internal IScriptable [] createRegExpWraps(Context cx, IScriptable scope)
{
if (idata.itsRegExpLiterals == null)
{
Context.CodeBug();
}
RegExpProxy rep = cx.RegExpProxy;
int N = idata.itsRegExpLiterals.Length;
IScriptable [] array = new IScriptable [N];
for (int i = 0; i != N; ++i)
{
array [i] = rep.Wrap(cx, scope, idata.itsRegExpLiterals [i]);
}
return(array);
}
public static void SetRegExpProxy(Context cx, RegExpProxy proxy)
{
if (proxy == null)
{
throw new ArgumentException();
}
cx.regExpProxy = proxy;
}
/*
* Used by js_split to find the next split point in target,
* starting at offset ip and looking either for the given
* separator substring, or for the next re match. ip and
* matchlen must be reference variables (assumed to be arrays of
* length 1) so they can be updated in the leading whitespace or
* re case.
*
* Return -1 on end of string, >= 0 for a valid index of the next
* separator occurrence if found, or the string length if no
* separator is found.
*/
private static int find_split(Context cx, IScriptable scope, string target, string separator, Context.Versions version, RegExpProxy reProxy, IScriptable re, int [] ip, int [] matchlen, bool [] matched, string [] [] parensp)
{
int i = ip [0];
int length = target.Length;
/*
* Perl4 special case for str.split(' '), only if the user has selected
* JavaScript1.2 explicitly. Split on whitespace, and skip leading w/s.
* Strange but true, apparently modeled after awk.
*/
if (version == Context.Versions.JS1_2 && re == null && separator.Length == 1 && separator [0] == ' ') {
/* Skip leading whitespace if at front of str. */
if (i == 0) {
while (i < length && char.IsWhiteSpace (target [i]))
i++;
ip [0] = i;
}
/* Don't delimit whitespace at end of string. */
if (i == length)
return -1;
/* Skip over the non-whitespace chars. */
while (i < length && !char.IsWhiteSpace (target [i]))
i++;
/* Now skip the next run of whitespace. */
int j = i;
while (j < length && char.IsWhiteSpace (target [j]))
j++;
/* Update matchlen to count delimiter chars. */
matchlen [0] = j - i;
return i;
}
/*
* Stop if past end of string. If at end of string, we will
* return target length, so that
*
* "ab,".split(',') => new Array("ab", "")
*
* and the resulting array converts back to the string "ab,"
* for symmetry. NB: This differs from perl, which drops the
* trailing empty substring if the LIMIT argument is omitted.
*/
if (i > length)
return -1;
/*
* Match a regular expression against the separator at or
* above index i. Return -1 at end of string instead of
* trying for a match, so we don't get stuck in a loop.
*/
if (re != null) {
return reProxy.FindSplit (cx, scope, target, separator, re, ip, matchlen, matched, parensp);
}
/*
* Deviate from ECMA by never splitting an empty string by any separator
* string into a non-empty array (an array of length 1 that contains the
* empty string).
*/
if (version != Context.Versions.Default && version < Context.Versions.JS1_3 && length == 0)
return -1;
/*
* Special case: if sep is the empty string, split str into
* one character substrings. Let our caller worry about
* whether to split once at end of string into an empty
* substring.
*
* For 1.2 compatibility, at the end of the string, we return the length as
* the result, and set the separator length to 1 -- this allows the caller
* to include an additional null string at the end of the substring list.
*/
if (separator.Length == 0) {
if (version == Context.Versions.JS1_2) {
if (i == length) {
matchlen [0] = 1;
return i;
}
return i + 1;
}
return (i == length) ? -1 : i + 1;
}
/* Punt to j.l.s.indexOf; return target length if seperator is
* not found.
*/
if (ip [0] >= length)
return length;
i = target.IndexOf (separator, ip [0]);
return (i != -1) ? i : length;
}
/*
* See ECMA 15.5.4.8. Modified to match JS 1.2 - optionally takes
* a limit argument and accepts a regular expression as the split
* argument.
