '{commandName}'", typeName, commandName);
return(result);
}
}
break;
case DoxygenSyntax.ArgumentKind.SingleWord:
{
// @TODO(final): IsWordStart()
bool foundWord = false;
if (!noMoreArgs && char.IsLetterOrDigit(Buffer.Peek()))
{
foundWord = true;
while (!Buffer.IsEOF)
{
// @TODO(final): IsWordPart()
if (char.IsWhiteSpace(Buffer.Peek()))
{
break;
}
Buffer.AdvanceColumn();
}
}
if (arg.IsOptional || foundWord)
{
DoxygenToken argToken = DoxygenTokenPool.Make(DoxygenTokenKind.ArgumentCaption, Buffer.LexemeRange, foundWord);
PushToken(argToken);
}
else if (arg.IsRequired)
{
AddError(Buffer.TextPosition, $"Unexpected character '{Buffer.Peek()}' for argument ({argNumber}:{arg}) in command '{commandName}'", typeName, commandName);
return(result);
}
}
break;
case DoxygenSyntax.ArgumentKind.QuotedString:
{
bool isComplete = false;
// @TODO(final): Make quotes configurable in the argument rule
bool hasQuote = Buffer.Peek() == '"' || Buffer.Peek() == '<';
char endQuote = char.MaxValue;
if (hasQuote && !noMoreArgs)
{
endQuote = Buffer.Peek() == '<' ? '>' : '"';
Buffer.AdvanceColumn();
while (!Buffer.IsEOF)
{
if (!hasQuote)
{
if (char.IsWhiteSpace(Buffer.Peek()))
{
break;
}
}
else
{
if (Buffer.Peek() == endQuote)
{
Buffer.AdvanceColumn();
isComplete = true;
break;
}
else if (SyntaxUtils.IsLineBreak(Buffer.Peek()) || Buffer.Peek() == TextStream.InvalidCharacter)
{
break;
}
}
Buffer.AdvanceColumn();
}
if (!isComplete)
{
AddError(Buffer.TextPosition, $"Unterminated quote string for argument ({argNumber}:{arg}) in command '{commandName}'", typeName, commandName);
return(result);
}
}
if (arg.IsOptional || isComplete)
{
DoxygenToken argToken = DoxygenTokenPool.Make(DoxygenTokenKind.ArgumentText, Buffer.LexemeRange, isComplete);
PushToken(argToken);
}
else if (arg.IsRequired)
{
AddError(Buffer.TextPosition, $"Unexpected character '{Buffer.Peek()}' for argument ({argNumber}:{arg}) in command '{commandName}'", typeName, commandName);
return(result);
}
}
break;
case DoxygenSyntax.ArgumentKind.UntilEndOfLine:
{
bool eolFound = false;
if (!noMoreArgs)
{
while (!Buffer.IsEOF)
{
if (SyntaxUtils.IsLineBreak(Buffer.Peek()))
{
eolFound = true;
break;
}
Buffer.AdvanceColumn();
}
if (Buffer.IsEOF)
{
eolFound = true;
}
}
if (arg.IsOptional || eolFound)
{
DoxygenToken argToken = DoxygenTokenPool.Make(DoxygenTokenKind.ArgumentText, Buffer.LexemeRange, true);
PushToken(argToken);
}
else if (arg.IsRequired)
{
AddError(Buffer.TextPosition, $"Unterminated end-of-line for argument ({argNumber}:{arg}) in command '{commandName}'", typeName, commandName);
return(result);
}
}
break;
case DoxygenSyntax.ArgumentKind.ComplexLine:
case DoxygenSyntax.ArgumentKind.ComplexBlock:
// @TODO(final): Implement complex line/block properly
goto CommandDone;
default:
AddError(Buffer.TextPosition, $"Unsupported argument ({argNumber}:{arg}) in command '{commandName}'", typeName, commandName);
