NumericArrayResize

[nasa2grlv]

Does anybody else have issues with NumericArrayResize?

I am running windows 7 x64, 32 bit labview, and creating a 32 bit dll.  There is a bug on the packing of the structure for doubles. I had to add a #pragma pack(1) to the code for my configuration.

I modified an example off the NI website/help as follows.  The example originally uses uint8, change the code for doubles and it did not work.  I think the "extcode.h" has improper #if/#defines for checking system configuration.

#include "extcode.h"

#pragma pack(1)

typedef struct {

long dimSize;

double elt[1];

} TD2;

typedef TD2 **TD2Hdl;

_declspec (dllexport) long sizeToTen(TD2Hdl array);

_declspec (dllexport) long sizeToTen(TD2Hdl array)

{

  double * temp;

    MgErr error;

    error = NumericArrayResize(fD, 1, (UHandle *)(&array), 10*sizeof(double));

   (*array)->dimSize = 10;

   temp = (*array)->elt;

   temp[3] = (double)1;

   temp[7] = (double)2;

   temp[9] = 255;

   return error;

}

[RavensFan]

This thread is for new Ideas about improving LabVIEW.  Your message certainly isn't that.  You say you are using LabVIEW, but yet you are talking about some C code.

Perhaps you want the LabWindows forum?

[nasa2grlv]

It is for calling external c-code in LABVIEW using LABVIEW's Call Library Function Node LABVIEW Block. Fixing LABVIEW is improving LABVIEW.

[Mr.Mike]

Could you clarify what eactly you think is wrong and why you think the packing is wrong? 

[nasa2grlv]

First off I am using VS2010 and labview 2010 32bit with a 32 bit dll to be called by Labview on 64bit Windows 7.

My release project defines are from the project properties:

WIN32;NDEBUG;_WINDOWS;_USRDLL;AES_UNWRAP_EXPORTS;%(PreprocessorDefinitions)

The results of the above code without the #pragma pack(1) as displayed on a front panel of a VI is

0, 0, 0, 0, 5.29981E-315, 0, 0, 0, 5.30499E-315, 0

[Chris_W1]

So, let me see if I understand:  you're saying that LabVIEW is wrong to not include the #pragma pack, as you have done manually, in the example?  Yes, I agree with you.

If I remember correctly, LabVIEW introduced the lv_prolog.h and lv_epilog.h headers in LV 2010 for this reason.  An easy way to see these headers in use is to right-click on a Call Library Node and choose "Create C File".

Here's the header I get from doing that on my machine:

/* Call Library source file */

#include "extcode.h"

/* lv_prolog.h and lv_epilog.h set up the correct alignment for LabVIEW data. */
#include "lv_prolog.h"

/* Typedefs */

typedef struct {
   int32_t dimSize;
   double elt[1];
   } TD1;
typedef TD1 **TD1Hdl;

#include "lv_epilog.h"

int32_t sizeToTen(TD1Hdl array);

int32_t sizeToTen(TD1Hdl array)
{

 /* Insert code here */

}

[nasa2grlv]

Thanks for your help, I will try constructing my cfile that way.  We'll it is their function, if you tell a function to do something it should work without any headaches.

[G-Money]

Programming bug issue. There is a workaround listed in the comments though.