std::shared_ptr<Object> read_character(std::istream &in) {
std::stringstream buf;
int c = in.get();
if (in.eof()) {
throw "EOF while reading character";
}
buf.put(c);
for (; ;) {
c = in.get();
if (in.eof() || iswhitespace(c) || isterminator(c)) {
in.unget();
break;
}
buf.put(c);
}
std::string token = buf.str();
if (token.size() == 1) {
return std::make_shared<Character>( token[0] );
} else if (token == "newline") {
return std::make_shared<Character>( '\n' );
} else if (token == "space") {
return std::make_shared<Character>( ' ' );
} else if (token == "tab") {
return std::make_shared<Character>( '\t' );
} else if (token == "backspace") {
return std::make_shared<Character>( '\b' );
} else if (token == "formfeed") {
return std::make_shared<Character>( '\f' );
} else if (token == "return") {
return std::make_shared<Character>( '\r' );
} else if (token[0] == 'u') {
long uc = read_unicode_char(token, 1, 4, 16);
if (c >= 0xD800 && c <= 0xDFFF) {
// TODO: java clojure actually prints u + the hex value of uc
// is this any different than token?
throw "Invalid character constant: \\" + token;
}
return std::make_shared<Character>( uc );
} else if (token[0] == 'o') {
int len = token.size() - 1;
if (len > 3) {
throw "Invalid octal escape sequence length: " + std::to_string(len);
}
long uc = read_unicode_char(token, 1, len, 8);
if (uc > 0377) {
throw "Octal escape sequence mst be in range [0, 377].";
}
return std::make_shared<Character>( uc );
}
throw "Unsupported character: \\" + token;
}
std::shared_ptr<Object> read_token(std::istream &in) {
std::stringstream buf;
std::string ns = "";
for (; ;) {
int c = in.get();
if (in.eof() || iswhitespace(c) || isterminator(c)) {
in.unget();
break;
}
buf.put(c);
if (c == '/') {
ns = buf.str();
}
}
std::string s = buf.str();
if (s.back() == '/') {
// TODO: this is the only case i can spot where we have an invalid token
throw "Invalid token: " + s;
}
if (s == "nil") {
return Object::nil;
}
if (s == "true") {
return Object::T;
}
if (s == "false") {
return Object::F;
}
// TODO: / = slash, clojure.core// = slash
if ((ns != "" && ns.substr(ns.size()-3) == ":/")
|| s.back() == ':'
|| s.find("::", 1) != std::string::npos) {
return nullptr;
}
if (s[0] == ':' && s[1] == ':') {
auto ks = Symbol::create(s.substr(2));
// TODO: handle namespace qualified Keywords
return nullptr;
}
bool iskey = s[0] == ':';
if (iskey) {
return Keyword::create(s.substr(1));
}
return std::make_shared<Symbol>(s);
}
std::shared_ptr<Object> read_number(std::istream &in) {
std::stringstream buf;
std::string start = "";
bool error = false;
number_type type = number_type::none;
int base = 10;
// lambda to tell if we've reached the end of the number
std::function<bool (int)> isend = [&in] (int i) -> bool {
return iswhitespace(i) || isterminator(i) || in.eof();
};
int c;
// process the first few chars
while (!isend(c = in.get())) {
if (c == '0') { // if the first character is 0, there's only a few posible options
c = in.get();
if (isend(c)) { // check if it's 0
type = number_type::integer;
buf.put('0');
break;
} else if (c == 'x' || c == 'X') { // check if it's a hex value
type = number_type::integer;
base = 16;
start += "0";
start += (char)c;
break;
} else if (c >= '0' && c <= '7') { // check if it's an oct value
type = number_type::integer;
buf.put(c);
base = 8;
start = "0";
break;
} else if (c == 'e' || c == 'E') {
type = number_type::scientific;
buf.put('0');
buf.put(c);
break;
} else {
error = true;
buf << "O" << (char)c;
break;
}
} else if (c == '-') { // add the sign to the buf if a -
buf.put(c);
} else if (c == '+') {
// ignore the sign if it's a +
} else {
buf.put(c);
break;
}
}
while (1) {
c = in.get();
if (isend(c)) {
// TODO: process buf and get the value
