> The original discussion was about inserting nondeletable labels via inline > assembly statements. Yes, but what on earth is the semantics that the author has in mind in inserting such labels, I can think of many possible purposes (and in each case, it is highly dubious that the purpose would be achieved).