defining a floating point value in data section from an equate

[Damos]

Can you help me please? i'm trying to do this:

Code Select Expand


ScreenResolutionX = 1024

.data
fScreenResolution      dd ScreenResolution.0


I need to generate a floating point value from a constant integer at assembly time.
Is this possible? A macro of some sort?

[jj2007]

No macro needed:

Code Select Expand

ScreenResolutionX = 1024
...
.data
fScreenResolution REAL4 @CatStr(%ScreenResolutionX, <.0>)

[Damos]

You're a star jj2007, thanks for the help. i'm working on a 2d tile game engine and i've just spent the last hour tring to do this, think I need to review some of my masm & macro knowledge. Thanks again.

[Neo]

Does it not work with just:

Code Select Expand

ScreenResolutionX = 1024
...
.data
fScreenResolution REAL4 ScreenResolutionX

Or does that give some sort of unnecessary error?

[jj2007]

Does it not work with just:

Code Select Expand

ScreenResolutionX = 1024
...
.data
fScreenResolution REAL4 ScreenResolutionX

Or does that give some sort of unnecessary error?

Try yourself:

Code Select Expand

include \masm32\include\masm32rt.inc

ScreenResolutionX = 1024

.code
fScreenResolution REAL4 ScreenResolutionX

start:
exit
end start


error A2187

[Neo]

error A2187

Darn, that sucks.  :(

It won't give an error in Inventor IDE, but I can't recall if it encodes it as the integer 1024 or the floating point 1024 (I should fix it if it's not floating point).  It'll also let you do things like doing operations on a float in arbitrary constant expressions, like:

Code Select Expand

mov  eax,1.0/(1 shl 32) ;moves the floating point 1/(2^32) into eax

I think it'll use it as an integer on operations other than + - * / though, so like:

Code Select Expand

mov  eax,1.0 shr 23 ;moves the exponent of 1.0 into eax (7Fh)

Edit: It'll treat it as a double instead of a float in some cases, so 1.0 shr 23 may not be 7Fh; I should check back over that code, since it's been a while.  :wink

[dedndave]

it doesn't allow "ScreenResolution*1.0"
it would be nice if the assembler could just do math - lol

[jj2007]

error A2187

Darn, that sucks.  :(

It's called type checking...

[Neo]

it doesn't allow "ScreenResolution*1.0"
it would be nice if the assembler could just do math - lol

That one's definitely supported in Inventor IDE.  In fact, you could define a constant ScreenResolutionFloat as "ScreenResolution*1.0", and when you hold Ctrl and run the mouse over the name, it'll show you "= 1024.0"  :8)

[jj2007]

it doesn't allow "ScreenResolution*1.0"
it would be nice if the assembler could just do math - lol

It does, it does...

Code Select Expand

MyCt = 5
MyDword dd MyCt*5
MyFloat REAL4 @CatStr(%MyCt*5, <.0>)


[dedndave]

yesssss - well i would hope it could multiply a real by an integer without catstr
could be useful for scaling
odd that i have never needed to do this
i guess it's because, once the assembler hiccuped on me, i would write code to do the conversion at runtime - lol