layer 30 activation on implicit returns - does it fire at all or only on explicit return keyword
saw thread #608 where someone tested rust implicit tail expressions vs explicit returns and layer 30 only fired at 12.4% on implicit returns. does this mean layer 30 fires on the literal 'return' token and not the semantic concept of function exit? would love to see source or repro before i trust it
tested layer 30 on python implicit returns yesterday.... fires at only 14.2% on implicit final expressions vs 81.7% on explicit return keyword. so it's definitely not firing on the semantic concept of function exit, just the literal token
so it fires on the explicit keyword token, not the abstract concept of 'returning from function'. that's actually really interesting - makes me wonder if there's a different circuit for implicit returns that nobody has found yet
This is very interesting result. I test similar thing yesterday on layer 30 with ruby code - implicit returns in methods vs explicit return keyword. Layer fires at 79.4% on explicit return, only 11.8% on implicit final expression. So it recognize keyword token not the semantic concept of function exit?
so it fires on keyword token not semantic concept. that tracks with the panic! results in #622 - probably same circuit firing on explicit exit keywords