Is there a bug in arrresize.c and in the documentation?

[kolan]

Hi all!

Here a some code from  C:\Program Files\National Instruments\LabVIEW 2009\CCodeGen\libsrc\lvanalysis\arrresize.c

NumericArrayResize function:

if (! *ap) return mgArgErr;                                  
p = **ap;                                                    
arr = (int32 *)p;                                            
                                                             
if (n == 0 && *ap == NULL) {                                 
    *ap = (UHandle)DSNewHClr( (size_t)(nd * sizeof(int32)) );
    if (! *ap) {                                            
        SetError(kArresizeModNum, kMemFail);                
        return mFullErr;                                    
    }                                                       
    return noErr;                                           
}                                                            

Does anybody see  an error?

And here NumericArrayResize (CIN Function):

If *dataHP is NULL, LabVIEW allocates a new array handle in *dataHP.

[kolan]

this is ok:

p = **ap;
arr =(int32 *)p;

compiler optimize it...

but here I'm weeping:

if ((dt == doubleDataType || dt == cmplx128DataType) && (nd % 2)) {
  newSize += 4; //allow for 8 byte alignment of doubles
}

[Wolfgang]

are you sure that int32 is a valid identifier in CVI?

[kolan]

Please read more closely!

[Wolfgang]

 it appears to me that your example code refers to LabView, not CVI

[kolan]

You are in part right! But all this functions developed to use in CVI. See the forum please.

[kolan]

If somebody bring me that LabWindows/CVI has NumericArrayResize implementation which is differ from LabView's one, I'll move the post with a Bug to LabView's branch immediately.

Message Edited by kolan on 02-27-2010 12:08 PM

[Lars_L]

Hey Kolan,

I did some searching on the CCodeGen directory as I don't have this or the file here on my machine so I could test the behavior. It looks like this directory is usually present, however, if you are using any of our LabVIEW Embedded products such as LV for ARM, LV for Blackfin, or Microprocessor SDK. If this is the case, then there is actually another forum where they should be able to help you further. Here is the link:

LabVIEW Embedded Forum

http://forums.ni.com/ni/board?board.id=420

Lars

[kolan]

Lars_L, Thanks! The topic continues here.