/*JSON{
"type" : "staticmethod",
"class" : "url",
"name" : "parse",
"generate" : "jswrap_url_parse",
"params" : [
["urlStr","JsVar","A URL to be parsed"],
["parseQuery","bool","Whether to parse the query string into an object not (default = false)"]
],
"return" : ["JsVar","An object containing options for ```http.request``` or ```http.get```. Contains `method`, `host`, `path`, `pathname`, `search`, `port` and `query`"]
}
A utility function to split a URL into parts
This is useful in web servers for instance when handling a request.
For instance `url.parse("/a?b=c&d=e",true)` returns `{"method":"GET","host":"","path":"/a?b=c&d=e","pathname":"/a","search":"?b=c&d=e","port":80,"query":{"b":"c","d":"e"}}`
*/
JsVar *jswrap_url_parse(JsVar *url, bool parseQuery) {
if (!jsvIsString(url)) return 0;
JsVar *obj = jsvNewWithFlags(JSV_OBJECT);
if (!obj) return 0; // out of memory
// scan string to try and pick stuff out
JsvStringIterator it;
jsvStringIteratorNew(&it, url, 0);
int slashes = 0;
int colons = 0;
int addrStart = -1;
int portStart = -1;
int pathStart = -1;
int searchStart = -1;
int charIdx = 0;
int portNumber = 0;
while (jsvStringIteratorHasChar(&it)) {
char ch = jsvStringIteratorGetChar(&it);
if (ch == '/') {
slashes++;
if (pathStart<0) pathStart = charIdx;
if (colons==1 && slashes==2 && addrStart<0) {
addrStart = charIdx;
pathStart = -1;
searchStart = -1;
}
}
if (ch == ':') {
colons++;
if (addrStart>=0 && pathStart<0)
portStart = charIdx;
}
if (portStart>=0 && charIdx>portStart && pathStart<0 && ch >= '0' && ch <= '9') {
portNumber = portNumber*10 + (ch-'0');
}
if (ch == '?' && pathStart>=0) {
searchStart = charIdx;
}
jsvStringIteratorNext(&it);
charIdx++;
}
jsvStringIteratorFree(&it);
// try and sort stuff out
if (pathStart<0) pathStart = charIdx;
if (pathStart<0) pathStart = charIdx;
int addrEnd = (portStart>=0) ? portStart : pathStart;
// pull out details
if (addrStart>0)
jsvObjectSetChildAndUnLock(obj, "protocol", jsvNewFromStringVar(url, 0, (size_t)addrStart-1));
jsvObjectSetChildAndUnLock(obj, "method", jsvNewFromString("GET"));
jsvObjectSetChildAndUnLock(obj, "host", jsvNewFromStringVar(url, (size_t)(addrStart+1), (size_t)(addrEnd-(addrStart+1))));
JsVar *v;
v = jsvNewFromStringVar(url, (size_t)pathStart, JSVAPPENDSTRINGVAR_MAXLENGTH);
if (jsvGetStringLength(v)==0) jsvAppendString(v, "/");
jsvObjectSetChildAndUnLock(obj, "path", v);
v = jsvNewFromStringVar(url, (size_t)pathStart, (size_t)((searchStart>=0)?(searchStart-pathStart):JSVAPPENDSTRINGVAR_MAXLENGTH));
if (jsvGetStringLength(v)==0) jsvAppendString(v, "/");
jsvObjectSetChildAndUnLock(obj, "pathname", v);
jsvObjectSetChildAndUnLock(obj, "search", (searchStart>=0)?jsvNewFromStringVar(url, (size_t)searchStart, JSVAPPENDSTRINGVAR_MAXLENGTH):jsvNewNull());
jsvObjectSetChildAndUnLock(obj, "port", (portNumber<=0 || portNumber>65535) ? jsvNewWithFlags(JSV_NULL) : jsvNewFromInteger(portNumber));