*/
private static object ImplSplit(Context cx, IScriptable scope, string target, object [] args)
{
// create an empty Array to return;
IScriptable top = GetTopLevelScope(scope);
IScriptable result = ScriptRuntime.NewObject(cx, top, "Array", null);
// return an array consisting of the target if no separator given
// don't check against undefined, because we want
// 'fooundefinedbar'.split(void 0) to split to ['foo', 'bar']
if (args.Length < 1)
{
result.Put(0, result, target);
return(result);
}
// Use the second argument as the split limit, if given.
bool limited = (args.Length > 1) && (args [1] != Undefined.Value);
long limit = 0; // Initialize to avoid warning.
if (limited)
{
/* Clamp limit between 0 and 1 + string length. */
limit = ScriptConvert.ToUint32(args [1]);
if (limit > target.Length)
{
limit = 1 + target.Length;
}
}
string separator = null;
int [] matchlen = new int [1];
IScriptable re = null;
RegExpProxy reProxy = null;
if (args [0] is IScriptable)
{
reProxy = cx.RegExpProxy;
if (reProxy != null)
{
IScriptable test = (IScriptable)args [0];
if (reProxy.IsRegExp(test))
{
re = test;
}
}
}
if (re == null)
{
separator = ScriptConvert.ToString(args [0]);
matchlen [0] = separator.Length;
}
// split target with separator or re
int [] ip = new int [] { 0 };
int match;
int len = 0;
bool [] matched = new bool [] { false };
string [] [] parens = new string [] [] { null };
Context.Versions version = cx.Version;
while ((match = find_split(cx, scope, target, separator, version, reProxy, re, ip, matchlen, matched, parens)) >= 0)
{
if ((limited && len >= limit) || (match > target.Length))
{
break;
}
string substr;
if (target.Length == 0)
{
substr = target;
}
else
{
substr = target.Substring(ip [0], (match) - (ip [0]));
}
result.Put(len, result, substr);
len++;
/*
* Imitate perl's feature of including parenthesized substrings
* that matched part of the delimiter in the new array, after the
* split substring that was delimited.
*/
// CB, 02.01.2007: Don't do this, causes bug #287630
// https://bugzilla.mozilla.org/show_bug.cgi?query_format=specific&order=relevance+desc&bug_status=__open__&id=287630
/*
* if (re != null && matched [0] == true) {
* int size = parens [0].Length;
* for (int num = 0; num < size; num++) {
* if (limited && len >= limit)
* break;
* result.Put (len, result, parens [0] [num]);
* len++;
* }
* matched [0] = false;
* }
*/
ip [0] = match + matchlen [0];
if (version < Context.Versions.JS1_3 && version != Context.Versions.Default)
{
/*
* Deviate from ECMA to imitate Perl, which omits a final
* split unless a limit argument is given and big enough.
*/
if (!limited && ip [0] == target.Length)
{
break;
}
}
}
return(result);
}
/*
* Used by js_split to find the next split point in target,
* starting at offset ip and looking either for the given
* separator substring, or for the next re match. ip and
* matchlen must be reference variables (assumed to be arrays of
* length 1) so they can be updated in the leading whitespace or
* re case.
*
* Return -1 on end of string, >= 0 for a valid index of the next
* separator occurrence if found, or the string length if no
* separator is found.
*/
private static int find_split(Context cx, IScriptable scope, string target, string separator, Context.Versions version, RegExpProxy reProxy, IScriptable re, int [] ip, int [] matchlen, bool [] matched, string [] [] parensp)
{
int i = ip [0];
int length = target.Length;
/*
* Perl4 special case for str.split(' '), only if the user has selected
* JavaScript1.2 explicitly. Split on whitespace, and skip leading w/s.
* Strange but true, apparently modeled after awk.
