development release newLISP v.9.2.12

[Lutz]

Feature complete for 9.3 release on January 15th.



Files and changes notes: http://newlisp.org/downloads/development/">http://newlisp.org/downloads/development/



Lutz

[cormullion]

Thanks Lutz!



Yes, I'm now finding a few 'list index out of bounds' errors in my code - such as in my first effort at a tokenizer (http://unbalanced-parentheses.nfshost.com/syntax.cgi?downloads/tokenizer.lsp.txt">//http://unbalanced-parentheses.nfshost.com/syntax.cgi?downloads/tokenizer.lsp.txt) !



A virtual beer token to anyone who can find the out of bounds list index!



(I've always claimed to be a beginner, so it's good that newLISP 9.3 is picking up more of my beginner's errors than before...!)

[newdep]

Lutz,



Could "list index out of bounds" be changed into 'nil or an empty list?



This error message does not feel right in the newlisp context.



Lots of functions do return 'nil when nothing found..

so should indexes on lists. That makes programming more logical too.

...but thats my view...





Btw.. (10 '( 1 2 3)) still returns  '().





Norman.

[Cyril]

It seems that I found a bug: 'randomize' on 2-element list doesn't randomize it. List is always reversed, mo matter how much times do you call the function.

Code Select Expand

> (randomize '(1 2))
(2 1)
> (randomize '(1 2))
(2 1)
> (randomize '(1 2))
(2 1)


Another related bug in documentation: in description of 'rand' function it said "When 0 (zero) is passed, the internal random generator is initialized using the current value returned by the time function". But 'time' is of no use here! Probably 'time-of-day' or 'date-value'?

[Cyril]

Could "list index out of bounds" be changed into 'nil or an empty list?

This way you cannot distinguish between ('(a b c) 3) and ('(a b c nil) 3)

Btw.. (10 '( 1 2 3)) still returns  '().

Python does it exactly this way.

[newdep]

Its still far from logic...

[Lutz]

This is intended behavior, 'randomize' keeps trying until a different sequence is found. Pass the extra parameter 'true' to force true randomization.

Code Select Expand


> (randomize '(1 2) true)
(2 1)
> (randomize '(1 2) true)
(1 2)
> (randomize '(1 2) true)
(2 1)
> (randomize '(1 2) true)
(2 1)
> (randomize '(1 2) true)
(2 1)
> (randomize '(1 2) true)
(2 1)
> (randomize '(1 2) true)
(1 2)
>


Lutz



ps: this is also described in the manual: http://newlisp.org/downloads/newlisp_manual.html#randomize">http://newlisp.org/downloads/newlisp_ma ... #randomize">http://newlisp.org/downloads/newlisp_manual.html#randomize

[cormullion]

I've found my bug, so I'm going to drink the beer myself... :)

Code Select Expand

(for (p 0 (length result))

should have been

Code Select Expand

(for (p 0 (- (length result) 2))

I have no excuse, and it's an ugly fix anyway... :) Still, you're lucky I don't do this for a living!