in.unget(); // TODO: make sure we want to unget here
if (error) {
buf.str("Invalid Number: " + start + buf.str());
throw buf.str();
} else {
if (type == number_type::integer || type == number_type::none) {
return std::make_shared<Integer>(buf.str(), base);
} else if (type == number_type::irrational || type == number_type::scientific) {
return std::make_shared<Irrational>(buf.str());
} else if (type == number_type::ratio) {
std::string r = buf.str();
size_t s = r.find('/');
return std::make_shared<Ratio>(r.substr(0, s), r.substr(s+1));
}
}
return Object::nil; // should never get here
}
if (error) {
buf.put(c);
} else if (c == 'r') {
if (type != number_type::none) { // we already have a base, so this is an invalid number
error = true;
buf.put(c);
} else {
std::string radix = buf.str();
buf.str("");
if (radix.size() > 2) {
error = true;
buf.put(c);
} else {
base = stoi(radix);
if (base > 36) {
// TODO: NumberFormatException
throw "Radix out of range";
}
type = number_type::integer;
}
}
} else if (c == '/') {
buf.put(c);
if (type != number_type::none) { // we already have a base, so this is an invalid number
error = true;
} else {
type = number_type::ratio;
c = in.get();
if (isend(c)) {
error = true;
in.unget();
} else {
buf.put(c);
}
}
} else if (c == '.' && type == number_type::none) {
type = number_type::irrational;
buf.put(c);
} else if ((c == 'e' || c == 'E') &&
(type == number_type::none || type == number_type::irrational)) {
type = number_type::scientific;
buf.put(c);
c = in.get();
if (isend(c)) {
error = true;
} else {
buf.put(c);
if (!(c == '-' || c == '+' || (c >= '0' && c <= '9'))) {
error = true;
}
}
} else if ((c == 'N' || c == 'M') && base == 10) { // M and N endings only possible with base 10
c = in.get();
if (!isend(c)) {
error = true;
buf.put(c);
} else {
in.unget();
}
} else if (c >= '0' && ((base < 11 && (c < '0' + base)) ||
(base > 10 && (c <= '9' ||
(c >= 'a' && c < 'a' + (base-10)) ||
(c >= 'A' && c < 'A' + (base-10))
)))) {
buf.put(c);
} else {
buf.put(c);
error = true;
}
}
return Object::nil;
}
/*
* cha_jfgets - fgets() for Japanese Text.
*
*/
char *
cha_jfgets(char *buffer, int bufsize, FILE * stream)
{
static unsigned char ibuf[INNER_BUFSIZE];
/* set to the end of line */
static unsigned char *pos = (unsigned char *) "";
unsigned char *q;
int count;
int kflag; /* kanji flag(0=not found, 1=found) */
if (pos == NULL &&
(pos = cha_fget_line(ibuf, sizeof(ibuf), stream)) == NULL)
return NULL;
kflag = 0;
q = (unsigned char *) buffer;
bufsize--;
for (count = bufsize; count > 0; count--) {
/*
* line is end without '\n', long string read more
*/
if (*pos == '\0')
if ((pos = cha_fget_line(ibuf, sizeof(ibuf), stream)) == NULL)
break;
/*
* KANJI
*/
if (*pos >= 0x80 && *(pos + 1)) {
if (count < 2)
break;
kflag = 1;
count--;
*q++ = *pos++;
*q++ = *pos++;
/*
* hit delimiter
*/
if (isterminator(pos - 2, jfgets_delimiter)) {
if (*pos == '\n')
pos++;
break;
}
}
/*
* not KANJI
*/
else {
/*
* line is end
*/
if (*pos == '\n') {
/*
* eliminate space characters at the end of line
*/
while (q > (unsigned char *) buffer
&& (q[-1] == ' ' || q[-1] == '\t'))
q--;
if ((pos =
cha_fget_line(ibuf, sizeof(ibuf), stream)) == NULL)
break;
while (*pos == ' ' || *pos == '\t')
pos++;
/*
* not have kanji or no space, return with this line
*/
if (count <= 0)
break;
/*
* have kanji, connect next line
*/
/*
* double '\n' is paragraph end. so it is delimiter
*/
if (*pos == '\n')
break;
/*
* "ASCII\nASCII" -> "ASCII ASCII"
*/
if (!kflag && !(*pos & 0x80))
*q++ = ' ';
} else {
if (*pos != ' ' && *pos != '\t')
kflag = 0;
*q++ = *pos++;
/*
* hit delimiter
*/
if (isterminator(pos - 1, jfgets_delimiter)) {
if (*pos == '\n')
pos++;
break;
}
}
}
}
*q = '\0';
return buffer;
}
