Path: csiph.com!eternal-september.org!feeder.eternal-september.org!news.eternal-september.org!.POSTED!not-for-mail
From: rickman <gnuarm@gmail.com>
Newsgroups: comp.lang.forth
Subject: Re: EXIT THEN
Date: Mon, 13 Feb 2017 00:02:01 -0500
Organization: A noiseless patient Spider
Lines: 56
Message-ID: <o7reh0$js7$1@dont-email.me>
References: <o5jrct$1r41$1@gioia.aioe.org> <o6v4tf$k59$1@cherry.spenarnc.xs4all.nl> <o7j45j$19ee$1@gioia.aioe.org> <f6928d15-aebe-4dcf-91f4-0fbc88b32827@googlegroups.com> <o7lq8n$1jm5$1@gioia.aioe.org> <db3573f7-a2f8-4a6f-8270-7adfcec20b83@googlegroups.com> <20170211075559.5d4b2fc3@_> <o7n2nd$fkf$1@dont-email.me> <-9WdneZj9upxHALFnZ2dnUU7-T3NnZ2d@supernews.com> <877f4wv22e.fsf@nightsong.com> <88caa00e-b9e6-4620-95a2-fd232a8f49a9@googlegroups.com> <20170212120909.5e42585f@t530.lan> <a1bf9577-6734-4bed-ad3b-1f5be6856dd3@googlegroups.com> <20170212164250.22867950@_> <6177f3b6-c5f1-4909-871a-2f4c533413d6@googlegroups.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=utf-8; format=flowed
Content-Transfer-Encoding: 7bit
Injection-Date: Mon, 13 Feb 2017 05:00:17 -0000 (UTC)
Injection-Info: mx02.eternal-september.org; posting-host="aa8b9ac927b6be8ad54db2c77eabed3e"; logging-data="20359"; mail-complaints-to="abuse@eternal-september.org";	posting-account="U2FsdGVkX1/FWR1Da1UD9rF1j1pZ1Yx+"
User-Agent: Mozilla/5.0 (Windows NT 6.2; WOW64; rv:45.0) Gecko/20100101 Thunderbird/45.7.1
In-Reply-To: <6177f3b6-c5f1-4909-871a-2f4c533413d6@googlegroups.com>
Cancel-Lock: sha1:TTt8W0UAqTwzVwdqGmhfD2xBT1U=
Xref: csiph.com comp.lang.forth:54829

On 2/12/2017 4:52 PM, Julian Fondren wrote:
> On Sunday, February 12, 2017 at 3:42:04 PM UTC-6, Rod Pemberton wrote:
>>
>> Apparently, you have objections to the following?
>>
>> : ROVER >R OVER R> SWAP ;
>>
>
> VFX Forth doesn't do as well with return-stack ops.
>
>   : third >r over r> swap ;  ok
>   see third
>   THIRD
>   ( 080C0A30    53 )                    PUSH      EBX
>   ( 080C0A31    5B )                    POP       EBX
>   ( 080C0A32    8D6DFC )                LEA       EBP, [EBP+-04]
>   ( 080C0A35    895D00 )                MOV       [EBP], EBX
>   ( 080C0A38    8B5D08 )                MOV       EBX, [EBP+08]
>   ( 080C0A3B    C3 )                    NEXT,
>   ( 12 bytes, 6 instructions )
>    ok
>   : third 2 pick ;
>   THIRD is redefined  ok
>   see third
>   THIRD
>   ( 080C0A60    8D6DFC )                LEA       EBP, [EBP+-04]
>   ( 080C0A63    895D00 )                MOV       [EBP], EBX
>   ( 080C0A66    8B5D08 )                MOV       EBX, [EBP+08]
>   ( 080C0A69    C3 )                    NEXT,
>   ( 10 bytes, 4 instructions )
>    ok
>   : third dup 2over drop nip ;
>   THIRD is redefined  ok
>   see third
>   THIRD
>   ( 080C0A90    8D6DFC )                LEA       EBP, [EBP+-04]
>   ( 080C0A93    895D00 )                MOV       [EBP], EBX
>   ( 080C0A96    8B5D08 )                MOV       EBX, [EBP+08]
>   ( 080C0A99    C3 )                    NEXT,
>   ( 10 bytes, 4 instructions )
>    ok

The "third" example of third is a bit much isn't it?  What about just

: third 2over nip ;

But I guess that doesn't know if there is a fourth item on the stack and 
could cause an error?  It wouldn't on the GA144, lol.

Odd that the third third is compiled efficiently.  Odder is that VFX 
adds pointless instructions to the first third.  I'd say that rises to 
the level of being  a bug.

-- 

Rick C