bug of newLISP v.10.6.0

[Jian Leng]

Hi NewLisper,



I found the predicate number? has a bug.

when input number? 2323abc, it will return

true. So I think its a bug.



David

[rickyboy]

It's not a bug.  2323 is a number.

[rickyboy]

:)



The TLDR is that newlisp parses 2323abc as two atoms, namely the integer 2323 and the symbol abc.  Finally, the primitive number? only reads its first argument (and ignores any additional arguments).  So saying

Code Select Expand

(symbol? 2323abc)
is the same as saying

Code Select Expand

(symbol? 2323 abc)

[Jian Leng]

I see. Thanks a lot.



David

[xytroxon]

That is not always a good "feature"...

Code Select Expand


> (setq abc 123OX 456)
456
> abc
123   ; abc  should be 1230
> X
nil   ; X should be 456

In the dyslexic process of anticipating to shift to type the "X" key, I hit the "O" (oh) key  instead of the "0"  (zero) key. (aka "toushie" typing ;o) Parser missed an error it might/should have caught by enforcing whitespace separation at the end of a numerical value.

Code Select Expand


> OX
> 456 ; There it is!!!


-- xytroxon

[Lutz]

The way newLISP parses numbers is practical when parsing normal text. Numbers are often followed by unit identifiers, e.g.:

Code Select Expand


> (parse "length 100cm")
("length" "100" "cm")
> (parse "temperature is 100°")
("temperature" "is" "100" "°")
> (parse "weight is 10lbs")
("weight" "is" "10" "lbs")
> (parse "100$")
("100" "$")
> (parse "100¥")
("100" "¥")
> (parse "100€")
("100" "€")
> (parse "100e3€")
("100e3" "€")
> 100e3
100000


[xytroxon]

Please add a few of those examples to the parse section of the newLISP manual.



We humans are not well-versed in these advanced logic designs of the Krell*...



-- xytroxon



*http://en.wikipedia.org/wiki/Krell">//http://en.wikipedia.org/wiki/Krell