| Bug in N_PARAMS? [message #64011] |
Mon, 24 November 2008 11:31  |
Kenneth P. Bowman
Messages: 585
Registered: May 2000
|
Senior Member |
|
|
I have been having some very odd problems with the N_PARAMS function lately.
Here is a simple example with a main program and a function.
PRO N_PARAMS_MAIN, a, b
COMPILE_OPT IDL2
PRINT, 'N_PARAMS_MAIN - Number of parameters : ', N_PARAMS()
HELP, a, b
value = N_PARAMS_BUG(a, b)
END
FUNCTION N_PARAMS_BUG, a, b
COMPILE_OPT IDL2
PRINT, 'N_PARAMS_BUG - Number of parameters : ', N_PARAMS()
HELP, a, b
RETURN, 0
END
Here is the output from a new IDL session:
stretch> idl
IDL Version 6.4.1, Mac OS X (darwin i386 m32). (c) 2007, ITT Visual Information Solutions
IDL> N_PARAMS_MAIN, 13
% Compiled module: N_PARAMS_MAIN.
N_PARAMS_MAIN - Number of parameters : 1
A LONG = 13
B UNDEFINED = <Undefined>
% Compiled module: N_PARAMS_BUG.
N_PARAMS_BUG - Number of parameters : 2 <-- returns 2, but b is undefined
A LONG = 13
B UNDEFINED = <Undefined>
In the main program N_PARAMS_MAIN, N_PARAMS correctly returns the number of defined arguments (one).
In the function N_PARAMS_BUG, the argument count is incorrectly given as two.
If I call the function N_PARAMS_BUG from the command line, it works correctly.
IDL> PRINT, N_PARAMS_BUG(10)
N_PARAMS_BUG - Number of parameters : 1
A LONG = 10
B UNDEFINED = <Undefined>
0
Am I missing something obvious here?
Ken
|
|
|
|
|
|
| Re: Bug in N_PARAMS? [message #64041 is a reply to message #64011] |
Mon, 01 December 2008 14:55  |
Mark[1]
Messages: 66
Registered: February 2008
|
Member |
|
|
On Dec 2, 3:11 am, Jeremy Bailin <astroco...@gmail.com> wrote:
> Except that there are valid reasons for passing in
> an undefined variable, if it's used for output. So
> it's perfectly legitimate to want to know whether
> it was passed in without caring whether it is
> defined.
To check that an argument has been passed, and can be written to, you
can use ARG_PRESENT.
To check that a argument has been passed, and is a defined variable,
you can use N_ELEMENTS > 0.
This leaves N_PARAMS for the cases where you want to know if an
argument has been passed, but otherwise don't care about its status.
In my experience. this is pretty rare, except in what I would call
wrapper routines.
|
|
|
|
| Re: Bug in N_PARAMS? [message #64048 is a reply to message #64011] |
Mon, 01 December 2008 06:11 |
Jeremy Bailin
Messages: 618
Registered: April 2008
|
Senior Member |
|
|
On Nov 30, 5:06 pm, Mark <mark.h...@gmail.com> wrote:
> On Nov 28, 11:02 pm, Reimar Bauer <R.Ba...@fz-juelich.de> wrote:
>
>> The usecase for me is to call a helper routine if not enough parameters
>> were used.
>
>> e.g.
>
>> if n_params() lt 2 then begin
>> MESSAGE,call_help(),/cont
>> return
>> endif
>
> Sure, but even if 2 parameters were supplied, you still don't know if
> the variables they're associated with are defined. If you're going to
> check for that anyway, with N_ELEMENTS, the N_PARAMS check is
> redundant.
>
> Eg.
>
> pro mypro, a, b
>
> if n_elements(a) eq 0 || n_elements(b) eq 0 then begin
> MESSAGE,call_help(),/cont
> return
> endif
>
> end
Except that there are valid reasons for passing in an undefined
variable, if it's used for output. So it's perfectly legitimate to
want to know whether it was passed in without caring whether it is
defined.
-Jeremy.
|
|
|
|
| Re: Bug in N_PARAMS? [message #64053 is a reply to message #64011] |
Sun, 30 November 2008 14:06 |
Mark[1]
Messages: 66
Registered: February 2008
|
Member |
|
|
On Nov 28, 11:02 pm, Reimar Bauer <R.Ba...@fz-juelich.de> wrote:
> The usecase for me is to call a helper routine if not enough parameters
> were used.
>
> e.g.
>
> if n_params() lt 2 then begin
> MESSAGE,call_help(),/cont
> return
> endif
>
>
Sure, but even if 2 parameters were supplied, you still don't know if
the variables they're associated with are defined. If you're going to
check for that anyway, with N_ELEMENTS, the N_PARAMS check is
redundant.
