| Re: How to get numbers into passed structure elements (pass-by-value/reference problem). [message #58951] |
Wed, 27 February 2008 07:56  |
![Paul Van Delst[1] is currently offline](theme/default/images/xoffline.png.pagespeed.ic.XRkd1fkXye.png "Paul Van Delst[1] is currently offline")
Paul Van Delst[1]
Messages: 1157
Registered: April 2002
|
Senior Member |
|
|
David Fanning wrote:
> Paul van Delst writes:
>
>> So, I realise this is one of those pass-by-reference or pass-by-value things, but how does
>> one get around it? Do I:
>> a) make the Type component of the Cloud structure a pointer? (Yuk!)
>> b) change the way I pass the Cloud structure into the CRTM_Read_Cloud_Record() fn?
>> i.e. not reference the cloud structure array via the index [n].
>>
>> I would much prefer (b), but will that entail copying entire structures? The number of
>> "Clouds" associated with any particular "Atm[m]" profile is variable.
>
> Too much misdirection for me to follow this early in the morning,
> but if you pass a structure array in by reference, what are you
> copying?
Well I meant if I do this:
FOR n = 0, Atm.n_Clouds-1 DO BEGIN
; Need to store data into (*Atm.Cloud)[n]
result = CRTM_Read_Cloud_Record( FileID, Cloud, DEBUG=Debug )
...Copy Cloud into (*Atm.Cloud)[n]
ENDFOR
> I think you are only "copying" if you pass something in
> by value. I would go with (b).
I just completed an experiment where I passed the entire pointer array and looped over
clouds in my reader,
result = CRTM_Read_Cloud_Record( FileID, *Atm.Cloud, DEBUG=Debug )
since it was passing (*Atm.Cloud)[n] that was buggering me up.
I find it a fundamental flaw of IDL that the argument passing mechanism (reference or
value) is so exposed to the user that they have to tailor their code to prevent data loss.
I understand why it's there, but even Fortran has moved beyond it (for about 18(!) years now).
cheers,
paulv
|
|
|
|
|
|
| Re: How to get numbers into passed structure elements (pass-by-value/reference problem). [message #59021 is a reply to message #58951] |
Thu, 28 February 2008 02:23  |
lecacheux.alain
Messages: 325
Registered: January 2008
|
Senior Member |
|
|
On 27 fév, 16:56, Paul van Delst <Paul.vanDe...@noaa.gov> wrote:
> David Fanning wrote:
>> Paul van Delst writes:
>
>>> So, I realise this is one of those pass-by-reference or pass-by-value things, but how does
>>> one get around it? Do I:
>>> a) make the Type component of the Cloud structure a pointer? (Yuk!)
>>> b) change the way I pass the Cloud structure into the CRTM_Read_Cloud_Record() fn?
>>> i.e. not reference the cloud structure array via the index [n].
>
>>> I would much prefer (b), but will that entail copying entire structures? The number of
>>> "Clouds" associated with any particular "Atm[m]" profile is variable.
>
>> Too much misdirection for me to follow this early in the morning,
>> but if you pass a structure array in by reference, what are you
>> copying?
>
> Well I meant if I do this:
>
> FOR n = 0, Atm.n_Clouds-1 DO BEGIN
> ; Need to store data into (*Atm.Cloud)[n]
> result = CRTM_Read_Cloud_Record( FileID, Cloud, DEBUG=Debug )
> ...Copy Cloud into (*Atm.Cloud)[n]
> ENDFOR
>
>> I think you are only "copying" if you pass something in
>> by value. I would go with (b).
>
> I just completed an experiment where I passed the entire pointer array and looped over
> clouds in my reader,
>
> result = CRTM_Read_Cloud_Record( FileID, *Atm.Cloud, DEBUG=Debug )
>
> since it was passing (*Atm.Cloud)[n] that was buggering me up.
>
> I find it a fundamental flaw of IDL that the argument passing mechanism (reference or
> value) is so exposed to the user that they have to tailor their code to prevent data loss.
> I understand why it's there, but even Fortran has moved beyond it (for about 18(!) years now).
>
> cheers,
>
> paulv
Since IDL pointers point to variables which are global in scope and
IDL structures are
global too, I would suggest, in any of your function, to just pass a
pointer on the used instance
of CRTM_Atmosphere...
alx
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
