Path: csiph.com!v102.xanadu-bbs.net!xanadu-bbs.net!feeder.erje.net!eu.feeder.erje.net!news.swapon.de!fu-berlin.de!uni-berlin.de!individual.net!not-for-mail
From: Helge Blischke <h.blischke@acm.org>
Newsgroups: comp.lang.postscript
Subject: Re: How does ghostscript make ints and reals equivalent as dictionary keys??!!!
Followup-To: comp.lang.postscript
Date: Fri, 23 Aug 2013 16:37:37 +0200
Lines: 48
Message-ID: <b7padjFj02tU1@mid.individual.net>
References: <5a1784ab-352f-403e-878d-39740b1b23cd@googlegroups.com>
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: 7Bit
X-Trace: individual.net zHQZFwj4VgWB/q9tcm9kzAcJ3TdAHfD+NoZas617vP4abDiKQZ
Cancel-Lock: sha1:DJotj4BV9CFLcUb1y62uCgGfQWk=
User-Agent: KNode/0.99.01
Xref: csiph.com comp.lang.postscript:1584

luser- -droog wrote:

> I've marveled at this behavior before. But I just can't imagine how
> they do that!
> 
> I suppose in many languages (PERL, Lua, JavaScript), the distinction
> between ints and reals is very blurry all the way down the pipeline.
> But postscript makes them distinct types for many operations
> (array and string indexing, stack indexing, various other 'count' or
> ordinal data), but blurs them during math. And C makes the distinction
> very explicit, but permits 'usual implicit conversions' to provide
> automatic promotions where appropriate.
> 
> All of this is fine, and where an operator needs one or the other
> it can check and signal an error; or when an operator can accept
> either, it can check and promote if necessary before merging
> the int and real cases. But...
> 
> But how do you make them hash equally so you can look them up in a dict?
> 
> My only guess is that ghostscript must merge them somehow into a single
> canonical form: some kind of extended floating-integer hybrid super-duper
> hackjob type.
> 
> But then, how do you do the 'super-duper' part?

Integers and reals don#t seem to be equivalent as dictionary keys.
For instance, if you define (in Ghostscript)

/1 (Heureka) def

currentdict/1 .knownget

puts 
(Heureka) true
onto the operand stack, but

currentdict/1.0 .knownget

yields only

false

on the operand stack.

Helge