Eg.
pro mypro, a, b
if n_elements(a) eq 0 || n_elements(b) eq 0 then begin
MESSAGE,call_help(),/cont
return
endif
end
|
|
|
|
| Re: Bug in N_PARAMS? [message #64066 is a reply to message #64011] |
Fri, 28 November 2008 02:02 |
R.Bauer
Messages: 1424
Registered: November 1998
|
Senior Member |
|
|
The usecase for me is to call a helper routine if not enough parameters
were used.
e.g.
if n_params() lt 2 then begin
MESSAGE,call_help(),/cont
return
endif
The user get's shown the CALLING SEQUENCE description
cheers
Reimar
Mark schrieb:
> On Nov 26, 3:29 am, "Kenneth P. Bowman" <k-bow...@null.edu> wrote:
>> Ah. That significantly restricts the utility of
>>
>> SWITCH N_PARAMS() OF
>> 0 : ...
>>
>> ENDSWITCH
>>
>> to provide default values for optional parameters.
>
> Indeed. N_ELEMENTS is what you need for that. The only use I have
> found for N_PARAMS is for writing wrapper routines
>
> pro wrapper, P1, P2, P3, P4, _REF_EXTRA=_extra
> ;; Do some other stuff
> case n_params() of
> 0: wrappee, _STRICT_EXTRA=_extra
> 1: wrappee, P1, _STRICT_EXTRA=_extra
> 2: wrappee, P1, P2, _STRICT_EXTRA=_extra
> 3: wrappee, P1, P2, P3, _STRICT_EXTRA=_extra
> 4: wrappee, P1, P2, P3, P4, _STRICT_EXTRA=_extra
> endcase
> end
>
> Ugly, but sometimes necessary.
|
|
|
|
| Re: Bug in N_PARAMS? [message #64139 is a reply to message #64011] |
Wed, 03 December 2008 11:15 |
pgrigis
Messages: 436
Registered: September 2007
|
Senior Member |
|
|
David, you don't really need all these keyboards,
it's enough to play a bit with the keyboard settings
and select other languages (this should be possible
on our triad of IDL OSes).
öäü
दाविद ्ाननिनग
だゔぃづ ふぁんんいぬぐ
δαωιδ φαννινγ
גשהןג כשממןמע
All written with an american keyboard!
Of course the hard part sometimes may be
finding out the meaning of what you just wrote ;-)
Ciao,
Paolo
David Fanning wrote:
> Reimar Bauer writes:
>
>> you need a german keyboard or is this also an english word
>> "doppelganger" <-> doppelg�nger?
>
> I need a German keyboard AND this is an English
> word. Well, it is in an English dictionary, but
> obviously stolen from the Germans. :-)
>
> Cheers,
>
> David
>
> --
> David Fanning, Ph.D.
> Coyote's Guide to IDL Programming (www.dfanning.com)
> Sepore ma de ni thui. ("Perhaps thou speakest truth.")
|
|
|
|
| Re: Bug in N_PARAMS? [message #64140 is a reply to message #64011] |
Wed, 03 December 2008 10:49 |
David Fanning
Messages: 11724
Registered: August 2001
|
Senior Member |
|
|
Reimar Bauer writes:
> you need a german keyboard or is this also an english word
> "doppelganger" <-> doppelg�nger?
I need a German keyboard AND this is an English
word. Well, it is in an English dictionary, but
obviously stolen from the Germans. :-)
Cheers,
David
--
David Fanning, Ph.D.
Coyote's Guide to IDL Programming (www.dfanning.com)
Sepore ma de ni thui. ("Perhaps thou speakest truth.")
|
|
|
|
| Re: Bug in N_PARAMS? [message #64141 is a reply to message #64011] |
Wed, 03 December 2008 09:59 |
R.Bauer
Messages: 1424
Registered: November 1998
|
Senior Member |
|
|
David Fanning schrieb:
> David Fanning writes:
>
>> Carsten Lechte writes:
>>
>>> Kenneth P. Bowman wrote:
>>>> to provide default values for optional parameters.
>>> I wrote this little procedure for setting default values
>>> that I use a zillion times a day:
>>>
>>>
>>> PRO defaultarg, arg, defval, VALUEGIVEN=valuegiven
>>>
>>> IF N_ELEMENTS( arg) EQ 0 THEN BEGIN
>>> arg = defval
>>> valuegiven = 0
>>> ENDIF ELSE valuegiven = 1
>>>
>>> END
>> This is a pretty good idea. I can see how it would
>> save some typing! :-)
>>
>> Here is a slightly modified version. I'll add this,
>> with documentation, to the Coyote Library later today.
>
> I have a doppelganger at work who takes all my good ideas
hehe
you need a german keyboard or is this also an english word
"doppelganger" <-> doppelg�nger?
cheers
Reimar
> as soon as I publish them and points out all their deficiencies
> and the obvious ways in which I could have made them better. :-(
>
> To that end, I've updated the SetDefaultValue program I published
> earlier to something a good deal simpler that also works in a more
> intuitive way. You can find the update, if you are interested in
> this sort of thing, here:
>
> http://www.dfanning.com/programs/setdefaultvalue.pro
>
> Cheers,
>
> David
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
