ok, I switched back to using the short-hand syntax and got the regex working for latin chars:

Code:
var rtl_match: RegExp = /[\u0041-\u007A]/g
sub in the hebrew range:

Code:
var rtl_match: RegExp = /[\u0590–\u05FF]/g;
and the damn thing stops working!

you can see for yourself if you copy and paste this into a new fla:

Code:
//var rtl_match: RegExp = /[\u0041-\u007A]/g;
var rtl_match: RegExp = /[\u0590–\u05FF]/g;

var heb: String = 'השידור לא זמין. אנא נסה שוב או בחר בשידור אחר.';
var lat: String = 'abcdefghijklmnopqrstuvwxyzABCDEFGHIJLKMNOPQRSTUVWXYZ1234567890';

trace(heb.replace(rtl_match, '!'));
trace(lat.replace(rtl_match, '!'))
switch between regex's by disabling/enabling lines 1/2

interestingly using the hebrew match on 'lat' matches characters: uF059 the characters used in the regex suggesting it is no longer treating it as unicode syntax