void SCsTranslator::dumpNode(pANTLR3_BASE_TREE node, std::ofstream &stream)
{
StringStream ss;
ss << node;
String s_root = ss.str().substr(3);
pANTLR3_COMMON_TOKEN tok = node->getToken(node);
if (tok)
{
stream << s_root << " [shape=box];" << std::endl;
String label((const char*) node->getText(node)->chars);
std::replace(label.begin(), label.end(), '"', '\'');
stream << s_root << " [label=\"" << label << "\"];" << std::endl;
}
else
stream << s_root << " [shape=circle];" << std::endl;
uint32 n = node->getChildCount(node);
for (uint32 i = 0; i < n; ++i)
{
pANTLR3_BASE_TREE child = (pANTLR3_BASE_TREE) node->getChild(node, i);
StringStream s1;
s1 << child;
stream << s_root << " -> " << s1.str().substr(3) << ";" << std::endl;
dumpNode(child, stream);
}
}
pANTLR3_STRING emerson_printAST(pANTLR3_BASE_TREE tree, pANTLR3_UINT8* parserTokenNames)
{
pANTLR3_STRING string;
ANTLR3_UINT32 i;
ANTLR3_UINT32 n;
pANTLR3_BASE_TREE t;
if(tree->children == NULL || tree->children->size(tree->children) == 0)
{
return tree->toString(tree);
}
// THis is how you get a new string. The string is blank
string = tree->strFactory->newRaw(tree->strFactory);
if(tree->isNilNode(tree) == ANTLR3_FALSE)
{
string->append8 (string, "(");
pANTLR3_COMMON_TOKEN token = tree->getToken(tree);
ANTLR3_UINT32 type = token->type;
string->append(string, (const char*)parserTokenNames[type]);
string->append8(string, " ");
}
if(tree->children != NULL)
{
n = tree->children->size(tree->children);
for (i = 0; i < n; i++)
{
t = (pANTLR3_BASE_TREE) tree->children->get(tree->children, i);
if (i > 0)
{
string->append8(string, " ");
}
string->appendS(string, emerson_printAST(t,parserTokenNames));
}
}
if(tree->isNilNode(tree) == ANTLR3_FALSE)
{
string->append8(string,")");
}
return string;
}
std::string MySQLRecognitionBase::dumpTree(pANTLR3_UINT8 *tokenNames, pANTLR3_BASE_TREE tree, const std::string &indentation)
{
std::string result;
ANTLR3_UINT32 char_pos = tree->getCharPositionInLine(tree);
ANTLR3_UINT32 line = tree->getLine(tree);
pANTLR3_STRING token_text = tree->getText(tree);
pANTLR3_COMMON_TOKEN token = tree->getToken(tree);
const char* utf8 = (const char*)token_text->chars;
if (token != NULL)
{
ANTLR3_UINT32 token_type = token->getType(token);
pANTLR3_UINT8 token_name;
if (token_type == EOF)
token_name = (pANTLR3_UINT8)"EOF";
else
token_name = tokenNames[token_type];
#ifdef ANTLR3_USE_64BIT
result = base::strfmt("%s(line: %i, offset: %i, length: %" PRId64 ", index: %" PRId64 ", %s[%i]) %s\n",
indentation.c_str(), line, char_pos, token->stop - token->start + 1, token->index, token_name,
token_type, utf8);
#else
result = base::strfmt("%s(line: %i, offset: %i, length: %i, index: %i, %s[%i]) %s\n",
indentation.c_str(), line, char_pos, token->stop - token->start + 1, token->index, token_name,
token_type, utf8);
#endif
}
else
{
result = base::strfmt("%s(line: %i, offset: %i, nil) %s\n", indentation.c_str(), line, char_pos, utf8);
}
for (ANTLR3_UINT32 index = 0; index < tree->getChildCount(tree); index++)
{
pANTLR3_BASE_TREE child = (pANTLR3_BASE_TREE)tree->getChild(tree, index);
std::string child_text = dumpTree(tokenNames, child, indentation + "\t");
result += child_text;
}
return result;
}
SCsTranslator::eSentenceType SCsTranslator::determineSentenceType(pANTLR3_BASE_TREE node)
{
pANTLR3_COMMON_TOKEN tok = node->getToken(node);
assert(tok);
if (tok->type == SEP_SIMPLE)
return SentenceLevel1;
if (tok->type == CONNECTORS)
return SentenceLevel2_7;
if (tok->type == SEP_ASSIGN)
return SentenceAssign;
if (tok->type == EOF)
return SentenceEOF;
return SentenceUnknown;
}
static pANTLR3_COMMON_TOKEN
getToken (pANTLR3_BASE_TREE_ADAPTOR adaptor, pANTLR3_BASE_TREE t)
{
return t->getToken(t);
}
/**
* Returns the text of the given node. The result depends on the type of the node. If it represents
* a quoted text entity then two consecutive quote chars are replaced by a single one and if
* escape sequence parsing is not switched off (as per sql mode) then such sequences are handled too.
*/
std::string MySQLRecognitionBase::token_text(pANTLR3_BASE_TREE node, bool keepQuotes)
{
pANTLR3_STRING text = node->getText(node);
if (text == NULL)
return "";
std::string chars;
pANTLR3_COMMON_TOKEN token = node->getToken(node);
ANTLR3_UINT32 type = (token != NULL) ? token->type : ANTLR3_TOKEN_INVALID;
if (type == STRING_TOKEN)
{
// STRING is the grouping subtree for multiple consecutive string literals, which
// must be concatenated.
for (ANTLR3_UINT32 index = 0; index < node->getChildCount(node); index++)
{
pANTLR3_BASE_TREE child = (pANTLR3_BASE_TREE)node->getChild(node, index);
chars += token_text(child, keepQuotes);
}
return chars;
}
chars = (const char*)text->chars;
std::string quote_char;
switch (type)
{
case BACK_TICK_QUOTED_ID:
quote_char = "`";
break;
case SINGLE_QUOTED_TEXT:
quote_char = "'";
break;
case DOUBLE_QUOTED_TEXT:
quote_char = "\"";
break;
default:
return chars;
}
// First unquote then handle escape squences and double quotes.
if (chars.size() < 3)
{
if (keepQuotes)
return chars;
return ""; // Also handles an invalid single quote char gracefully.
}
// Need to unquote even if keepQuotes is true to avoid trouble with replacing the outer quotes.
// Add them back below.
chars = base::unquote(chars);
if ((d->_sql_mode & SQL_MODE_NO_BACKSLASH_ESCAPES) == 0)
chars = base::unescape_sql_string(chars, quote_char[0]);
else
if (token->user1 > 0)
{
// The field user1 is set by the parser to the number of quote char pairs it found.
// So we can use it to shortcut our handling here.
base::replace(chars, quote_char + quote_char, quote_char);
}
if (keepQuotes)
return quote_char + chars + quote_char;
return chars;
}
pANTLR3_COMMON_TOKEN GetToken(pANTLR3_BASE_TREE node)
{
return node->getToken(node);
}