return(result);
}
// Postfix
if (!noMoreArgs && (hadPrefix && !string.IsNullOrWhiteSpace(postfix) && arg.Kind != DoxygenSyntax.ArgumentKind.PrefixToPostfix))
{
if (Buffer.CompareText(0, postfix) == 0)
{
Buffer.AdvanceColumns(prefix.Length);
}
else
{
AddError(Buffer.TextPosition, $"Expected postfix '{postfix}' for pp-argument({argNumber}:{arg}) in command '{commandName}'", typeName, commandName);
return(result);
}
}
++argNumber;
}
}
CommandDone:
result.IsValid = true;
return(result);
}
protected override bool LexNext(State hiddenState)
{
CppLexerState state = (CppLexerState)hiddenState;
bool allowWhitespaces = !state.IsInsidePreprocessor;
if (allowWhitespaces)
{
Buffer.SkipAllWhitespaces();
}
if (Buffer.IsEOF)
{
return(false);
}
int line = Buffer.TextPosition.Line;
char first = Buffer.Peek();
char second = Buffer.Peek(1);
char third = Buffer.Peek(2);
Buffer.StartLexeme();
LexResult lexRes = new LexResult(CppTokenKind.Unknown, true);
switch (first)
{
case '&':
{
if (second == '&')
{
lexRes.Kind = CppTokenKind.LogicalAndOp;
Buffer.AdvanceColumns(2);
}
else if (second == '=')
{
lexRes.Kind = CppTokenKind.AndAssign;
Buffer.AdvanceColumns(2);
}
else
{
lexRes.Kind = CppTokenKind.AndOp;
Buffer.AdvanceColumn();
}
}
break;
case '|':
{
if (second == '|')
{
lexRes.Kind = CppTokenKind.LogicalOrOp;
Buffer.AdvanceColumns(2);
}
else if (second == '=')
{
lexRes.Kind = CppTokenKind.OrAssign;
Buffer.AdvanceColumns(2);
}
else
{
lexRes.Kind = CppTokenKind.OrOp;
Buffer.AdvanceColumn();
}
}
break;
case '=':
{
if (second == '=')
{
lexRes.Kind = CppTokenKind.LogicalEqualsOp;
Buffer.AdvanceColumns(2);
}
else
{
lexRes.Kind = CppTokenKind.EqOp;
Buffer.AdvanceColumn();
}
}
break;
case '!':
{
if (second == '=')
{
lexRes.Kind = CppTokenKind.LogicalNotEqualsOp;
Buffer.AdvanceColumns(2);
}
else
{
lexRes.Kind = CppTokenKind.ExclationMark;
Buffer.AdvanceColumn();
}
}
break;
case '<':
{
if (second == '<')
{
if (third == '=')
{
lexRes.Kind = CppTokenKind.LeftShiftAssign;
Buffer.AdvanceColumns(3);
}
else
{
lexRes.Kind = CppTokenKind.LeftShiftOp;
Buffer.AdvanceColumns(2);
}
}
else if (second == '=')
{
lexRes.Kind = CppTokenKind.LessOrEqualOp;
Buffer.AdvanceColumns(2);
}
else
{
lexRes.Kind = CppTokenKind.LessThanOp;
Buffer.AdvanceColumn();
}
}
break;
case '>':
{
if (second == '>')
{
if (third == '=')
{
lexRes.Kind = CppTokenKind.RightShiftAssign;
Buffer.AdvanceColumns(3);
}
else
{
lexRes.Kind = CppTokenKind.RightShiftOp;
Buffer.AdvanceColumns(2);
}
}
else if (second == '=')
{
lexRes.Kind = CppTokenKind.GreaterOrEqualOp;
Buffer.AdvanceColumns(2);
}
else
{
lexRes.Kind = CppTokenKind.GreaterThanOp;
Buffer.AdvanceColumn();
}
}
break;
case '+':
{
if (second == '+')
{
lexRes.Kind = CppTokenKind.IncOp;
Buffer.AdvanceColumns(2);
}
else if (second == '=')
{
lexRes.Kind = CppTokenKind.AddAssign;
Buffer.AdvanceColumns(2);
}
else
{
lexRes.Kind = CppTokenKind.AddOp;
