public void ConsumeToAny()
{
CharacterReader r = new CharacterReader("One &bar; qux");
Assert.AreEqual("One ", r.ConsumeToAny('&', ';'));
Assert.IsTrue(r.Matches('&'));
Assert.IsTrue(r.Matches("&bar;"));
Assert.AreEqual('&', r.Consume());
Assert.AreEqual("bar", r.ConsumeToAny('&', ';'));
Assert.AreEqual(';', r.Consume());
Assert.AreEqual(" qux", r.ConsumeToAny('&', ';'));
}
// in data state, gather characters until a character reference or tag is found
internal override void Read(Tokeniser t, CharacterReader r)
{
switch (r.Current())
{
case '&':
t.AdvanceTransition(TokeniserState.CharacterReferenceInData);
break;
case '<':
t.AdvanceTransition(TokeniserState.TagOpen);
break;
case TokeniserState.nullChar:
t.Error(this);
// NOT replacement character (oddly?)
t.Emit(r.Consume());
break;
case TokeniserState.eof:
t.Emit(new Token.EOF());
break;
default:
string data = r.ConsumeToAny('&', '<', TokeniserState.nullChar);
t.Emit(data);
break;
}
}
public void MatchesAny()
{
char[] scan = { ' ', '\n', '\t' };
CharacterReader r = new CharacterReader("One\nTwo\tThree");
Assert.IsFalse(r.MatchesAny(scan));
Assert.AreEqual("One", r.ConsumeToAny(scan));
Assert.IsTrue(r.MatchesAny(scan));
Assert.AreEqual('\n', r.Consume());
Assert.IsFalse(r.MatchesAny(scan));
}
internal override void Read(Tokeniser t, CharacterReader r)
{
char c = r.Current();
switch (c)
{
case '-':
t.Emit(c);
t.AdvanceTransition(TokeniserState.ScriptDataDoubleEscapedDash);
break;
case '<':
t.Emit(c);
t.AdvanceTransition(TokeniserState.ScriptDataDoubleEscapedLessthanSign
);
break;
case TokeniserState.nullChar:
t.Error(this);
r.Advance();
t.Emit(TokeniserState.replacementChar);
break;
case TokeniserState.eof:
t.EofError(this);
t.Transition(TokeniserState.Data);
break;
default:
string data = r.ConsumeToAny('-', '<', TokeniserState.nullChar);
t.Emit(data);
break;
}
}
/// handles data in title, textarea etc
internal override void Read(Tokeniser t, CharacterReader r)
{
switch (r.Current())
{
case '&':
t.AdvanceTransition(TokeniserState.CharacterReferenceInRcdata);
break;
case '<':
t.AdvanceTransition(TokeniserState.RcdataLessthanSign);
break;
case TokeniserState.nullChar:
t.Error(this);
r.Advance();
t.Emit(TokeniserState.replacementChar);
break;
case TokeniserState.eof:
t.Emit(new Token.EOF());
break;
default:
string data = r.ConsumeToAny('&', '<', TokeniserState.nullChar);
t.Emit(data);
break;
}
}
// from < or </ in data, will have start or end tag pending
internal override void Read(Tokeniser t, CharacterReader r)
{
// previous TagOpen state did NOT consume, will have a letter char in current
string tagName = r.ConsumeToAny('\t', '\n', '\r', '\f', ' ', '/', '>', TokeniserState.nullChar).ToLower();
t.tagPending.AppendTagName(tagName);
switch (r.Consume())
{
case '\t':
case '\n':
case '\r':
case '\f':
case ' ':
t.Transition(TokeniserState.BeforeAttributeName);
break;
case '/':
t.Transition(TokeniserState.SelfClosingStartTag);
break;
case '>':
t.EmitTagPending();
t.Transition(TokeniserState.Data);
break;
case TokeniserState.nullChar:
// replacement
t.tagPending.AppendTagName(TokeniserState.replacementStr);
break;
case TokeniserState.eof:
// should emit pending tag?
