🚀 go-pugleaf

In article <4a73n0$jfu@newshost.lanl.gov>, tanmoy@qcd.lanl.gov (Tanmoy Bhattacharya) writes:
|>In article <4a6r9u$hq5@rdsunx.crd.ge.com>, volpe@bart.crd.ge.com (Christopher
|>R. Volpe) writes: 
|><snip>
|>|> |>this to satisfy the _unsatisfiable_ edict of `statements are evaluated
|>|> |>in sequence': after all `f()+f();' is a statement --- is it done
|>|> |>before or after the statements comprising f? 
|>|> 
|>|> It's not unsatisfiable, because the edict includes the phrase "except where
|>|> otherwise specified", and function calls are one of those exceptions.
|>
|>And 6.3 `the order of evaluation of subexpressions and ... are both
|>unspecified' (i.e. the clause which allowed interleaved expressions in the
|>first place) does not count as `except where otherwise specified'?

In my opinion, no. What does the order of evaluation of subexpressions have to
do with statement sequencing? (Except in the case of function calls, which I
already mentioned.)

|> If the
|>sequence point on both sides of i++ still leaves i + (a,i++,b) undefined; why
|>should i + f() where f() is a; i++; b; be any different?

Because the commas in "(a,i++,b)" do not delimit statements.

|>
|>Or rather, not why should it be different: it can be different by fiat. But,
|>I do not see even the genesis of the claim that the standard currently
|>specifies it clearly.

I admit that it's not so cut and dry. There is definitely room for argument
here. But in my opinion, Once control is transferred to the function, leaving
the function to go  compute something else in the calling expression
constitutes a break in the statement sequence.

--

Chris Volpe				Phone: (518) 387-7766 (Dial Comm 8*833
GE Corporate R&D			Fax:   (518) 387-6560
PO Box 8, Schenectady, NY 12301		Email: volpecr@crd.ge.com