*/
if (version == Context.Versions.JS1_2 && re == null && separator.Length == 1 && separator [0] == ' ')
{
/* Skip leading whitespace if at front of str. */
if (i == 0)
{
while (i < length && char.IsWhiteSpace(target [i]))
{
i++;
}
ip [0] = i;
}
/* Don't delimit whitespace at end of string. */
if (i == length)
{
return(-1);
}
/* Skip over the non-whitespace chars. */
while (i < length && !char.IsWhiteSpace(target [i]))
{
i++;
}
/* Now skip the next run of whitespace. */
int j = i;
while (j < length && char.IsWhiteSpace(target [j]))
{
j++;
}
/* Update matchlen to count delimiter chars. */
matchlen [0] = j - i;
return(i);
}
/*
* Stop if past end of string. If at end of string, we will
* return target length, so that
*
* "ab,".split(',') => new Array("ab", "")
*
* and the resulting array converts back to the string "ab,"
* for symmetry. NB: This differs from perl, which drops the
* trailing empty substring if the LIMIT argument is omitted.
*/
if (i > length)
{
return(-1);
}
/*
* Match a regular expression against the separator at or
* above index i. Return -1 at end of string instead of
* trying for a match, so we don't get stuck in a loop.
*/
if (re != null)
{
return(reProxy.FindSplit(cx, scope, target, separator, re, ip, matchlen, matched, parensp));
}
/*
* Deviate from ECMA by never splitting an empty string by any separator
* string into a non-empty array (an array of length 1 that contains the
* empty string).
*/
if (version != Context.Versions.Default && version < Context.Versions.JS1_3 && length == 0)
{
return(-1);
}
/*
* Special case: if sep is the empty string, split str into
* one character substrings. Let our caller worry about
* whether to split once at end of string into an empty
* substring.
*
* For 1.2 compatibility, at the end of the string, we return the length as
* the result, and set the separator length to 1 -- this allows the caller
* to include an additional null string at the end of the substring list.
*/
if (separator.Length == 0)
{
if (version == Context.Versions.JS1_2)
{
if (i == length)
{
matchlen [0] = 1;
return(i);
}
return(i + 1);
}
return((i == length) ? -1 : i + 1);
}
/* Punt to j.l.s.indexOf; return target length if seperator is
* not found.
*/
if (ip [0] >= length)
{
return(length);
}
i = target.IndexOf(separator, ip [0]);
return((i != -1) ? i : length);
}
private static int Find_split(Context cx, Scriptable scope, string target, string separator, int version, RegExpProxy reProxy, Scriptable re, int[] ip, int[] matchlen, bool[] matched, string[][] parensp)
{
int i = ip[0];
int length = target.Length;
if (version == Context.VERSION_1_2 && re == null && separator.Length == 1 && separator[0] == ' ')
{
if (i == 0)
{
while (i < length && char.IsWhiteSpace(target[i]))
{
i++;
}
ip[0] = i;
}
if (i == length)
{
return -1;
}
while (i < length && !char.IsWhiteSpace(target[i]))
{
i++;
}
int j = i;
while (j < length && char.IsWhiteSpace(target[j]))
{
j++;
}
matchlen[0] = j - i;
return i;
}
if (i > length)
{
return -1;
}
if (re != null)
{
return reProxy.Find_split(cx, scope, target, separator, re, ip, matchlen, matched, parensp);
}
if (version != Context.VERSION_DEFAULT && version < Context.VERSION_1_3 && length == 0)
{
return -1;
}
if (separator.Length == 0)
{
if (version == Context.VERSION_1_2)
{
if (i == length)
{
matchlen[0] = 1;
return i;
}
return i + 1;
}
return (i == length) ? -1 : i + 1;
}
if (ip[0] >= length)
{
return length;
}
i = target.IndexOf(separator, ip[0]);
return (i != -1) ? i : length;
}
