/**
* Does the actual parsing. Perform this immediately
* after creating the parser object.
*/
private void Go(TextReader reader)
{
doc.StartDocument();
while (true)
{
// read a new character
if (previousCharacter == -1)
{
character = reader.Read();
}
// or re-examin the previous character
else
{
character = previousCharacter;
previousCharacter = -1;
}
// the end of the file was reached
if (character == -1)
{
if (html)
{
if (html && state == TEXT)
{
Flush();
}
doc.EndDocument();
}
else
{
ThrowException(MessageLocalization.GetComposedMessage("missing.end.tag"));
}
return;
}
// dealing with \n and \r
if (character == '\n' && eol)
{
eol = false;
continue;
}
else if (eol)
{
eol = false;
}
else if (character == '\n')
{
lines++;
columns = 0;
}
else if (character == '\r')
{
eol = true;
character = '\n';
lines++;
columns = 0;
}
else
{
columns++;
}
switch (state)
{
// we are in an unknown state before there's actual content
case UNKNOWN:
if (character == '<')
{
SaveState(TEXT);
state = TAG_ENCOUNTERED;
}
break;
// we can encounter any content
case TEXT:
if (character == '<')
{
Flush();
SaveState(state);
state = TAG_ENCOUNTERED;
}
else if (character == '&')
{
SaveState(state);
entity.Length = 0;
state = ENTITY;
nowhite = true;
}
else if (character == ' ')
{
if (html && nowhite)
{
text.Append(' ');
nowhite = false;
}
else
{
if (nowhite)
{
text.Append((char)character);
}
nowhite = false;
}
}
else if (Char.IsWhiteSpace((char)character))
{
if (html)
{
// totally ignore other whitespace
}
else
{
if (nowhite)
{
text.Append((char)character);
}
nowhite = false;
}
}
else
{
text.Append((char)character);
nowhite = true;
}
break;
// we have just seen a < and are wondering what we are looking at
// <foo>, </foo>, <!-- ... --->, etc.
case TAG_ENCOUNTERED:
InitTag();
if (character == '/')
{
state = IN_CLOSETAG;
}
else if (character == '?')
{
RestoreState();
state = PI;
}
else
{
text.Append((char)character);
state = EXAMIN_TAG;
}
break;
// we are processing something like this <foo ... >.
// It could still be a <!-- ... --> or something.
case EXAMIN_TAG:
if (character == '>')
{
DoTag();
ProcessTag(true);
InitTag();
state = RestoreState();
}
else if (character == '/')
{
state = SINGLE_TAG;
}
else if (character == '-' && text.ToString().Equals("!-"))
{
Flush();
state = COMMENT;
}
else if (character == '[' && text.ToString().Equals("![CDATA"))
{
Flush();
state = CDATA;
}
else if (character == 'E' && text.ToString().Equals("!DOCTYP"))
{
Flush();
state = PI;
}
else if (char.IsWhiteSpace((char)character))
{
DoTag();
state = TAG_EXAMINED;
}
else
{
text.Append((char)character);
}
break;
// we know the name of the tag now.
case TAG_EXAMINED:
if (character == '>')
{
ProcessTag(true);
InitTag();
state = RestoreState();
}
else if (character == '/')
{
state = SINGLE_TAG;
}
else if (char.IsWhiteSpace((char)character))
{
// empty
}
else
{
text.Append((char)character);
state = ATTRIBUTE_KEY;
}
break;
// we are processing a closing tag: e.g. </foo>
case IN_CLOSETAG:
if (character == '>')
{
DoTag();
ProcessTag(false);
if (!html && nested == 0)
{
return;
}
state = RestoreState();
}
else
{
if (!char.IsWhiteSpace((char)character))
{
text.Append((char)character);
}
}
break;
// we have just seen something like this: <foo a="b"/
// and are looking for the final >.
case SINGLE_TAG:
if (character != '>')
{
ThrowException(MessageLocalization.GetComposedMessage("expected.gt.for.tag.lt.1.gt", tag));
}
DoTag();
ProcessTag(true);
ProcessTag(false);
InitTag();
if (!html && nested == 0)
{
doc.EndDocument();
return;
}
state = RestoreState();
break;
// we are processing CDATA
case CDATA:
if (character == '>' &&
text.ToString().EndsWith("]]"))
{
text.Length = text.Length - 2;
Flush();
state = RestoreState();
}
else
{
text.Append((char)character);
}
break;
// we are processing a comment. We are inside
// the <!-- .... --> looking for the -->.
case COMMENT:
if (character == '>' &&
text.ToString().EndsWith("--"))
{
text.Length = text.Length - 2;
Flush();
state = RestoreState();
}
else
{
text.Append((char)character);
}
break;
// We are inside one of these <? ... ?> or one of these <!DOCTYPE ... >
case PI:
if (character == '>')
{
state = RestoreState();
if (state == TEXT)
{
state = UNKNOWN;
}
}
break;
// we are processing an entity, e.g. <, », etc.