JsVar *query = (searchStart>=0)?jsvNewFromStringVar(url, (size_t)(searchStart+1), JSVAPPENDSTRINGVAR_MAXLENGTH):jsvNewNull();
if (parseQuery && !jsvIsNull(query)) {
JsVar *queryStr = query;
jsvStringIteratorNew(&it, query, 0);
query = jsvNewWithFlags(JSV_OBJECT);
JsVar *key = jsvNewFromEmptyString();
JsVar *val = jsvNewFromEmptyString();
bool hadEquals = false;
while (jsvStringIteratorHasChar(&it)) {
char ch = jsvStringIteratorGetChar(&it);
if (ch=='&') {
if (jsvGetStringLength(key)>0 || jsvGetStringLength(val)>0) {
key = jsvAsArrayIndexAndUnLock(key); // make sure "0" gets made into 0
jsvMakeIntoVariableName(key, val);
jsvAddName(query, key);
jsvUnLock2(key, val);
key = jsvNewFromEmptyString();
val = jsvNewFromEmptyString();
hadEquals = false;
}
} else if (!hadEquals && ch=='=') {
hadEquals = true;
} else {
// decode percent escape chars
if (ch=='%') {
jsvStringIteratorNext(&it);
ch = jsvStringIteratorGetChar(&it);
jsvStringIteratorNext(&it);
ch = (char)((chtod(ch)<<4) | chtod(jsvStringIteratorGetChar(&it)));
}
if (hadEquals) jsvAppendCharacter(val, ch);
else jsvAppendCharacter(key, ch);
}
jsvStringIteratorNext(&it);
charIdx++;
}
jsvStringIteratorFree(&it);
jsvUnLock(queryStr);
if (jsvGetStringLength(key)>0 || jsvGetStringLength(val)>0) {
key = jsvAsArrayIndexAndUnLock(key); // make sure "0" gets made into 0
jsvMakeIntoVariableName(key, val);
jsvAddName(query, key);
}
jsvUnLock2(key, val);
}
jsvObjectSetChildAndUnLock(obj, "query", query);
return obj;
}
void jslGetNextToken(JsLex *lex) {
jslGetNextToken_start:
// Skip whitespace
while (isWhitespace(lex->currCh)) jslGetNextCh(lex);
// Search for comments
if (lex->currCh=='/') {
// newline comments
if (jslNextCh(lex)=='/') {
while (lex->currCh && lex->currCh!='\n') jslGetNextCh(lex);
jslGetNextCh(lex);
goto jslGetNextToken_start;
}
// block comments
if (jslNextCh(lex)=='*') {
while (lex->currCh && !(lex->currCh=='*' && jslNextCh(lex)=='/'))
jslGetNextCh(lex);
if (!lex->currCh) {
lex->tk = LEX_UNFINISHED_COMMENT;
return; /* an unfinished multi-line comment. When in interactive console,
detect this and make sure we accept new lines */
}
jslGetNextCh(lex);
jslGetNextCh(lex);
goto jslGetNextToken_start;
}
}
lex->tk = LEX_EOF;
lex->tokenl = 0; // clear token string
if (lex->tokenValue) {
jsvUnLock(lex->tokenValue);
lex->tokenValue = 0;
}
// record beginning of this token
lex->tokenLastStart = jsvStringIteratorGetIndex(&lex->tokenStart.it) - 1;
/* we don't lock here, because we know that the string itself will be locked
* because of lex->sourceVar */
lex->tokenStart.it = lex->it;
lex->tokenStart.currCh = lex->currCh;
// tokens
if (((unsigned char)lex->currCh) < jslJumpTableStart ||
((unsigned char)lex->currCh) > jslJumpTableEnd) {
// if unhandled by the jump table, just pass it through as a single character
jslSingleChar(lex);
} else {
switch(jslJumpTable[((unsigned char)lex->currCh) - jslJumpTableStart]) {
case JSLJT_ID: {
while (isAlpha(lex->currCh) || isNumeric(lex->currCh) || lex->currCh=='$') {
jslTokenAppendChar(lex, lex->currCh);
