The following are the details of the data structure the function reference as defined in the API documentation. I want to know how to pass the details to the dll and get the information back for this structure.
typedef struct _BLUETOOTH_DEVICE_INFO_EX {
DWORD dwSize;
BYTE address[DEVICE_ADDRESS_LENGTH];
BYTE classOfDevice[DEVICE_CLASS_LENGTH];
CHAR szName[MAX_DEVICE_NAME_LENGTH];
BOOL bPaired;
UCHAR ucLmpVersion;
WORD wManuName;
WORD wLmpSubversion;
BYTE reserved[16];
WORD wClockOffset;
BOOL bConnected;
DWORD dwDataRecvBytes;
DWORD dwDataSentBytes;
CHAR cSignalStrength;
} BLUETOOTH_DEVICE_INFO_EX, *PBLUETOOTH_DEVICE_INFO_EX;
DWORD BT_GetLocalDeviceInfo (
/* [in] */ DWORD dwMask,
/* [out] */ PBLUETOOTH_DEVICE_INFO_EX pDevInfo
);
There are a few approaches to take here, with various levels of tedium / non-LabVIEW effort. It's unfortunate that LV's built-in Bluetooth functions don't offer this level of detail - as far as I know.
Approach I:
Write a wrapper DLL of your own. The roles of the DLL are to act as a data structure translator and to cope with situations LV can't support directly from the diagram (e.g. passing callback functions to functions, etc.). The main drawback, of course, is dealing in C code with data types like LStrHandle and whatnot.
Approach II:
Rely on a C struct being the same as a LV cluster. In this particular case, you have to deal with fixed-sized blocks within the C structure, such as BYTE address[DEVICE_ADDRESS_LENGTH]. The way to do this is to treat each fixed-sized array as a cluster as well. So the 'address' element of your LV cluster would itself be a cluster containing DEVICE_ADDRESS_LENGTH uInt8-sized numeric values. You'll also have to be sure to set the dwSize parameter to the size IN BYTES of the entire cluster's data. I believe in Win32 the BOOL type is a 32-bit integer sized quantity - the rest should be apparent. Provided you accurately compute and set the dwSize element, this approach should work. In this case, you're taking advantage of the fact that a struct containing N of the same elements of type T looks the same in memory as a fixed-sized array of N elements of the data type T.
Approach III:
Make an array of bytes large enough for the structure and pass it as a simple data pointer. You'll have to fill in the first four bytes correctly with the size of the C structure, just as in Approach II, but you could make the array larger than necessary. After the call, extract the separate structure members by hand, possibly needing to fix up the byte ordering (it's little endian in Windows).Though this sounds really onerous, if you only need one or two things out of the structure, it may be less work to implement than the other two suggestions.
intvsteve
LabVIEW R&D
