Well, this 'natural' behavior is a result of how binary addition works. Say you have an 8-bit binary integer 11111110 (254)
and you add another 8-bit binary integer to it 00000011 (3) you would get 100000001 (257) Since the integer can only hold 8 bits, that first bit is truncated and you are left with 00000001 (1) This is referred to as integer overflow. While it seems natural, the behavior for overflow in integers is undefined in the C standards. This means you shouldn't rely on that behavior if you want your code to remain portable. --Palrich. On Wed, 2004-11-17 at 10:57, Richard Lindsey wrote: > I know this is going to sound like a stupid question, but I'm just > seeking confirmation of this for my own peace of mind... the "natural" > behavior of variables in a programming environment (and since it may > differ from language to language, I'll specifically name C++ for this; > and by "natural" behavior, I mean how the variable would behave without > any special handlers written for bounds checking or anything, just > exactly how the code would be interpreted under normal circumstances) > is, as far as I've ever observed, that if the value exceeds the upper or > lower limit of the value range for the declared type, that wraps around > to the other end of the spectrum... i.e. in a short int variable that > can hold a range of 65536 values (unsigned), if I were to declare it as > 60536 and then add 10000 to it, it's current value should equal 5000, > correct? Or in a signed variable of the same size, ranging from -32767 > to 32767, if I declared it as being equal to 30767 and added 4000 to it, > it would then equal -30767, right? And if they were initialized to say, > -30767 and I subtracted 4000 then it should equal 30767?... I can't find > a definition on the natural behavior of integers anywhere online, and > since I've never taken an actual C or C++ course, all of my knowledge of > it is based on observation, so can someone more learned than I confirm > that? Or if it's incorrect, tell me why, or if it's quirky, point out > certain instances that this wouldn't be the case? Thanks in advance :) > > Richard Lindsey