jslGetNextCh(lex);
}
lex->tk = LEX_ID;
// We do fancy stuff here to reduce number of compares (hopefully GCC creates a jump table)
switch (lex->token[0]) {
case 'b': if (jslIsToken(lex,"break", 1)) lex->tk = LEX_R_BREAK;
break;
case 'c': if (jslIsToken(lex,"case", 1)) lex->tk = LEX_R_CASE;
else if (jslIsToken(lex,"continue", 1)) lex->tk = LEX_R_CONTINUE;
break;
case 'd': if (jslIsToken(lex,"default", 1)) lex->tk = LEX_R_DEFAULT;
else if (jslIsToken(lex,"do", 1)) lex->tk = LEX_R_DO;
break;
case 'e': if (jslIsToken(lex,"else", 1)) lex->tk = LEX_R_ELSE;
break;
case 'f': if (jslIsToken(lex,"false", 1)) lex->tk = LEX_R_FALSE;
else if (jslIsToken(lex,"for", 1)) lex->tk = LEX_R_FOR;
else if (jslIsToken(lex,"function", 1)) lex->tk = LEX_R_FUNCTION;
break;
case 'i': if (jslIsToken(lex,"if", 1)) lex->tk = LEX_R_IF;
else if (jslIsToken(lex,"in", 1)) lex->tk = LEX_R_IN;
else if (jslIsToken(lex,"instanceof", 1)) lex->tk = LEX_R_INSTANCEOF;
break;
case 'n': if (jslIsToken(lex,"new", 1)) lex->tk = LEX_R_NEW;
else if (jslIsToken(lex,"null", 1)) lex->tk = LEX_R_NULL;
break;
case 'r': if (jslIsToken(lex,"return", 1)) lex->tk = LEX_R_RETURN;
break;
case 's': if (jslIsToken(lex,"switch", 1)) lex->tk = LEX_R_SWITCH;
break;
case 't': if (jslIsToken(lex,"this", 1)) lex->tk = LEX_R_THIS;
else if (jslIsToken(lex,"true", 1)) lex->tk = LEX_R_TRUE;
else if (jslIsToken(lex,"typeof", 1)) lex->tk = LEX_R_TYPEOF;
break;
case 'u': if (jslIsToken(lex,"undefined", 1)) lex->tk = LEX_R_UNDEFINED;
break;
case 'w': if (jslIsToken(lex,"while", 1)) lex->tk = LEX_R_WHILE;
break;
case 'v': if (jslIsToken(lex,"var", 1)) lex->tk = LEX_R_VAR;
else if (jslIsToken(lex,"void", 1)) lex->tk = LEX_R_VOID;
break;
default: break;
} break;
case JSLJT_NUMBER: {
// TODO: check numbers aren't the wrong format
bool canBeFloating = true;
if (lex->currCh=='0') {
jslTokenAppendChar(lex, lex->currCh);
jslGetNextCh(lex);
}
if ((lex->currCh=='x' || lex->currCh=='X') ||
(lex->currCh=='b' || lex->currCh=='B') ||
(lex->currCh=='o' || lex->currCh=='O')) {
canBeFloating = false;
jslTokenAppendChar(lex, lex->currCh); jslGetNextCh(lex);
}
lex->tk = LEX_INT;
while (isNumeric(lex->currCh) || (!canBeFloating && isHexadecimal(lex->currCh))) {
jslTokenAppendChar(lex, lex->currCh);
jslGetNextCh(lex);
}
if (canBeFloating && lex->currCh=='.') {
lex->tk = LEX_FLOAT;
jslTokenAppendChar(lex, '.');
jslGetNextCh(lex);
while (isNumeric(lex->currCh)) {
jslTokenAppendChar(lex, lex->currCh);
jslGetNextCh(lex);
}
}
// do fancy e-style floating point
if (canBeFloating && (lex->currCh=='e'||lex->currCh=='E')) {
lex->tk = LEX_FLOAT;
jslTokenAppendChar(lex, lex->currCh); jslGetNextCh(lex);
if (lex->currCh=='-' || lex->currCh=='+') { jslTokenAppendChar(lex, lex->currCh); jslGetNextCh(lex); }
while (isNumeric(lex->currCh)) {
jslTokenAppendChar(lex, lex->currCh); jslGetNextCh(lex);
}
}
} break;
case JSLJT_STRING:
{
char delim = lex->currCh;
lex->tokenValue = jsvNewFromEmptyString();
// strings...
