Background: The ECL language ref indicates that the regex functions support unicode and point towards the ICU documentation for information about regular expression metacharacters and operators (http://userguide.icu-project.org/strings/regexp). That documentation, in turn, references another page regarding the properties that are supported via metacharacters (http://userguide.icu-project.org/strings/properties).
Here is a snippet of code testing that support. The intention is to create a 'data pattern' result (uppercase characters become 'A', lowercase becomes 'a', etc.) that is unicode-friendly.
This fails during execution but does not show an error in the workunit. The error occurred in the compiling phase:
So there may be two items here: 1) Failure of an error message appearing in the workunit, and 2) failure to properly compile the metacharacter support for an underlying ICU call.