On Mon, Apr 28, 2008 at 12:07:24AM +0100, Donal K. Fellows wrote:
>  Larry McVoy wrote:
>  >If you don't return but one char at a time, isn't it cheaper?
>  
>  Maybe. I'd be interested in seeing real measurements. (I only know what
>  I did to [split] was a win because I measured it.)

I'm a big fan of measurements.  See http://bitmover.com/lm/langbench.shar :)

>  >Shouldn't that just maintain a reference on buf throughout the loop (so
>  >noone can change it) and internally, in the state for the foreach,
>  >maintain an index into buf?  Module unicode char widths, etc.
>  >
>  >It seems to me that is mucho cheaper than internally splitting it into
>  >a bunch of char objects.  Just use one.  Would that not work?
>  
>  Theoretically yes, but you're spreading a lot of complexity to handle
>  these indexed (and reference-counted) strings properly. 

Can you explain that a bit?  I'm thinking something like

    when we detect a foreach (c in buf) 
	bump the refcount on buf
	set an index at 0
	start the loop
	each iteration
		c = buf[index++]
		loop body
		decr refcount on c
        decr refcount on buf

What's so hard about that?  I realize I'm asking this naively so I'm not
say "you wuss, go do it", I'm say "teach me why this is hard, please".
-- 
---
Larry McVoy                lm at bitmover.com           http://www.bitkeeper.com

-------------------------------------------------------------------------
This SF.net email is sponsored by the 2008 JavaOne(SM) Conference 
Don't miss this year's exciting event. There's still time to save $100. 
Use priority code J8TL2D2. 
http://ad.doubleclick.net/clk;198757673;13503038;p?http://java.sun.com/javaone
_______________________________________________
Tcl-Core mailing list
Tcl-Core@[...].net
https://lists.sourceforge.net/lists/listinfo/tcl-core