case ENTITY:
if (character == ';')
{
state = RestoreState();
String cent = entity.ToString();
entity.Length = 0;
char ce = EntitiesToUnicode.DecodeEntity(cent);
if (ce == '\0')
{
text.Append('&').Append(cent).Append(';');
}
else
{
text.Append(ce);
}
}
else if ((character != '#' && (character < '0' || character > '9') && (character < 'a' || character > 'z') &&
(character < 'A' || character > 'Z')) || entity.Length >= 7)
{
state = RestoreState();
previousCharacter = character;
text.Append('&').Append(entity.ToString());
entity.Length = 0;
}
else
{
entity.Append((char)character);
}
break;
// We are processing the quoted right-hand side of an element's attribute.
case QUOTE:
if (html && quoteCharacter == ' ' && character == '>')
{
Flush();
ProcessTag(true);
InitTag();
state = RestoreState();
}
else if (html && quoteCharacter == ' ' && char.IsWhiteSpace((char)character))
{
Flush();
state = TAG_EXAMINED;
}
else if (html && quoteCharacter == ' ')
{
text.Append((char)character);
}
else if (character == quoteCharacter)
{
Flush();
state = TAG_EXAMINED;
}
else if (" \r\n\u0009".IndexOf((char)character) >= 0)
{
text.Append(' ');
}
else if (character == '&')
{
SaveState(state);
state = ENTITY;
entity.Length = 0;
}
else
{
text.Append((char)character);
}
break;
case ATTRIBUTE_KEY:
if (char.IsWhiteSpace((char)character))
{
Flush();
state = ATTRIBUTE_EQUAL;
}
else if (character == '=')
{
Flush();
state = ATTRIBUTE_VALUE;
}
else if (html && character == '>')
{
text.Length = 0;
ProcessTag(true);
InitTag();
state = RestoreState();
}
else
{
text.Append((char)character);
}
break;
case ATTRIBUTE_EQUAL:
if (character == '=')
{
state = ATTRIBUTE_VALUE;
}
else if (char.IsWhiteSpace((char)character))
{
// empty
}
else if (html && character == '>')
{
text.Length = 0;
ProcessTag(true);
InitTag();
state = RestoreState();
}
else if (html && character == '/')
{
Flush();
state = SINGLE_TAG;
}
else if (html)
{
Flush();
text.Append((char)character);
state = ATTRIBUTE_KEY;
}
else
{
ThrowException(MessageLocalization.GetComposedMessage("error.in.attribute.processing"));
}
break;
case ATTRIBUTE_VALUE:
if (character == '"' || character == '\'')
{
quoteCharacter = character;
state = QUOTE;
}
else if (char.IsWhiteSpace((char)character))
{
// empty
}
else if (html && character == '>')
{
Flush();
ProcessTag(true);
InitTag();
state = RestoreState();
}
else if (html)
{
text.Append((char)character);
quoteCharacter = ' ';
state = QUOTE;
}
else
{
ThrowException(MessageLocalization.GetComposedMessage("error.in.attribute.processing"));
}
break;
}
}
}
/// <summary>
/// Does the actual parsing. Perform this immediately
/// after creating the parser object.
/// </summary>
private void go(TextReader reader)
{
Doc.StartDocument();
while (true)
{
// read a new character
if (PreviousCharacter == -1)
{
Character = reader.Read();
}
// or re-examin the previous character
else
{
Character = PreviousCharacter;
PreviousCharacter = -1;
}
// the end of the file was reached
if (Character == -1)
{
if (Html)
{
if (Html && State == Text)
{
flush();
}
Doc.EndDocument();
}
else
{
throwException("Missing end tag");
}
return;
}
// dealing with \n and \r
if (Character == '\n' && Eol)
{
Eol = false;
continue;
}
else if (Eol)
{
Eol = false;
}
else if (Character == '\n')
{
Lines++;
Columns = 0;
}
else if (Character == '\r')
{
Eol = true;
Character = '\n';
Lines++;
Columns = 0;
}
else
{
Columns++;
}
switch (State)
{
// we are in an unknown state before there's actual content
case Unknown:
if (Character == '<')
{
saveState(Text);
State = TagEncountered;
}
break;
// we can encounter any content
case Text:
if (Character == '<')
{
flush();
saveState(State);
State = TagEncountered;
}
else if (Character == '&')
{
saveState(State);
entity.Length = 0;
State = Entity;
}
else if (char.IsWhiteSpace((char)Character))
{
if (Nowhite)
{
text.Append((char)Character);
}
Nowhite = false;
}
else
{
text.Append((char)Character);
Nowhite = true;
}
break;
// we have just seen a < and are wondering what we are looking at
// <foo>, </foo>, <!-- ... --->, etc.
case TagEncountered:
initTag();
if (Character == '/')
{
State = InClosetag;
}
else if (Character == '?')
{
restoreState();
State = Pi;
}
else
{
text.Append((char)Character);
State = ExaminTag;
}
break;
// we are processing something like this <foo ... >.
