Ok this may have been asked before, so I appologize if it's been answered already.. I been working with FASM and MASM32 for about 3 months now, which is why I'm surprised the question for me is now just coming up, but it dawned on me... what in the world is the difference in declaring a variable of some type with ? or not using it? For example:

hwnd dd ? ; is pretty typical... but what about...
ps PAINTSTRUCT

Now I seem to remember certain things about the use of call (I think it was) that variables initialized on declaration get loaded into memory differently, and a call can jump backwards in memory but not forwards, and unintialized variables get loaded after the code and certain functions can't access them so you have to use another set of functions, but I think that might have been more of an MASM thing more so than anything... So is that really the only difference? If so is there any real main advantage or reason why I should choose one over the other?

-- StakFallT

? data should not be assigned to any value and should run the risk of being filled with garbage data(i believe)

but out of curiosity i went and tested, declaring data with ? just sets it to 0 from the tests i did, but it may run that garbage risk

That is true - ? usually gets defined as 0 but you can't count on it. Just to be sure always fill it yourself or if you NEED it to be ZERO, then:


ps PAINTSTRUCT is a structure that you must fill in yourself and you can't expect it to be 0 (although it might be )

Offtopic: I've always wondered how to tell which is best:
1) rd 1
2) dd ?

3) dd 1 dup ?
4) dd 1 dup (?)


When it happens that uninitialized data has to be anyway written into the file, it is filled with zeros - but you shouldn't rely even on this. If I changed it to put 90h or 0CCh bytes there, it still would be consistent with the presumptions.

aaah ok that makes sense, thanks for replying guys

im gonna have to go with madis here, i prefer the fasm syntax to the masm syntax (or tasm or whatever >_>)