t.EofError(this);
t.Transition(TokeniserState.Data);
break;
}
}
internal override void Read(Tokeniser t, CharacterReader r)
{
char c = r.Current();
switch (c)
{
case '-':
t.AdvanceTransition(TokeniserState.CommentEndDash);
break;
case TokeniserState.nullChar:
t.Error(this);
r.Advance();
t.commentPending.data.Append(TokeniserState.replacementChar);
break;
case TokeniserState.eof:
t.EofError(this);
t.EmitCommentPending();
t.Transition(TokeniserState.Data);
break;
default:
t.commentPending.data.Append(r.ConsumeToAny('-', TokeniserState.nullChar
));
break;
}
}
// no default, handled in consume to any above
internal override void Read(Tokeniser t, CharacterReader r)
{
string value = r.ConsumeToAny('\t', '\n', '\r', '\f', ' ', '&', '>', TokeniserState.nullChar, '"', '\'', '<', '=', '`');
if (value.Length > 0)
{
t.tagPending.AppendAttributeValue(value);
}
char c = r.Consume();
switch (c)
{
case '\t':
case '\n':
case '\r':
case '\f':
case ' ':
t.Transition(TokeniserState.BeforeAttributeName);
break;
case '&':
char[] @ref = t.ConsumeCharacterReference('>', true);
if (@ref != null)
{
t.tagPending.AppendAttributeValue(@ref);
}
else
{
t.tagPending.AppendAttributeValue('&');
}
break;
case '>':
t.EmitTagPending();
t.Transition(TokeniserState.Data);
break;
case TokeniserState.nullChar:
t.Error(this);
t.tagPending.AppendAttributeValue(TokeniserState.replacementChar);
break;
case TokeniserState.eof:
t.EofError(this);
t.Transition(TokeniserState.Data);
break;
case '"':
case '\'':
case '<':
case '=':
case '`':
t.Error(this);
t.tagPending.AppendAttributeValue(c);
break;
}
}
// no default, handled in consume to any above
internal override void Read(Tokeniser t, CharacterReader r)
{
string value = r.ConsumeToAny('\'', '&', TokeniserState.nullChar);
if (value.Length > 0)
{
t.tagPending.AppendAttributeValue(value);
}
char c = r.Consume();
switch (c)
{
case '\'':
t.Transition(TokeniserState.AfterAttributeValue_quoted);
break;
case '&':
char[] @ref = t.ConsumeCharacterReference('\'', true);
if (@ref != null)
{
t.tagPending.AppendAttributeValue(@ref);
}
else
{
t.tagPending.AppendAttributeValue('&');
}
break;
case TokeniserState.nullChar:
t.Error(this);
t.tagPending.AppendAttributeValue(TokeniserState.replacementChar);
break;
case TokeniserState.eof:
t.EofError(this);
t.Transition(TokeniserState.Data);
break;
}
}
// from before attribute name
internal override void Read(Tokeniser t, CharacterReader r)
{
string name = r.ConsumeToAny('\t', '\n', '\r', '\f', ' ', '/', '=', '>', TokeniserState.nullChar, '"', '\'', '<');
t.tagPending.AppendAttributeName(name.ToLower());
char c = r.Consume();
switch (c)
{
case '\t':
case '\n':
case '\r':
case '\f':
case ' ':
t.Transition(TokeniserState.AfterAttributeName);
break;
case '/':
t.Transition(TokeniserState.SelfClosingStartTag);
break;
case '=':
t.Transition(TokeniserState.BeforeAttributeValue);
break;
case '>':
t.EmitTagPending();
t.Transition(TokeniserState.Data);
break;
case TokeniserState.nullChar:
t.Error(this);
t.tagPending.AppendAttributeName(TokeniserState.replacementChar);
break;
case TokeniserState.eof:
t.EofError(this);
t.Transition(TokeniserState.Data);
break;
case '"':
case '\'':
case '<':
t.Error(this);
t.tagPending.AppendAttributeName(c);
break;
}
}