llStripCharacters

With the introduction of Unicode font support in 1.6, it occurs to me that at least a few of my scripts simply can't have complete input filtering any longer. There's no way I can make a string machine-readable and meet all eventualities of punctuation anymore.

With that in mind, I suggest llStripCharacters:
string llStripCharacters(string src, integer charset) where 'charset' is ASCII, alphanumeric, alphabetical, uppercase, lowercase, numerical, unicode, and so on. This would allow the removal of all characters not meeting the criteria and would allow for far better data checking.

For example, if I currently were to try to look for the words "hello how are you" in a larger string, case isn't an issue, but there could be any number of punctuation marks, depending on the source's familiarity with the English language. Yes, this is sadly the best example I can think of at the moment, but if it's come up more than once with me, surely such a function would appeal to someone else as well.

I can see this being a problem. I endorse this feature.

with a title like llStripCharacter, i thought of something else lol

I hope we get something like llUnicodeValue(string) that gets the integer value of the string's characters (the total, if there is more then one character passed to it). Unicode escape strings (\uXXXX) would be nice too.

Heck, just put in support for the full set of Perl regular-expression operations.

I agree with Carnildo. Hook up a regular expression library like PCRE...

llRegexMatch(string str, string regex, integer modifiers);
string llRegexReplace(string str, string regex, string replacement, integer modifiers);

...where modifiers is a bitmask for things like REGEX_CASE_INSENSITIVE.

# matching in Perl
$str = 'foobarbaz';
$foo = 'bar';

if ($str =~ /\w{3}$foo\w{3}/)
{
   # do something
}

// matching in LSL
string str = "foobarbaz";
string foo = "bar";

if (llRegexMatch(str, "\w{3}" + foo + "\w{3}", 0))
{
   // do something
}

# replacement in perl
$str = 'The Philadelphia Experiment';
$str =~ s/^(an?|the)\s*(.*)$/$2, $1/i;

// replacement in LSL
string str = "The Philadelphia Experiment";
str = llRegexReplace(str, "^(an?|the)\s*(.*)$", "$2, $1", REGEX_CASE_INSENSITIVE);

Thoughts? I think an approach like this would give us a lot of power. Using methods instead of providing first-class support for regexes in LSL also make the change much less drastic.

Yes, regex is really what I want, but I've always come away with the feeling that it's a little much to hope for.

Oh wow regex in SL would be so SEXY!!

Please do it Lindens! PLEASE, for the love of Perl!

--Jack Lambert

and here I was thinking llStripCharacters was going to let us see underneath an avatar's clothes.

I guess that would be llStripAvatar