// It could still be a <!-- ... --> or something.
case ExaminTag:
if (Character == '>')
{
doTag();
processTag(true);
initTag();
State = restoreState();
}
else if (Character == '/')
{
State = SingleTag;
}
else if (Character == '-' && text.ToString().Equals("!-"))
{
flush();
State = Comment;
}
else if (Character == '[' && text.ToString().Equals("![CDATA"))
{
flush();
State = Cdata;
}
else if (Character == 'E' && text.ToString().Equals("!DOCTYP"))
{
flush();
State = Pi;
}
else if (char.IsWhiteSpace((char)Character))
{
doTag();
State = TagExamined;
}
else
{
text.Append((char)Character);
}
break;
// we know the name of the tag now.
case TagExamined:
if (Character == '>')
{
processTag(true);
initTag();
State = restoreState();
}
else if (Character == '/')
{
State = SingleTag;
}
else if (char.IsWhiteSpace((char)Character))
{
// empty
}
else
{
text.Append((char)Character);
State = AttributeKey;
}
break;
// we are processing a closing tag: e.g. </foo>
case InClosetag:
if (Character == '>')
{
doTag();
processTag(false);
if (!Html && Nested == 0)
{
return;
}
State = restoreState();
}
else
{
if (!char.IsWhiteSpace((char)Character))
{
text.Append((char)Character);
}
}
break;
// we have just seen something like this: <foo a="b"/
// and are looking for the final >.
case SingleTag:
if (Character != '>')
{
throwException($"Expected > for tag: <{Tag}/>");
}
doTag();
processTag(true);
processTag(false);
initTag();
if (!Html && Nested == 0)
{
Doc.EndDocument();
return;
}
State = restoreState();
break;
// we are processing CDATA
case Cdata:
if (Character == '>' &&
text.ToString().EndsWith("]]"))
{
text.Length = text.Length - 2;
flush();
State = restoreState();
}
else
{
text.Append((char)Character);
}
break;
// we are processing a comment. We are inside
// the <!-- .... --> looking for the -->.
case Comment:
if (Character == '>' &&
text.ToString().EndsWith("--"))
{
text.Length = text.Length - 2;
flush();
State = restoreState();
}
else
{
text.Append((char)Character);
}
break;
// We are inside one of these <? ... ?> or one of these <!DOCTYPE ... >
case Pi:
if (Character == '>')
{
State = restoreState();
if (State == Text)
{
State = Unknown;
}
}
break;
// we are processing an entity, e.g. <, », etc.
case Entity:
if (Character == ';')
{
State = restoreState();
var cent = entity.ToString();
entity.Length = 0;
var ce = EntitiesToUnicode.DecodeEntity(cent);
if (ce == '\0')
{
text.Append('&').Append(cent).Append(';');
}
else
{
text.Append(ce);
}
}
else if ((Character != '#' && (Character < '0' || Character > '9') && (Character < 'a' || Character > 'z') &&
(Character < 'A' || Character > 'Z')) || entity.Length >= 7)
{
State = restoreState();
PreviousCharacter = Character;
text.Append('&').Append(entity);
entity.Length = 0;
}
else
{
entity.Append((char)Character);
}
break;
// We are processing the quoted right-hand side of an element's attribute.
case Quote:
if (Html && QuoteCharacter == ' ' && Character == '>')
{
flush();
processTag(true);
initTag();
State = restoreState();
}
else if (Html && QuoteCharacter == ' ' && char.IsWhiteSpace((char)Character))
{
flush();
State = TagExamined;
}
else if (Html && QuoteCharacter == ' ')
{
text.Append((char)Character);
}
else if (Character == QuoteCharacter)
{
flush();
State = TagExamined;
}
else if (" \r\n\u0009".IndexOf(((char)Character).ToString(), StringComparison.Ordinal) >= 0)
{
text.Append(' ');
}
else if (Character == '&')
{
saveState(State);
State = Entity;
entity.Length = 0;
}
else
{
text.Append((char)Character);
}
break;
case AttributeKey:
if (char.IsWhiteSpace((char)Character))
{
flush();
State = AttributeEqual;
}
else if (Character == '=')
{
flush();
State = AttributeValue;
}
else if (Html && Character == '>')
{
text.Length = 0;
processTag(true);
initTag();
State = restoreState();
}
else
{
text.Append((char)Character);
}
break;
case AttributeEqual:
if (Character == '=')
{
State = AttributeValue;
}
else if (char.IsWhiteSpace((char)Character))
{
// empty
}
else if (Html && Character == '>')
{
text.Length = 0;
processTag(true);
initTag();
State = restoreState();
}
else if (Html && Character == '/')
{
flush();
State = SingleTag;
}
else if (Html)
{
flush();
text.Append((char)Character);
State = AttributeKey;
}
else
{
throwException("Error in attribute processing.");
}
break;
case AttributeValue:
if (Character == '"' || Character == '\'')
{
QuoteCharacter = Character;
State = Quote;
}
else if (char.IsWhiteSpace((char)Character))
{
// empty
}
else if (Html && Character == '>')
{
flush();
processTag(true);
initTag();
State = restoreState();
}
else if (Html)
{
text.Append((char)Character);
QuoteCharacter = ' ';
State = Quote;
}
else
{
throwException("Error in attribute processing");
}
break;
}
}
}
