Element Parser::Parse(std::istream& in)
{
std::vector<Element> lexed_elements;
LexResult lex_result(Lex(in, lexed_elements));
strine::Element element;
if (lex_result.Success())
{
TokenStream token_stream(lexed_elements);
strine::Imperative imperative;
while(token_stream.HasTokens())
{
bool success = ParseAny( token_stream, element );
if (false == success)
{
Element token = token_stream.NextToken();
SourceLocation const& location = token.GetSourceLocation();
std::stringstream ss;
ss << "parser("
<< location.GetRow()
<< "): Unrecognized start \""
<< token.ToString()
<< "\"";
Error error(ss.str());
error.SetRow(location.GetRow());
error.SetColumn(location.GetColumn());
// Okay, not sure if I want to mess with this yet.
// This means that we received some un-intelligeable code.
return error;
}
else if (token_stream.HasTokens())
{
// If we have more tokens, then we start doing something
// fun, like putting into an imperative.
imperative.Add(element);
}
else
{
if (imperative.NumberOfElements() > 0)
{
imperative.Add(element);
}
}
}
if (imperative.NumberOfElements() > 0)
{
// Set the element to be returned to this imperative
// element.
element = imperative;
}
}
return element;
}
bool XSDParser::ParseContent(DTDElement& node, bool extended /*=false*/)
{
DTDElement::EType curr_type;
int emb=0;
bool eatEOT= false;
bool hasContents= false;
TToken tok;
for ( tok=GetNextToken(); ; tok=GetNextToken()) {
emb= node.GetContent().size();
if (tok != T_EOF &&
tok != K_ENDOFTAG &&
tok != K_ANNOTATION) {
hasContents= true;
}
switch (tok) {
case T_EOF:
return hasContents;
case K_ENDOFTAG:
if (eatEOT) {
eatEOT= false;
break;
}
FixEmbeddedNames(node);
return hasContents;
case K_COMPLEXTYPE:
ParseComplexType(node);
break;
case K_SIMPLECONTENT:
ParseSimpleContent(node);
break;
case K_EXTENSION:
ParseExtension(node);
break;
case K_RESTRICTION:
ParseRestriction(node);
break;
case K_ATTRIBUTE:
ParseAttribute(node);
break;
case K_ATTRIBUTEGROUP:
ParseAttributeGroup(node);
break;
case K_ANY:
node.SetTypeIfUnknown(DTDElement::eSequence);
{
string name = CreateTmpEmbeddedName(node.GetName(), emb);
DTDElement& elem = m_MapElement[name];
elem.SetName(name);
elem.SetSourceLine(Lexer().CurrentLine());
elem.SetEmbedded();
elem.SetType(DTDElement::eAny);
elem.SetQualified(node.IsQualified());
ParseAny(elem);
AddElementContent(node,name);
}
break;
case K_SEQUENCE:
emb= node.GetContent().size();
if (emb != 0 && extended) {
node.SetTypeIfUnknown(DTDElement::eSequence);
if (node.GetType() != DTDElement::eSequence) {
ParseError("sequence");
}
tok = GetRawAttributeSet();
eatEOT = true;
break;
}
curr_type = node.GetType();
if (curr_type == DTDElement::eUnknown ||
curr_type == DTDElement::eUnknownGroup ||
(m_ResolveTypes && curr_type == DTDElement::eEmpty)) {
node.SetType(DTDElement::eSequence);
ParseContainer(node);
if (node.GetContent().empty()) {
node.ResetType(curr_type);
}
} else {
string name = CreateTmpEmbeddedName(node.GetName(), emb);
DTDElement& elem = m_MapElement[name];
elem.SetName(name);
elem.SetSourceLine(Lexer().CurrentLine());
elem.SetEmbedded();
elem.SetType(DTDElement::eSequence);
elem.SetQualified(node.IsQualified());
ParseContainer(elem);
AddElementContent(node,name);
}
break;
case K_CHOICE:
curr_type = node.GetType();
if (curr_type == DTDElement::eUnknown ||
curr_type == DTDElement::eUnknownGroup ||
(m_ResolveTypes && curr_type == DTDElement::eEmpty)) {