Buffer.AdvanceColumn();
}
}
break;
case '-':
{
if (second == '-')
{
lexRes.Kind = CppTokenKind.DecOp;
Buffer.AdvanceColumns(2);
}
else if (second == '=')
{
lexRes.Kind = CppTokenKind.SubAssign;
Buffer.AdvanceColumns(2);
}
else if (second == '>')
{
lexRes.Kind = CppTokenKind.PtrOp;
Buffer.AdvanceColumns(2);
}
else
{
lexRes.Kind = CppTokenKind.SubOp;
Buffer.AdvanceColumn();
}
}
break;
case '/':
{
if (second == '=')
{
lexRes.Kind = CppTokenKind.DivAssign;
Buffer.AdvanceColumns(2);
}
else if (second == '/')
{
lexRes = LexSingleLineComment(Buffer, true);
if (!lexRes.IsComplete)
{
AddError(Buffer.LexemeStart, $"Unterminated single-line comment, expect '\n' or '\r' but found '{Buffer.Peek()}'", lexRes.Kind.ToString());
}
}
else if (second == '*')
{
lexRes = LexMultiLineComment(Buffer, true);
if (!lexRes.IsComplete)
{
AddError(Buffer.LexemeStart, $"Unterminated single-line comment, expect '*/' but found '{Buffer.Peek()}'", lexRes.Kind.ToString());
}
}
else
{
Buffer.AdvanceColumn();
lexRes.Kind = CppTokenKind.DivOp;
}
}
break;
case '*':
{
if (second == '=')
{
lexRes.Kind = CppTokenKind.MulAssign;
Buffer.AdvanceColumns(2);
}
else
{
lexRes.Kind = CppTokenKind.MulOp;
Buffer.AdvanceColumn();
}
}
break;
case '%':
{
if (second == '=')
{
lexRes.Kind = CppTokenKind.ModAssign;
Buffer.AdvanceColumns(2);
}
else
{
lexRes.Kind = CppTokenKind.ModOp;
Buffer.AdvanceColumn();
}
}
break;
case '.':
{
if (second == '.' && third == '.')
{
lexRes.Kind = CppTokenKind.Ellipsis;
Buffer.AdvanceColumns(3);
}
else if (SyntaxUtils.IsNumeric(second))
{
lexRes = LexNumber();
}
else
{
lexRes.Kind = CppTokenKind.Dot;
Buffer.AdvanceColumn();
}
}
break;
case '^':
{
if (second == '=')
{
lexRes.Kind = CppTokenKind.XorAssign;
Buffer.AdvanceColumns(2);
}
else
{
lexRes.Kind = CppTokenKind.XorOp;
Buffer.AdvanceColumn();
}
}
break;
case '#':
return(LexPreprocessor(state));
case '"':
lexRes = LexString("string");
break;
case '\'':
lexRes = LexString("char");
break;
case '~':
lexRes.Kind = CppTokenKind.Tilde;
Buffer.AdvanceColumn();
break;
case '\\':
lexRes.Kind = CppTokenKind.Backslash;
Buffer.AdvanceColumn();
break;
case ',':
lexRes.Kind = CppTokenKind.Comma;
Buffer.AdvanceColumn();
break;
case ';':
lexRes.Kind = CppTokenKind.Semicolon;
Buffer.AdvanceColumn();
break;
case ':':
lexRes.Kind = CppTokenKind.Colon;
Buffer.AdvanceColumn();
break;
case '?':
lexRes.Kind = CppTokenKind.QuestionMark;
Buffer.AdvanceColumn();
break;
case '{':
lexRes.Kind = CppTokenKind.LeftBrace;
Buffer.AdvanceColumn();
break;
case '}':
lexRes.Kind = CppTokenKind.RightBrace;
Buffer.AdvanceColumn();
break;
case '[':
lexRes.Kind = CppTokenKind.LeftBracket;
Buffer.AdvanceColumn();
break;
case ']':
lexRes.Kind = CppTokenKind.RightBracket;
Buffer.AdvanceColumn();
break;
case '(':
lexRes.Kind = CppTokenKind.LeftParen;
Buffer.AdvanceColumn();
break;
case ')':
lexRes.Kind = CppTokenKind.RightParen;
