| Re: negation operator [message #49296] |
Fri, 14 July 2006 08:02 |
greg michael
Messages: 163
Registered: January 2006
|
Senior Member |
|
|
Yes, you're right - I'd got used to the quirk with exponents. I've only
just realised, though, that it appears to mean that there's no such
thing as a negative constant in IDL. They're always calculated values.
Maybe that's normal - I don't know.
greg
JD Smith wrote:
> On Thu, 13 Jul 2006 06:31:18 -0700, greg michael wrote:
>
>> This is odd...
>>
>> IDL> print,-100 > -90 < 90
>>
>> print,-100 > -90 < 90
>> ^
>> Syntax error.
>> IDL> print,-100 > (-90) < 90
>> -90
>>
>> Is there any reason why a negation operator should ever come lower than
>> anything else?
>
> You get the same problem with exponentiation, which is 2 steps up in
> the operator precedence chain (see
> http://www.dfanning.com/misc_tips/precedence.html).
>
> IDL> print, 2.^-2
>
> print, 2.^-2
> ^
> % Syntax error.
> IDL> print, 2.^(-2)
> 0.250000
>
> So it's not that fact that `-' shares precedence with '>', just that
> both are left-right associative, so you can't put two adjacent. This
> doesn't always hold though... for example, how about `++' (2nd highest
> precedence) and '~' (2nd to last precedence):
>
> IDL> a=1
> IDL> print,~++a
> 0
>
> but how about:
>
> IDL> print,~-a
>
> print,~-a
> ^
> % Syntax error.
>
> This *should* work. Negation is priority 5, logical negation is
> priority 8, so it should evaluate -a first, and then ~(-a). Not sure
> if this should be considered a bug or just a weakness of the
> interpreter.
>
> JD
|
|
|
|
| Re: negation operator [message #49320 is a reply to message #49296] |
Thu, 13 July 2006 09:35  |
Foldy Lajos
Messages: 268
Registered: October 2001
|
Senior Member |
|
|
Hi,
On Thu, 13 Jul 2006, JD Smith wrote:
> IDL> a=1
> IDL> print,~++a
> 0
>
> but how about:
>
> IDL> print,~-a
>
> print,~-a
> ^
> % Syntax error.
>
> This *should* work. Negation is priority 5, logical negation is
> priority 8, so it should evaluate -a first, and then ~(-a). Not sure
> if this should be considered a bug or just a weakness of the
> interpreter.
>
well, this should work according to the documentation, but does not.
I think this is a bug in the documentation, logical negation has the
same precedence as bitwise negation and unary -. IDL/GDL/FL agree on this.
regards,
lajos
|
|
|
|
| Re: negation operator [message #49324 is a reply to message #49320] |
Thu, 13 July 2006 09:17 |
JD Smith
Messages: 850
Registered: December 1999
|
Senior Member |
|
|
On Thu, 13 Jul 2006 06:31:18 -0700, greg michael wrote:
> This is odd...
>
> IDL> print,-100 > -90 < 90
>
> print,-100 > -90 < 90
> ^
> Syntax error.
> IDL> print,-100 > (-90) < 90
> -90
>
> Is there any reason why a negation operator should ever come lower than
> anything else?
You get the same problem with exponentiation, which is 2 steps up in
the operator precedence chain (see
http://www.dfanning.com/misc_tips/precedence.html).
IDL> print, 2.^-2
print, 2.^-2
^
% Syntax error.
IDL> print, 2.^(-2)
0.250000
So it's not that fact that `-' shares precedence with '>', just that
both are left-right associative, so you can't put two adjacent. This
doesn't always hold though... for example, how about `++' (2nd highest
precedence) and '~' (2nd to last precedence):
IDL> a=1
IDL> print,~++a
0
but how about:
IDL> print,~-a
print,~-a
^
% Syntax error.
This *should* work. Negation is priority 5, logical negation is
priority 8, so it should evaluate -a first, and then ~(-a). Not sure
if this should be considered a bug or just a weakness of the
interpreter.
JD
|
|
|
|
| Re: negation operator [message #49328 is a reply to message #49324] |
Thu, 13 July 2006 07:15 |
Foldy Lajos
Messages: 268
Registered: October 2001
|
Senior Member |
|
|
Hi,
On Thu, 13 Jul 2006, greg michael wrote:
> This is odd...
>
> IDL> print,-100 > -90 < 90
>
> print,-100 > -90 < 90
> ^
> Syntax error.
> IDL> print,-100 > (-90) < 90
> -90
>
> Is there any reason why a negation operator should ever come lower than
> anything else?
>
> Greg
>
>
yes, operator precedence. <, > and unary - has the same precedence in IDL,
so they can not follow each other immediately.
regards,
lajos
ps: in this respect I have followed the C precedence rules in FL, where
unary + and - have higher precedence than * (multiplication). So in FL
the above statements are correct and give the same result.
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