node.SetType(DTDElement::eChoice);
ParseContainer(node);
if (node.GetContent().empty()) {
node.ResetType(curr_type);
}
} else {
string name = CreateTmpEmbeddedName(node.GetName(), emb);
DTDElement& elem = m_MapElement[name];
elem.SetName(name);
elem.SetSourceLine(Lexer().CurrentLine());
elem.SetEmbedded();
elem.SetType(DTDElement::eChoice);
elem.SetQualified(node.IsQualified());
ParseContainer(elem);
AddElementContent(node,name);
}
break;
case K_SET:
curr_type = node.GetType();
if (curr_type == DTDElement::eUnknown ||
curr_type == DTDElement::eUnknownGroup ||
(m_ResolveTypes && curr_type == DTDElement::eEmpty)) {
node.SetType(DTDElement::eSet);
ParseContainer(node);
if (node.GetContent().empty()) {
node.ResetType(curr_type);
}
} else {
string name = CreateTmpEmbeddedName(node.GetName(), emb);
DTDElement& elem = m_MapElement[name];
elem.SetName(name);
elem.SetSourceLine(Lexer().CurrentLine());
elem.SetEmbedded();
elem.SetType(DTDElement::eSet);
elem.SetQualified(node.IsQualified());
ParseContainer(elem);
AddElementContent(node,name);
}
break;
case K_ELEMENT:
{
string name = ParseElementContent(&node,emb);
AddElementContent(node,name);
}
break;
case K_GROUP:
{
string name = ParseGroup(&node,emb);
AddElementContent(node,name);
}
break;
case K_ANNOTATION:
SetCommentsIfEmpty(&(node.Comments()));
ParseAnnotation();
break;
case K_UNION:
ParseUnion(node);
break;
case K_LIST:
ParseList(node);
break;
default:
for ( tok = GetNextToken(); tok == K_ATTPAIR || tok == K_XMLNS; tok = GetNextToken())
;
if (tok == K_CLOSING) {
ParseContent(node);
}
break;
}
}
FixEmbeddedNames(node);
return hasContents;
}
/**
* One of the key methods inside of parser can take a general description
* of what should be on the token stream and attempts to figure it out.
*
* @param in -- The input token stream.
* @param rules -- The set of input rules we tokenize by.
* @param elements -- The set of elements.
*
* @return true if the set of rules can be parsed, false otherwise.
*/
bool Parser::CanParse( TokenStream& in
, std::vector<ParseRule> const& rules
, std::vector<Element>& elements) const
{
bool success = true;
elements.clear();
size_t rule_index = 0;
while(in.HasTokens() && rule_index < rules.size())
{
// Grab the current rule.
ParseRule const& rule = rules[rule_index];
// Check to see if the current rule states that we expect a token,
// if this is the case, then try to match the next element as a
// token.
if (rule.Type() == strine::Types::TOKEN)
{
strine::Element current_element = in.NextToken();
std::string const rule_token(rule.Token());
std::string const element_string(current_element.ToString());
success = (rule_token == element_string);
}
// If the rule states that the type should be ANY, try to parse
// out any element by using ParseAny.
else if (rule.Type() == ParseRule::ANY)
{
strine::Element current_element;
// Push a marker onto the token stream.
in.Push();
success = ParseAny( in, current_element );
if (success)
{
elements.push_back(current_element);
}
else
{
in.Rollback();
}
}
// If the rule.Type() hasn't been specified, then go ahead and try
// to match up the types.
else
{
strine::Element current_element = in.NextToken();
if (rule.Type() == current_element.Type())
{
success = true;
elements.push_back(current_element);
}
else
{
success = false;
}
}
if (false == success)
{
break;
}
rule_index += 1;
}
if (rule_index < rules.size())
{
success = false;
}
return success;
}