Buffer.AdvanceColumn();
break;
default:
{
if (SyntaxUtils.IsLineBreak(first) && allowWhitespaces)
{
lexRes.Kind = CppTokenKind.EndOfLine;
int nb = SyntaxUtils.GetLineBreakChars(first, second);
Buffer.AdvanceLine(nb);
}
else if (first == '\t' && allowWhitespaces)
{
lexRes.Kind = CppTokenKind.Spacings;
while (!Buffer.IsEOF)
{
if (Buffer.Peek() != '\t')
{
break;
}
Buffer.AdvanceTab();
}
}
else if (SyntaxUtils.IsSpacing(first) && allowWhitespaces)
{
lexRes.Kind = CppTokenKind.Spacings;
Buffer.AdvanceColumnsWhile(SyntaxUtils.IsSpacing);
}
else if (SyntaxUtils.IsIdentStart(first))
{
Debug.Assert(!state.IsInsidePreprocessor);
lexRes = LexIdent(false);
}
else if (SyntaxUtils.IsNumeric(first))
{
lexRes = LexNumber();
}
else
{
AddError(Buffer.TextPosition, $"Unexpected character '{first}'", "Character");
return(false);
}
}
break;
}
return(PushToken(CppTokenPool.Make(lexRes.Kind, Buffer.LexemeRange, lexRes.IsComplete)));
}
private LexResult LexNumber()
{
Debug.Assert(SyntaxUtils.IsNumeric(Buffer.Peek()) || Buffer.Peek() == '.');
CppTokenKind kind;
char first = Buffer.Peek(0);
char second = Buffer.Peek(1);
bool dotSeen = false;
if (first == '0')
{
if (second == 'x' || second == 'X')
{
// Hex
kind = CppTokenKind.HexLiteral;
Buffer.AdvanceColumns(2); // Skip 0[xX]
}
else if (second == 'b' || second == 'B')
{
// Binary
kind = CppTokenKind.BinaryLiteral;
Buffer.AdvanceColumns(2); // Skip 0[bB]
}
else
{
// Octal
kind = CppTokenKind.OctalLiteral;
}
}
else if (first == '.')
{
Debug.Assert(SyntaxUtils.IsNumeric(second));
kind = CppTokenKind.IntegerFloatLiteral;
Buffer.AdvanceColumn();
dotSeen = true;
}
else
{
Debug.Assert(SyntaxUtils.IsNumeric(first));
kind = CppTokenKind.IntegerLiteral;
}
// @NOTE(final): We never set the DecimalHexLiteral kind initially,
// as every hex decimal always starts as a normal hex literal!
Debug.Assert(kind != CppTokenKind.HexadecimalFloatLiteral);
// First number part
int firstLiteralPos = Buffer.TextPosition.Index;
bool readNextLiteral = false;
do
{
readNextLiteral = false;
int s = Buffer.TextPosition.Index;
switch (kind)
{
case CppTokenKind.IntegerLiteral:
case CppTokenKind.IntegerFloatLiteral:
if (SyntaxUtils.IsNumeric(Buffer.Peek()))
{
Buffer.AdvanceColumnsWhile(SyntaxUtils.IsNumeric);
}
else
{
AddError(Buffer.TextPosition, $"Expect integer literal, but got '{Buffer.Peek()}'", kind.ToString());
}
break;
case CppTokenKind.OctalLiteral:
if (SyntaxUtils.IsOctal(Buffer.Peek()))
{
Buffer.AdvanceColumnsWhile(SyntaxUtils.IsOctal);
}
else
{
AddError(Buffer.TextPosition, $"Expect octal literal, but got '{Buffer.Peek()}'", kind.ToString());
}
break;
case CppTokenKind.HexLiteral:
if (SyntaxUtils.IsHex(Buffer.Peek()))
{
Buffer.AdvanceColumnsWhile(SyntaxUtils.IsHex);
}
else
{
AddError(Buffer.TextPosition, $"Expect hex literal, but got '{Buffer.Peek()}'", kind.ToString());
}
break;
case CppTokenKind.BinaryLiteral:
