if bug in Ritchie C compiler

[utzoo!decvax!yale-com!leichter]

The text you quote explicitly gives the compiler to RE-ARRANGE expressions
and evaluate their component parts in any order it likes.  It DOES NOT, as
it stands, give it permission to skip the evaluation of any part.  I think
if you look up the discussion of, say, & vs. &&, you will find some remarks
to the effect that && only evaluates as much as it needs, but & evaluates
both arguments.  What, then, is one to make of 0&foo()?
							-- Jerry
						decvax!yale-comix!leichter