Possible bug in ToLittleEndian32 in CVI 2009

[jtocwru]

Hello,

I ran the following sample code:

//////////////////////////

OutboundMsg.Header.Msg_ID = ToBigEndian32(FOO_MSG_ID);

DebugPrintf(“FOO_MSG_ID:        %d”,   FOO_MSG_ID);

DebugPrintf(“NATIVE:            %08X”, FOO_MSG_ID);

DebugPrintf(“BIG ENDIAN:        %08X”, OutboundMsg.Header.Msg_ID);

DebugPrintf(“LITTLE ENDIAN:     %08X”, ToLittleEndian32(OutboundMsg.Header.Msg_ID));

DebugPrintf(“LITTLE ENDIAN???   %08X”, ToBigEndian32(OutboundMsg.Header.Msg_ID));

///////////////////////////////

And here is the output:

FOO_MSG_ID:           30

NATIVE:               0000001E

BIG ENDIAN:           1E000000

LITTLE ENDIAN:        1E000000

LITTLE ENDIAN???      0000001E

The line in red shows that the call to ToLittleEndian32 did no byte swapping.  Instead, only ToBigEndian32 actually did any byte swapping.  CVI says that a failed call to these endian functions will return 0.  It should not return exactly what was passed into it, as I’m seeing here. 

Or am I missing something?

Thanks!

Tim O'Leary

[RobertoBozzolo · Accepted solution]

Functions dealing for endianness are located in toolbox instrument, whose code is distributed so you can check them by yourself.

As you can see here:

unsigned int CVIFUNC ToLittleEndian32(unsigned int n)
{
    if (HostIsLittleEndian())
        return n;
    return ToOtherEndian32(n);
}

 the function tests for system endianness and returns original number if the system is already little endian, which appears to be your case. This appears a correct behaviour to me.

The only condition for returning 0 is that ToOtherEndian32 () cannot discriminate system endianness; in every other case a number is returned.