C++: set value to floating point variable by bits

Question

I have a large message which must be initialized to default values. The message is a structure with many sub-structures, enums, etc. I have an automatically generated description file of the message, which contains the field-by-field parse of the message: field start byte, start bit, length, type, default value. I've implemented a mechanism which treats the message as a bits buffer and initializes it according to the description file. But.... I have a problem with floating point defaults. How do I make the automation of setting floating point variable bitwise?

Thanks!

Answer 1

Well, I've managed to solve it by myself. Assume the following:

MSG_STRUCT_TYPE OutputMsg; // e.g.750 bytes
int startByte = 28; // the start byte of floating point field in the message
float defVal = -1.75f; // the default value for the field above

Now, using the reinterpret_cast we can do the following manipulation:

unsigned char* bytes_to_set = reinterpret_cast<unsigned char*>(&defVal);
unsigned char* msgBuff = reinterpret_cast<unsigned char*>(&OutputMsg);
*reinterpret_cast<float*>(&msgBuf[startByte]) = *reinterpret_cast<float*>(bytes_to_set);

That do the job.

---

After implementing and checking njuffa 's comment bellow, I'm adding it as another (simplier and cleaner) answer:

unsigned char* msgBuff = reinterpret_cast<unsigned char*>(&OutputMsg);
memcpy(&msgBuf[startByte], &defVal, sizeof(float);