static ejsval
Encode (ejsval string, ejsval unescaped)
{
EJSPrimString *stringObj = EJSVAL_TO_STRING(string);
jschar *unescapedStr = EJSVAL_TO_FLAT_STRING(unescaped);
/* 1. Let strLen be the number of code units in string. */
int32_t strLen = stringObj->length;
/* 2. Let R be the empty String. */
ejsval R = _ejs_atom_empty;
/* 3. Let k be 0. */
int32_t k = 0;
/* 4. Repeat. */
for (;;) {
/* a. If k equals strLen, return R. */
if (k == strLen)
return R;
/* b. Let C be the code unit at index k within string. */
jschar C = _ejs_string_ucs2_at (stringObj, k);
jschar C_arr [2] = { C, 0 };
/* c. If C is in unescapedSet, then */
jschar *p = ucs2_strstr (unescapedStr, C_arr);
if (p) {
/* i. Let S be a String containing only the code unit C. */
ejsval S = _ejs_string_new_substring (string, k, 1);
/* ii. Let R be a new String value computed by concatenating the previous value of R and S. */
R = _ejs_string_concat (R, S);
}
/* d. Else C is not in unescapedSet, */
else {
/* i. If the code unit value of C is not less than 0xDC00 and not greater than 0xDFFF, throw a URIError exception. */
if (C >= 0xDC00 && C <= 0xDFFF)
_ejs_throw_nativeerror_utf8 (EJS_URI_ERROR, "URI malformed");
/* ii. If the code unit value of C is less than 0xD800 or greater than 0xDBFF, then Let V be the code unit value of C. */
jschar V;
if (C < 0xD800 || C > 0xDBFF)
V = C;
/* iii. Else, */
else {
/* 1. Increase k by 1. */
k++;
/* 2. If k equals strLen, throw a URIError exception. */
if (k == strLen)
_ejs_throw_nativeerror_utf8 (EJS_URI_ERROR, "URI malformed");
/* 3. Let kChar be the code unit value of the code unit at index k within string. */
jschar kChar = _ejs_string_ucs2_at (stringObj, k);
/* 4. If kChar is less than 0xDC00 or greater than 0xDFFF, throw a URIError exception. */
if (kChar < 0xDC00 || kChar > 0xDFFF)
_ejs_throw_nativeerror_utf8 (EJS_URI_ERROR, "URI malformed");
/* 5. Let V be (((the code unit value of C) – 0xD800) × 0x400 + (kChar – 0xDC00) + 0x10000). */
V = (C - 0xD800) * 0x400 + (kChar - 0xDC00) + 0x10000;
}
/* iv. Let Octets be the array of octets resulting by applying the UTF-8 transformation to V, and let L be the array size. */
char octets[4];
int32_t L = ucs2_to_utf8_char (V, octets);
/* v. Let j be 0. */
int32_t j = 0;
/* vi. Repeat, while j < L. */
while (j < L) {
/* 1. Let jOctet be the value at index j within Octets. */
char jOctet = octets [j];
/* 2. Let S be a String containing three code units “%XY” where XY are two uppercase hexadecimal
* digits encoding the value of jOctet. */
char buff[4];
sprintf(buff, "%%%X", jOctet);
ejsval S = _ejs_string_new_utf8 (buff);
/* 3. Let R be a new String value computed by concatenating the previous value of R and S. */
R = _ejs_string_concat (R, S);
/* 4. Increase j by 1. */
j++;
}
}
/* e. Increase k by 1. */
k++;
}
}
ejsval
_ejs_regexp_replace(ejsval str, ejsval search_re, ejsval replace)
{
EJSRegExp* re = (EJSRegExp*)EJSVAL_TO_OBJECT(search_re);
pcre16_extra extra;
memset (&extra, 0, sizeof(extra));
pcre16* code = (pcre16*)re->compiled_pattern;
int capture_count;
pcre16_fullinfo (code, NULL, PCRE_INFO_CAPTURECOUNT, &capture_count);
int ovec_count = 3 * (1 + capture_count);
int* ovec = malloc(sizeof(int) * ovec_count);
int cur_off = 0;
do {
EJSPrimString *flat_str = _ejs_string_flatten (str);
jschar *chars_str = flat_str->data.flat;
int rv = pcre16_exec(code, &extra,
chars_str, flat_str->length, cur_off,
PCRE_NO_UTF16_CHECK, ovec, ovec_count);
if (rv < 0)
break;
ejsval replaceval;
if (EJSVAL_IS_FUNCTION(replace)) {
ejsval substr_match = _ejs_string_new_substring (str, ovec[0], ovec[1] - ovec[0]);
ejsval capture = _ejs_string_new_substring (str, ovec[2], ovec[3] - ovec[2]);
_ejs_log ("substring match is %s\n", ucs2_to_utf8(_ejs_string_flatten(substr_match)->data.flat));
_ejs_log ("capture is %s\n", ucs2_to_utf8(_ejs_string_flatten(capture)->data.flat));
int argc = 3;
ejsval args[3];
args[0] = substr_match;
args[1] = capture;
args[2] = _ejs_undefined;
replaceval = ToString(_ejs_invoke_closure (replace, _ejs_undefined, argc, args));
}
else {
replaceval = ToString(replace);
}
if (ovec[0] == 0) {
// we matched from the beginning of the string, so nothing from there to prepend
str = _ejs_string_concat (replaceval, _ejs_string_new_substring (str, ovec[1], flat_str->length - ovec[1]));
}
else {
str = _ejs_string_concatv (_ejs_string_new_substring (str, 0, ovec[0]),
replaceval,
_ejs_string_new_substring (str, ovec[1], flat_str->length - ovec[1]),
_ejs_null);
}
cur_off = ovec[1];
// if the RegExp object was created without a 'g' flag, only replace the first match
if (!re->global)
break;
} while (EJS_TRUE);
free (ovec);
return str;
}