jslGetNextCh(lex);
while (lex->currCh && lex->currCh!=delim) {
if (lex->currCh == '\\') {
jslGetNextCh(lex);
char ch = lex->currCh;
switch (lex->currCh) {
case 'n' : ch = '\n'; jslGetNextCh(lex); break;
case 'a' : ch = '\a'; jslGetNextCh(lex); break;
case 'r' : ch = '\r'; jslGetNextCh(lex); break;
case 't' : ch = '\t'; jslGetNextCh(lex); break;
case 'x' : { // hex digits
char buf[5] = "0x??";
jslGetNextCh(lex);
buf[2] = lex->currCh; jslGetNextCh(lex);
buf[3] = lex->currCh; jslGetNextCh(lex);
ch = (char)stringToInt(buf);
} break;
default:
if (lex->currCh>='0' && lex->currCh<='7') {
// octal digits
char buf[5] = "0";
buf[1] = lex->currCh;
int n=2;
jslGetNextCh(lex);
if (lex->currCh>='0' && lex->currCh<='7') {
buf[n++] = lex->currCh; jslGetNextCh(lex);
if (lex->currCh>='0' && lex->currCh<='7') {
buf[n++] = lex->currCh; jslGetNextCh(lex);
}
}
buf[n]=0;
ch = (char)stringToInt(buf);
} else {
// for anything else, just push the character through
jslGetNextCh(lex);
}
break;
}
if (lex->tokenValue) {
jslTokenAppendChar(lex, ch);
jsvAppendCharacter(lex->tokenValue, ch);
}
} else {
if (lex->tokenValue) {
jslTokenAppendChar(lex, lex->currCh);
jsvAppendCharacter(lex->tokenValue, lex->currCh);
}
jslGetNextCh(lex);
}
}
jslGetNextCh(lex);
lex->tk = LEX_STR;
} break;
case JSLJT_EXCLAMATION: jslSingleChar(lex);
if (lex->currCh=='=') { // !=
lex->tk = LEX_NEQUAL;
jslGetNextCh(lex);
if (lex->currCh=='=') { // !==
lex->tk = LEX_NTYPEEQUAL;
jslGetNextCh(lex);
}
} break;
case JSLJT_PLUS: jslSingleChar(lex);
if (lex->currCh=='=') {
lex->tk = LEX_PLUSEQUAL;
jslGetNextCh(lex);
} else if (lex->currCh=='+') {
lex->tk = LEX_PLUSPLUS;
jslGetNextCh(lex);
} break;
case JSLJT_MINUS: jslSingleChar(lex);
if (lex->currCh=='=') {
lex->tk = LEX_MINUSEQUAL;
jslGetNextCh(lex);
} else if (lex->currCh=='-') {
lex->tk = LEX_MINUSMINUS;
jslGetNextCh(lex);
} break;
case JSLJT_AND: jslSingleChar(lex);
if (lex->currCh=='=') {
lex->tk = LEX_ANDEQUAL;
jslGetNextCh(lex);
} else if (lex->currCh=='&') {
lex->tk = LEX_ANDAND;
jslGetNextCh(lex);
} break;
case JSLJT_OR: jslSingleChar(lex);
if (lex->currCh=='=') {
lex->tk = LEX_OREQUAL;
jslGetNextCh(lex);
} else if (lex->currCh=='|') {
lex->tk = LEX_OROR;
jslGetNextCh(lex);
} break;
case JSLJT_TOPHAT: jslSingleChar(lex);
if (lex->currCh=='=') {
lex->tk = LEX_XOREQUAL;
jslGetNextCh(lex);
} break;
case JSLJT_STAR: jslSingleChar(lex);
if (lex->currCh=='=') {
lex->tk = LEX_MULEQUAL;
jslGetNextCh(lex);
} break;
case JSLJT_FORWARDSLASH: jslSingleChar(lex);
if (lex->currCh=='=') {
lex->tk = LEX_DIVEQUAL;
jslGetNextCh(lex);
} break;
case JSLJT_PERCENT: jslSingleChar(lex);
if (lex->currCh=='=') {
lex->tk = LEX_MODEQUAL;
jslGetNextCh(lex);
} break;
case JSLJT_EQUAL: jslSingleChar(lex);
if (lex->currCh=='=') { // ==
lex->tk = LEX_EQUAL;
jslGetNextCh(lex);
if (lex->currCh=='=') { // ===
lex->tk = LEX_TYPEEQUAL;
jslGetNextCh(lex);
}
} break;
case JSLJT_LESSTHAN: jslSingleChar(lex);
if (lex->currCh=='=') { // <=
lex->tk = LEX_LEQUAL;
jslGetNextCh(lex);
} else if (lex->currCh=='<') { // <<
lex->tk = LEX_LSHIFT;
jslGetNextCh(lex);
if (lex->currCh=='=') { // <<=
lex->tk = LEX_LSHIFTEQUAL;
jslGetNextCh(lex);
}
} break;
case JSLJT_GREATERTHAN: jslSingleChar(lex);
if (lex->currCh=='=') { // >=
lex->tk = LEX_GEQUAL;
jslGetNextCh(lex);
} else if (lex->currCh=='>') { // >>
lex->tk = LEX_RSHIFT;
jslGetNextCh(lex);
if (lex->currCh=='=') { // >>=
lex->tk = LEX_RSHIFTEQUAL;
jslGetNextCh(lex);
} else if (lex->currCh=='>') { // >>>
jslGetNextCh(lex);
if (lex->currCh=='=') { // >>>=
lex->tk = LEX_RSHIFTUNSIGNEDEQUAL;
jslGetNextCh(lex);
} else {
lex->tk = LEX_RSHIFTUNSIGNED;
}
}
} break;
case JSLJT_SINGLECHAR: jslSingleChar(lex); break;
default: assert(0);break;
}
}
}
}