if (SyntaxUtils.IsBinary(Buffer.Peek()))
{
Buffer.AdvanceColumnsWhile(SyntaxUtils.IsBinary);
}
else
{
AddError(Buffer.TextPosition, $"Expect binary literal, but got '{Buffer.Peek()}'", kind.ToString());
}
break;
default:
AddError(Buffer.TextPosition, $"Unsupported token kind '{kind}' for integer literal on {Buffer}", kind.ToString());
break;
}
bool hadIntegerLiteral = Buffer.TextPosition.Index > s;
if (kind != CppTokenKind.IntegerFloatLiteral && kind != CppTokenKind.HexadecimalFloatLiteral)
{
// @NOTE(final): Single quotes (') are allowed as separators for any non-decimal literal
char check0 = Buffer.Peek();
if (check0 == '\'')
{
if (!hadIntegerLiteral)
{
AddError(Buffer.TextPosition, $"Too many single quote escape in integer literal, expect any integer literal but got '{Buffer.Peek()}'", kind.ToString());
return(new LexResult(kind, false));
}
Buffer.AdvanceColumn();
readNextLiteral = true;
}
}
} while (!Buffer.IsEOF && readNextLiteral);
// Validate any literal after starting dot
if (dotSeen)
{
if (firstLiteralPos == Buffer.TextPosition.Index)
{
AddError(Buffer.TextPosition, $"Expect any integer literal after starting dot, but got '{Buffer.Peek()}'", kind.ToString());
return(new LexResult(kind, false));
}
}
// Dot separator
if ((!dotSeen) &&
((kind == CppTokenKind.IntegerLiteral) ||
(kind == CppTokenKind.HexLiteral) ||
(kind == CppTokenKind.OctalLiteral)
))
{
char check0 = Buffer.Peek();
if (check0 == '.')
{
dotSeen = true;
Buffer.AdvanceColumn();
if (kind == CppTokenKind.IntegerLiteral || kind == CppTokenKind.OctalLiteral)
{
kind = CppTokenKind.IntegerFloatLiteral;
}
else
{
Debug.Assert(kind == CppTokenKind.HexLiteral);
kind = CppTokenKind.HexadecimalFloatLiteral;
}
}
else if (SyntaxUtils.IsExponentPrefix(check0))
{
if (kind == CppTokenKind.IntegerLiteral || kind == CppTokenKind.OctalLiteral)
{
kind = CppTokenKind.IntegerFloatLiteral;
}
else
{
Debug.Assert(kind == CppTokenKind.HexLiteral);
kind = CppTokenKind.HexadecimalFloatLiteral;
}
}
}
// Decimal after dot separator
if ((kind != CppTokenKind.IntegerFloatLiteral) &&
(kind != CppTokenKind.HexadecimalFloatLiteral))
{
// Integer suffix
if (SyntaxUtils.IsIntegerSuffix(Buffer.Peek()))
{
Buffer.AdvanceColumnsWhile(SyntaxUtils.IsIntegerSuffix, 3);
}
}
else
{
if (kind == CppTokenKind.IntegerFloatLiteral)
{
// Float decimal
if (SyntaxUtils.IsNumeric(Buffer.Peek()))
{
Buffer.AdvanceColumnsWhile(SyntaxUtils.IsNumeric);
}
if (Buffer.Peek() == 'e' || Buffer.Peek() == 'E')
{
AdvanceExponent('e');
}
}
else
{
// Hex decimal
Debug.Assert(kind == CppTokenKind.HexadecimalFloatLiteral);
if (SyntaxUtils.IsHex(Buffer.Peek()))
{
Buffer.AdvanceColumnsWhile(SyntaxUtils.IsHex);
}
if (Buffer.Peek() == 'p' || Buffer.Peek() == 'P')
{
AdvanceExponent('e');
}
}
// Float suffix
if (SyntaxUtils.IsFloatSuffix(Buffer.Peek()))
{
Buffer.AdvanceColumn();
}
}
return(new LexResult(kind, true));
}