[E-Lang] E Project, Documentation, and new programmers

Dan Bornstein danfuzz@milk.com
Thu, 12 Apr 2001 13:46:20 -0700 (PDT)


I remember that discussion, but I thought the discussion on the table then
was about having to explicitly reveal values inside a method (I think
that's why the term "reveal" was bandied about, that is, to make it
explicit that it wasn't just for returning values from methods), not about
just having a return-like thing. If it's that, then we have a new point of
debate here. If not, then, well, it's a year and a half since we last
discussed this, and maybe it's time to reexamine the issue.

-dan

Mark S. Miller writes:
>At 11:43 AM Thursday 4/12/01, Dan Bornstein wrote:
>>Okay, then, here's possibly a better way to state the proposal:
>>
>>    Every method definition should have an implicit binding of the
>>    identifier "return" to an escape function which, if called, returns
>>    from the method with the given value, and there should not be any
>>    implicit return value from a method (that is, falling off the end
>>    of a *method*--not just any block--is tantamount to returning null).
>
>Well, it's certainly a better way to state this, as I didn't even understand 
>that you were proposing this.  In any case, isn't this one of the proposals 
>we considered and rejected back in the unary-^ discussions?  (Rooted 
>http://www.eros-os.org/~majordomo/e-lang/0844.html , 
>http://www.eros-os.org/~majordomo/e-lang/0895.html , and 
>http://www.eros-os.org/~majordomo/e-lang/0969.html .)