| str[s,false]
{ Debug("parser-sygus") << "Sygus grammar " << fun << " : string literal \""
<< s << "\"" << std::endl;
- sgt.d_expr = MK_CONST( ::CVC4::String(s) );
+ sgt.d_expr = MK_CONST( ::CVC4::String(s, true) );
sgt.d_name = s;
sgt.d_gterm_type = SygusGTerm::gterm_op;
}
expr = MK_CONST( BitVector(binString, 2) ); }
| str[s,false]
- { expr = MK_CONST( ::CVC4::String(s) ); }
+ { expr = MK_CONST( ::CVC4::String(s, true) ); }
| FP_RNE_TOK { expr = MK_CONST(roundNearestTiesToEven); }
| FP_RNA_TOK { expr = MK_CONST(roundNearestTiesToAway); }
| FP_RTP_TOK { expr = MK_CONST(roundTowardPositive); }
static bool stringifyRegexp(Node n, stringstream& ss) {
if(n.getKind() == kind::STRING_TO_REGEXP) {
- ss << n[0].getConst<String>().toString();
+ ss << n[0].getConst<String>().toString(true);
} else if(n.getKind() == kind::REGEXP_CONCAT) {
for(unsigned i = 0; i < n.getNumChildren(); ++i) {
if(!stringifyRegexp(n[i], ss)) {
case kind::CONST_STRING: {
//const std::vector<unsigned int>& s = n.getConst<String>().getVec();
- std::string s = n.getConst<String>().toString();
+ std::string s = n.getConst<String>().toString(true);
out << '"';
for(size_t i = 0; i < s.size(); ++i) {
//char c = String::convertUnsignedIntToChar(s[i]);
return true;
}
-std::vector<unsigned> String::toInternal(const std::string &s) {
+std::vector<unsigned> String::toInternal(const std::string &s,
+ bool useEscSequences) {
std::vector<unsigned> str;
unsigned i = 0;
while (i < s.size()) {
- if (s[i] == '\\') {
+ if (s[i] == '\\' && useEscSequences) {
i++;
if (i < s.size()) {
switch (s[i]) {
} break;
default: {
if (isdigit(s[i])) {
+ // octal escape sequences TODO : revisit (issue #1251).
int num = (int)s[i] - (int)'0';
bool flag = num < 4;
if (i + 1 < s.size() && num < 8 && isdigit(s[i + 1]) &&
throw CVC4::Exception("should be handled by lexer: \"" + s + "\"");
// str.push_back( convertCharToUnsignedInt('\\') );
}
- } else if ((unsigned)s[i] > 127) {
+ } else if ((unsigned)s[i] > 127 && useEscSequences) {
throw CVC4::Exception("Illegal String Literal: \"" + s +
"\", must use escaped sequence");
} else {
return i;
}
-std::string String::toString() const {
+std::string String::toString(bool useEscSequences) const {
std::string str;
for (unsigned int i = 0; i < size(); ++i) {
unsigned char c = convertUnsignedIntToChar(d_str[i]);
- if (isprint(c)) {
+ if (!useEscSequences) {
+ str += c;
+ } else if (isprint(c)) {
if (c == '\\') {
str += "\\\\";
}
}
std::ostream &operator<<(std::ostream &os, const String &s) {
- return os << "\"" << s.toString() << "\"";
+ return os << "\"" << s.toString(true) << "\"";
}
std::ostream &operator<<(std::ostream &out, const RegExp &s) {
return (c >= ' ' && c <= '~'); // isprint( (int)c );
}
+ /** constructors for String
+ *
+ * Internally, a CVC4::String is represented by a vector of unsigned
+ * integers (d_str), where the correspondence between C++ characters
+ * to and from unsigned integers is determined by
+ * by convertCharToUnsignedInt and convertUnsignedIntToChar.
+ *
+ * If useEscSequences is true, then the escape sequences in the input
+ * are converted to the corresponding character. This constructor may
+ * throw an exception if the input contains unrecognized escape sequences.
+ * Currently supported escape sequences are \n, \t, \v, \b, \r, \f, \a, \\,
+ * \x[N] where N is a hexidecimal, and octal escape sequences of the
+ * form \[c1]([c2]([c3])?)? where c1, c2, c3 are digits from 0 to 7.
+ *
+ * If useEscSequences is false, then the characters of the constructed
+ * CVC4::String correspond one-to-one with the input string.
+ */
String() = default;
- explicit String(const std::string& s) : d_str(toInternal(s)) {}
- explicit String(const char* s) : d_str(toInternal(std::string(s))) {}
+ explicit String(const std::string& s, bool useEscSequences = false)
+ : d_str(toInternal(s, useEscSequences)) {}
+ explicit String(const char* s, bool useEscSequences = false)
+ : d_str(toInternal(std::string(s), useEscSequences)) {}
explicit String(const unsigned char c)
: d_str({convertCharToUnsignedInt(c)}) {}
explicit String(const std::vector<unsigned>& s) : d_str(s) {}
bool strncmp(const String& y, const std::size_t np) const;
bool rstrncmp(const String& y, const std::size_t np) const;
- /*
- * Convenience functions
- */
- std::string toString() const;
+ /* toString
+ * Converts this string to a std::string.
+ *
+ * If useEscSequences is true, then unprintable characters
+ * are converted to escape sequences. The escape sequences
+ * \n, \t, \v, \b, \r, \f, \a, \\ are printed in this way.
+ * For all other unprintable characters, we print \x[N] where
+ * [N] is the 2 digit hexidecimal corresponding to value of
+ * the character.
+ *
+ * If useEscSequences is false, the returned std::string's characters
+ * map one-to-one with the characters in this string.
+ * Notice that for all std::string s, we have that
+ * CVC4::String( s ).toString() = s.
+ */
+ std::string toString(bool useEscSequences = false) const;
+ /** is this the empty string? */
bool empty() const { return d_str.empty(); }
+ /** is this the empty string? */
bool isEmptyString() const { return empty(); }
+ /** Return the length of the string */
std::size_t size() const { return d_str.size(); }
unsigned char getFirstChar() const { return getUnsignedCharAt(0); }
// guarded
static unsigned char hexToDec(unsigned char c);
- static std::vector<unsigned> toInternal(const std::string& s);
+ static std::vector<unsigned> toInternal(const std::string& s,
+ bool useEscSequences = true);
unsigned char getUnsignedCharAt(size_t pos) const;
/**