🚀 go-pugleaf

In article <14382@ksr.com>, jfw@ksr.com (John F. Woods) writes:
|> volpe@bart.NoSubdomain.NoDomain (Christopher R Volpe) writes:
|> >There are no "two interpretations". That's the whole point. The "other
|> >interpretation" was shot down by the strictly conforming program that
|> >prints different results under the non-literal interpretation.
|> 
|> No.  The program given was strictly conforming if and only if the "implied"
|> forbidding of "unnecessary" padding is real.  I claim that there is nothing

True. I wasn't clear enough. The "other interpretation" that I was
referring to (the invalid one) is the one that says that unnecessary
padding can be show to be acceptable via the "as-if" rule.

|> in the rest of the Standard that allows one to deduce this (unlike the "all
|> structure pointers are the same size" rule), and that the words "may not" do
|> not appear in either 
|> 
|> 	"There may be unnamed padding within a structure object, but not
|> 	at its beginning, as necessary to achieve the appropriate alignment."

Ok, there is definitely room for interpretation here. For some reason
I remember the wording being something like "...only as necessary to achieve",
which would forbid unnecessary padding. Anyway, I stand corrected.

|> As I said, whoever wrote the Rationale also appears to believe that structures
|> may have internal padding that strict parsimony would reject.

Agreed.

-- 
==================
Chris Volpe
G.E. Corporate R&D
volpecr@crd.ge.com