> "Paul Fisher" <[EMAIL PROTECTED]> writes:
>
> "John Keiser" <[EMAIL PROTECTED]> writes:
>
> > The only plausible solution, then, that ports to javac is the use of
> > some sort of no-kiev comment style.
>
> Why bother with require/ensure equivalents for javac? If you use Kiev
> you win, if you use javac you lose. I don't see a big problem with
> that.
>
I was considering require/ensure to be analogous to checking for abnormal
conditions, which many parts of java.* do. If this is the case, then if we
*only* check for the condition in require/ensure, then it won't get checked
for in javac-compiled versions that need to do so. If we *do* put an
explicit check in in addition to a require/ensure check, we check twice.
The only solution then is never to put something that is *supposed* to be
checked by the library into a require statement, which as I understand it
loses some of the benefits of require.
> I also don't particular like this whole "comment this and Kiev will
> understand it" and "comment this and Kiev will ignore it". It makes
> code ugly.
>
> Why not just use require/ensure in the normal Kiev manner and use sed
> to augment the source if Classpath is being compiled with a normal
> java compiler?
>
If we need the code in require/ensure to show up on javac, a sed script will
not cut it since require/ensure statements are inherited. Of course, my
fallacy may be that we don't need require/ensure statements to have
analogous representations in the javac-compiled version.
--John Keiser
