You mean, I dreamed the bugs, the confirmation mails, and the patches?
Of course I do. It has compiled my whole Linux work environment, which hasn't crashed once during the last years. "Doesn't work" looks different.
Even if I had tested it against Plum Hall, that would only prove that it correctly compiles Plum Hall. I don't want to compile Plum Hall. I want to compile my code.
"What constitutes an access to an object that has volatile-qualified type is implementation-defined." Most people who say that the compiler is wrong just have a different imagination of what is an access. For example, I would never think 'i;' were one (however, the 'WATCHDOG = WATCHDOG' example is a clear bug under all definitions of volatile I know).
That aside, I've always said that if you want assembler sequence 'X' generated, you write assembler sequence 'X', and not some C fragment that happens to generate assembler sequence 'X'.
Neither is it a guarantee if the instance it appears to work in is Plum Hall.
Can you still buy Turbo C? Coherent?
Stefan