/**
// Recursively calculate nullable for this node.
*/
void RegexNode::calculate_nullable()
{
for ( std::vector<ptr<RegexNode> >::const_iterator i = nodes_.begin(); i != nodes_.end(); ++i )
{
RegexNode* node = i->get();
SWEET_ASSERT( node );
node->calculate_nullable();
}
switch ( type_ )
{
case LEXER_NODE_NULL:
nullable_ = true;
break;
case LEXER_NODE_CAT:
SWEET_ASSERT( nodes_.size() == 2 );
nullable_ = nodes_[0]->is_nullable() && nodes_[1]->is_nullable();
break;
case LEXER_NODE_OR:
SWEET_ASSERT( nodes_.size() == 2 );
nullable_ = nodes_[0]->is_nullable() || nodes_[1]->is_nullable();
break;
case LEXER_NODE_STAR:
case LEXER_NODE_OPTIONAL:
nullable_ = true;
break;
case LEXER_NODE_PLUS:
SWEET_ASSERT( nodes_.size() == 1 );
nullable_ = nodes_[0]->is_nullable();
break;
case LEXER_NODE_SYMBOL:
nullable_ = begin_character_ == INVALID_BEGIN_CHARACTER && end_character_ == INVALID_END_CHARACTER;
break;
case LEXER_NODE_ACTION:
nullable_ = false;
break;
default:
SWEET_ASSERT( false );
nullable_ = false;
break;
}
}
/**
// Recursively calculate the last positions at this node.
*/
void RegexNode::calculate_last_positions()
{
for ( std::vector<ptr<RegexNode> >::const_iterator i = nodes_.begin(); i != nodes_.end(); ++i )
{
RegexNode* node = i->get();
SWEET_ASSERT( node );
node->calculate_last_positions();
}
switch ( type_ )
{
case LEXER_NODE_NULL:
break;
case LEXER_NODE_CAT:
{
SWEET_ASSERT( nodes_.size() == 2 );
last_positions_.insert( nodes_[1]->last_positions_.begin(), nodes_[1]->last_positions_.end() );
if ( nodes_[1]->is_nullable() )
{
last_positions_.insert( nodes_[0]->last_positions_.begin(), nodes_[0]->last_positions_.end() );
}
break;
}
case LEXER_NODE_OR:
SWEET_ASSERT( nodes_.size() == 2 );
last_positions_.insert( nodes_[0]->last_positions_.begin(), nodes_[0]->last_positions_.end() );
last_positions_.insert( nodes_[1]->last_positions_.begin(), nodes_[1]->last_positions_.end() );
break;
case LEXER_NODE_STAR:
case LEXER_NODE_PLUS:
case LEXER_NODE_OPTIONAL:
SWEET_ASSERT( nodes_.size() == 1 );
last_positions_.insert( nodes_[0]->last_positions_.begin(), nodes_[0]->last_positions_.end() );
break;
case LEXER_NODE_SYMBOL:
case LEXER_NODE_ACTION:
last_positions_.insert( this );
break;
default:
SWEET_ASSERT( false );
break;
}
}
/**
// Print the RegexNodes \e nodes.
//
// @param nodes
// The RegexNodes to print.
//
// @param level
// The recursion level to use when identing lines.
*/
void RegexSyntaxTree::print_nodes( const vector<std::shared_ptr<RegexNode> >& nodes, int level ) const
{
for ( vector<std::shared_ptr<RegexNode> >::const_iterator i = nodes.begin(); i != nodes.end(); ++i )
{
static const char* LEXER_NODE_TYPES [LEXER_NODE_COUNT] =
{
"LEXER_NODE_NULL",
"LEXER_NODE_CAT",
"LEXER_NODE_OR",
"LEXER_NODE_STAR",
"LEXER_NODE_PLUS",
"LEXER_NODE_OPTIONAL",
"LEXER_NODE_SYMBOL",
"LEXER_NODE_ACTION"
};
RegexNode* node = i->get();
LALR_ASSERT( node != NULL );
for ( int i = 0; i < level; ++i )
{
printf( " " );
}
printf( "%d, %s, %s, [%d, %d), nullable=%s", node->get_index(), LEXER_NODE_TYPES[node->get_type()], node->get_lexeme(), node->get_begin_character(), node->get_end_character(), node->is_nullable() ? "true" : "false" );
printf( ", first={" );
print_positions( node->get_first_positions() );
printf( "}, last={" );
print_positions( node->get_last_positions() );
printf( "}, follow={" );
print_positions( node->get_follow_positions() );
printf( "}\n" );
if ( !node->get_nodes().empty() )
{
print_nodes( node->get_nodes(), level + 1 );
}
}
}
/**
// Recursively calculate the follow positions at this node.
*/
void RegexNode::calculate_follow_positions()
{
for ( std::vector<ptr<RegexNode> >::const_iterator i = nodes_.begin(); i != nodes_.end(); ++i )
{
RegexNode* node = i->get();
SWEET_ASSERT( node );
node->calculate_follow_positions();
}
switch ( type_ )
{
case LEXER_NODE_NULL:
break;
case LEXER_NODE_CAT:
{
SWEET_ASSERT( nodes_.size() == 2 );
const std::set<RegexNode*, RegexNodeLess>& last_positions = nodes_[0]->last_positions_;
const std::set<RegexNode*, RegexNodeLess>& first_positions = nodes_[1]->first_positions_;
for ( std::set<RegexNode*, RegexNodeLess>::const_iterator i = last_positions.begin(); i != last_positions.end(); ++i )
{
RegexNode* node = *i;
SWEET_ASSERT( node );
if ( node->get_type() == LEXER_NODE_SYMBOL || node->get_type() == LEXER_NODE_ACTION )
{
node->follow_positions_.insert( first_positions.begin(), first_positions.end() );
}
}
break;
}
case LEXER_NODE_STAR:
case LEXER_NODE_PLUS:
{
SWEET_ASSERT( nodes_.size() == 1 );
const std::set<RegexNode*, RegexNodeLess>& last_positions = last_positions_;
const std::set<RegexNode*, RegexNodeLess>& first_positions = first_positions_;
for ( std::set<RegexNode*, RegexNodeLess>::const_iterator i = last_positions.begin(); i != last_positions.end(); ++i )
{
RegexNode* node = *i;
SWEET_ASSERT( node );
if ( node->get_type() == LEXER_NODE_SYMBOL || node->get_type() == LEXER_NODE_ACTION )
{
node->follow_positions_.insert( first_positions.begin(), first_positions.end() );
}
}
break;
}
case LEXER_NODE_OR:
case LEXER_NODE_OPTIONAL:
case LEXER_NODE_SYMBOL:
case LEXER_NODE_ACTION:
break;
default:
SWEET_ASSERT( false );
break;
}
}