It's therefore quite surprising that this doesn't actually happen if EXECLINE_STRICT is set in the environment, which it may very well be in the likely case that wait is invoked from an execline script:
You're right, it's an oversight on my part - and also proof that every time I go against a design principle for convenience, it comes back to bite me. Here, the design principle is "library functions should not depend on global values", and el_semicolon() breaks this because it fails when EXECLINE_STRICT is set and it can't find a block. I will fix it for the next release; this requires some refactoring. Thanks for the report! -- Laurent
