Re: [colorforth] Dare I say ANS!

["Samuel A. Falvo II" <kc5tja@xxxxxxxx>]

On Thursday 15 January 2004 05:22 am, Howerd Oakford wrote:
> Make i return an index from 0 to n-1 instead of 1 to n. I presume this
> was intentional - does any one know why?

Because implementing FOR/NEXT is substantially easier to implement than 
DO/LOOP.

> ?dup - what is this?

This is part of Chuck's peephole optimizer.  It detects if a DROP 
preceeded the word, and if so, "backspaces over the DROP," so that the 
literal can be loaded into EAX directly, instead of taking the time to 
first DROP, then DUP, then load EAX.

> I think we should use whatever the ANS form is for 1@, 2@ etc ( c@ w@
> I think ).

ANSI defines only @ and C@.  There is no equivalent to 1@ or 2@.  
Frankly, I prefer 1@ and 2@ over C@ -- the former indicates firmly how 
many bytes are being fetched, while the latter is rather ambiguous and 
platform dependent (e.g., are we running the Forth on a Unicode system?  
If so, a character may be 16-bits wide.  Or, even 32-bits.  Otherwise, 
it's generally accepted to be only 8-bits.  Moreover, IIRC, ANSI Forth 
has some ambiguities over whether C@ is to fetch a *character* versus a 
*byte*.  E.g., the word COUNT, for example, is defined for strings that 
do not exceed 255 characters; it's behavior isn't well defined for 
strings exceeding 255 characters.  This implies that C@ fetches a byte.  
Yet, ANSI insists that C@ may fetch characters wider than 8-bits.  Had 
enough yet?)

> Does anyone know if "port fetch" has an ANS name? polyForth uses INPUT
> and OUTPUT, p@ and p! is fine by me.

No.  Again, port-fetch is extremely x86 specific.

--
Samuel A. Falvo II


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