Thread: Creating Random hex numbers

Originally Posted by disruptor108

7fff

Ok well that confirms your assumption about the value of RAND_MAX, and the print function is the problem. I would try some version of sprintf(). So instead of:

Code:

>_tcscpy(ProductID, _T("{%8X-%4X-%4X-%4X-%12X}", rand(), rand(), rand(), rand(), rand()));

Try:

Code:

_tsprintf(ProductID, _T("{%8X-%4X-%4X-%4X-%12X}", rand(), rand(), rand(), rand(), rand()));

And if that doesn't work or there is no such animal, then try sprintf():

Code:

_sprintf(ProductID, "{%8X-%4X-%4X-%4X-%12X}", rand(), rand(), rand(), rand(), rand()));

And if you're creating a GUID, use the WinAPI function CornedBee suggests.

Originally Posted by Ehtyar

Personally I'm very much not a fan of CoCreateGuid, instead, i initialize a variable to the size of _GUID and fill it with CryptGenRandom. Advertising my MAC address is not something I'm willing to do in the name of uniqueness.

Ehtyar.

Then don't call it GUID

Originally Posted by Ehtyar

Personally I'm very much not a fan of CoCreateGuid, instead, i initialize a variable to the size of _GUID and fill it with CryptGenRandom. Advertising my MAC address is not something I'm willing to do in the name of uniqueness.

Ehtyar.

I strongly suggest you rethink that strategy. Using a different algorithm to everyone else gives a possibility that the GUIDs you pick will clash with GUIDs created by others, and even potentially with your own.
The idea you speak of means that you can no longer call them GUIDs, as they are not GLOBALLY unique. They are in fact only LOCALLY unique, with very high probability.

Your escuse about not advertising your MAC address is absurdly paranoid.

I think <ole2.h>, but MSDN will tell you for sure. Just search for the function and look at the bottom of its page. It tells you what header file to include and what library to link against.