Path: csiph.com!x330-a1.tempe.blueboxinc.net!usenet.pasdenom.info!news.chainon-marquant.org!nntpfeed.proxad.net!proxad.net!feeder2-2.proxad.net!news.osn.de!diablo2.news.osn.de!news.belwue.de!rz.uni-karlsruhe.de!feed.news.schlund.de!schlund.de!news.online.de!not-for-mail
From: Bernd Paysan <bernd.paysan@gmx.de>
Newsgroups: comp.lang.forth
Subject: Re: return address manipulation
Date: Wed, 14 Dec 2011 14:08:34 +0100
Organization: 1&1 Internet AG
Lines: 30
Message-ID: <jca74j$o1f$1@online.de>
References: <jbds1r$1v8$1@dont-email.me> <VLKdnayez7QlZEHTnZ2dnUVZ_gOdnZ2d@supernews.com> <jbki7m$uq3$2@dont-email.me> <4eddf07f.57544369@192.168.0.50> <jbl28o$c6e$1@dont-email.me> <v-adnQ8WrLMA30PTnZ2dnUVZ_qmdnZ2d@supernews.com> <jbman1$nu6$1@online.de> <vNKdnTQRr_RwrkLTnZ2dnUVZ_hidnZ2d@supernews.com> <4edf3e69.143030702@192.168.0.50> <2011Dec7.142922@mips.complang.tuwien.ac.at> <OvadnU1Pp7QENULTnZ2dnUVZ_rydnZ2d@supernews.com> <2011Dec8.165153@mips.complang.tuwien.ac.at> <If2dnZAUYMzJYn3TnZ2dnUVZ_v2dnZ2d@supernews.com> <6MqdnZRLdI0XP3rTnZ2dnUVZ_u6dnZ2d@supernews.com>
NNTP-Posting-Host: p5dcd67af.dip0.t-ipconnect.de
Mime-Version: 1.0
Content-Type: text/plain; charset="ISO-8859-1"
Content-Transfer-Encoding: 7Bit
X-Trace: online.de 1323868115 24623 93.205.103.175 (14 Dec 2011 13:08:35 GMT)
X-Complaints-To: abuse@einsundeins.com
NNTP-Posting-Date: Wed, 14 Dec 2011 13:08:35 +0000 (UTC)
User-Agent: KNode/4.7.2
Xref: x330-a1.tempe.blueboxinc.net comp.lang.forth:8062

Andrew Haley wrote:
> Turns out it's pretty easy on x86 SwiftForth, but not exactly trivial:
> 
> icode (quotation)
>    here $400 + call
>    ret   end-code
> 
> : [:   postpone (quotation)  postpone begin ;  immediate
> : ;]   postpone exit  postpone then
>     postpone r>  postpone code> ;   immediate
> 
> This takes advantage of the fact that the return stack really is the
> processor's return stack and the rather nice ICODE allows the easy
> insertion of assembly language fragments.  It even generates pretty
> efficient code.  RECURSE doesn't work properly, though; is that a
> problem?  Do we even know what RECURSE should to inside a quotation?

Probably recurse the nested definition.

> The capture of locals' values would be nice to have, but perhaps
> that's rather too much.

Ah, no closures.  Just nested anonymous definitions.  What would be nice 
is when the nested definition could have their own locals.

-- 
Bernd Paysan
"If you want it done right, you have to do it yourself"
http://bernd-paysan.de/