public override void ParseChunk(Document document, char[] block, Action<Comment> onComment)
{
int index = 0;
int len = block.Length;
int state = document.State;
#if USE_UNSAFE
// nice work, buffer overruns are now possible
// are you proud of yourself?
unsafe
{
fixed (char* buffer = block)
{
#else
char[] buffer = block;
#endif
for (; index <= len; index++)
{
// the bounds checks are real.
Char c = buffer[index];
switch (state)
{
case PARSING:
switch (c)
{
case '/': state = AWAITING_COMMENT_COMPLETION; break;
case '"': state = STRING_LITERAL; break;
case '\'': state = CHAR_LITERAL; break;
case '@': state = AWAITING_VERBATIM_STRING; break;
case '\n': document.IncrementLineCount(); break;
default: break; // still parsing
} break;
case STRING_LITERAL:
if (c == '"') state = PARSING;// close out the string and continue.
else if (c == '\\') state = LITERAL_ESCAPE_SEQUENCE;
break;
case AWAITING_COMMENT_COMPLETION:
switch (c)
{
case '/':
state = INLINE_COMMENT;
document.MarkCommentBegin();
break;
case '*':
state = BLOCK_COMMENT;
document.MarkCommentBegin();
break;
default: state = PARSING; break;
} break;
case BLOCK_COMMENT:
if (c == '*')
{
state = AWAITING_BLOCK_COMMENT_COMPLETION;
}
else
{
document.AppendChar(c);
}
break;
case AWAITING_BLOCK_COMMENT_COMPLETION:
if (c == '/')
{
onComment(document.CloseComment());
state = PARSING;
}
else
{
// we have no completed the block.
document.AppendChar('*');
document.AppendChar(c);
state = BLOCK_COMMENT;
}
break;
case INLINE_COMMENT:
{
if (c == '\n')
{
onComment(document.CloseComment());
document.IncrementLineCount();
state = PARSING;
}
else
{
document.AppendChar(c);
}
break;
}
case LITERAL_ESCAPE_SEQUENCE:
{
// what was escaped? do we actually care?
// doesn't handle
state = STRING_LITERAL;
break;
}
case AWAITING_VERBATIM_STRING:
{
if (c == '"') state = VERBATIM_STRING;
else state = PARSING;
break;
}
case VERBATIM_STRING:
switch (c)
{
case '"': state = VERBATIM_ESCAPE_QUOTE; break;
case '\n': document.IncrementLineCount(); break;
default: break;
} break;
case VERBATIM_ESCAPE_QUOTE:
if (c == '"')
state = VERBATIM_STRING;
else state = PARSING;
// in this case though, should we increment?
break;
case CHAR_LITERAL:
switch (c)
{
case '\'': state = PARSING; break;
case '\\': state = CHAR_LITERAL_ESCAPE; break;
} break;
case CHAR_LITERAL_ESCAPE:
// this is a more complicated case, but I don't really
// feel like implementing this.
state = CHAR_LITERAL;
break;
default: throw new ApplicationException();
}
}
document.State = state;
#if USE_UNSAFE
}
}
#endif
}
