Re: [colorforth] Dare I say ANS!

["Roman Pavlyuk \(personal\)" <john@xxxxxxxxxxxxx>]

>>> > > ?dup          ????
>>> >
>>> > ?DUP is ISO

>>> Sure it's ISO, but it's not part of the compiler's peep-hole
optimization
>>> scheme.  However, I would like to point out that Chuck doesn't
>>> particularly like words that return a variable number of results on the
>>> stack.  Indeed, they put dataflow-based optimizers into fits.

Necessaity of F-83 style ?DUP is obsoleted by non-dropping IF. And cyan
?dup/drop are quite different from ?DUP ( N -- N N | 0 )

My 2c to all this discussion: I do not care about compatibility with FIG,
F83, ANSI, whatever, because for more or less long term projects (we all
like these, don't we? :) there's a big sense in not using any legacy code
nor any third party components (at least, w/o thorough code review). So, for
me it's better have all primitives documented rather that rely of knowing
that semantics has not changed.

Roman


---------------------------------------------------------------------
To unsubscribe, e-mail: colorforth-unsubscribe@xxxxxxxxxxxxxxxxxx
For additional commands, e-mail: colorforth-help@xxxxxxxxxxxxxxxxxx
Main web page - http://www.colorforth.com