another question about macro's.

[gelatine1]

So this question is about some %echo that's acting weird and I don't know why. I have the 2 small macro's below:

Code Select Expand


FLT4 MACRO float_number:REQ ;taken from siekmanski ;)
LOCAL float_num
.data
align 4
float_num real4 float_number
.code
EXITM <float_num>
ENDM

ech MACRO vars:VARARG
LOCAL temp
FOR item, <vars>
temp TEXTEQU FLT4(item)
%echo item
%echo temp
    ENDM
ENDM


If i try to use these macro's as follows:

Code Select Expand


ech(1.0)
%echo FLT4(1.0)


then I get this as output while assembling:

Code Select Expand


(1.0)
??001A
??001B

real4.asm(103) : error A2187: must use floating-point initializer
FLT4(4): Macro Called From
  MacroLoop(0): iteration 1: Macro Called From
   ech(6): Macro Called From
    real8.asm(103): Main Line Code


As far as I know I never call something similar to FLT4(4) and I don't understand the meaning of those ??001x. And how should I echo the DWORD value representing 1.0 in Real4 format ? because all my attempts have failed so far..

Thanks in advance

[mabdelouahab]

Try

Code Select Expand


ech 1.0
%echo FLT4(1.0)


[jj2007]

The ??001A are labels and variables generated for local use by the macro. You can "visualise" them with the % operator. See the guide for macro expansion operator.

[gelatine1]

Try

Code Select Expand


ech 1.0
%echo FLT4(1.0)


Didn't make any difference :/

The ??001A are labels and variables generated for local use by the macro. You can "visualise" them with the % operator. See the guide for macro expansion operator.

So I had read about this expansion operator (the % sign) (on http://www.sxlist.com/techref/language/masm/masmc09.htm) and I tried it out (maybe I am doing something wrong ?) but it doesn't work either... this is my new code and my new output:

Code Select Expand


ech MACRO vars:VARARG
LOCAL temp
FOR item, <vars>
temp TEXTEQU FLT4(item)
% echo temp
% echo %temp
% echo %(temp)
echo temp
echo %temp
echo %(temp)
    ENDM
ENDM


Code Select Expand


??001A
%??001A
%(??001A)
??0019
%??0019
%(??0019)


[qWord]

What should the output show? Your macro does return an anonym label, so the output is correct.

[gelatine1]

I'd like to see this: 0 01111111 00000000000000000000000b (or 3F800000h in hex notation or in decimal ofcourse)

I've also tried this now:

Code Select Expand

ech MACRO vars:VARARG
LOCAL temp
FOR item, <vars>
temp TEXTEQU %(FLT4(item))
% echo temp
    ENDM
ENDM

ech 1.0


but that gave the following output while assembling:

Code Select Expand


real4.asm(89) : error A2026: constant expected
MacroLoop(1): iteration 1: Macro Called From
  ech(5): Macro Called From
   real8.asm(89): Main Line Code
0


[qWord]

I'd like to see this: 0 01111111 00000000000000000000000b (or 3F800000h in hex notation or in decimal ofcourse)

That's impossible with MASM, because you can't read from segments while assembling.

[Zen]

...because you can't read from segments while assembling.

...I didn't know that,...:icon_eek:
In fact,...I know so little about writing macros, it's pathetic,...

[gelatine1]

...because you can't read from segments while assembling.

...I didn't know that,...:icon_eek:
In fact,...I know so little about writing macros, it's pathetic,...

Well .. me neither :o But I already have another small question...

Code Select Expand


.data
align 4
fl1 REAL4 1.0
fl0 REAL4 0.0
flm1 REAL4 -1.0

FLT4 MACRO float_number:REQ
LOCAL float_num,temp
temp TEXTEQU <1.0>
IF (float_number TEXTEQ temp) ;some frequent cases to save memory
EXITM fl1
ENDIF
ENDM

push FLT4(1.0)


Why does the above code give me the following error ?

Code Select Expand


real4.asm(108) : error A2050: real or BCD number not allowed
FLT4(3): Macro Called From
  real4.asm(108): Main Line Code
real4.asm(108) : error A2008: syntax error : in instruction


[mabdelouahab]

Code Select Expand

FLT4 MACRO float_number:REQ
temp SUBSTR  <float_number> ,1,3
IFIDN temp,<1.0>
EXITM <fl1>
ENDIF
IFDIF temp,<1.0>
EXITM <0> ; <---- do anyting else
ENDIF
ENDM

[gelatine1]

Code Select Expand

FLT4 MACRO float_number:REQ
temp SUBSTR  <float_number> ,1,3
IFIDN temp,<1.0>
EXITM <fl1>
ENDIF
IFDIF temp,<1.0>
EXITM <0> ; <---- do anyting else
ENDIF
ENDM

I don't understand why that SUBSTR Is necessary? And what if the user passed 1.056 for example?  your code would return fl1 which is not correct

[qWord]

Code Select Expand


IF (float_number TEXTEQ temp) ;some frequent cases to save memory

Why does the above code give me the following error ?

Because you invented new syntax elements which MASM is not aware of.

From your previous posts I deduce that you want to avoid redundant floating point constants in memory(?). To do so, you need to record the FP literals to compare them:

Code Select Expand

gfl_glb_cntr = 0
GET_REAL4_CONST macro lit
LOCAL lbl

grc_const_found = 0
grc_cntr = 0
WHILE grc_cntr LT gfl_glb_cntr
% IFIDN <&lit>,<@CatStr(<const_flt__>,%grc_cntr)>
grc_const_found = -1
EXITM
ENDIF
grc_cntr = grc_cntr + 1
ENDM
IF grc_const_found EQ 0
.const
lbl REAl4 lit
.code
@CatStr(<const_flt__>,%gfl_glb_cntr) TEXTEQU <&lit>
@CatStr(<const_flt_lbl__>,%gfl_glb_cntr) TEXTEQU <lbl>
gfl_glb_cntr = gfl_glb_cntr + 1
EXITM <lbl>
ELSE
EXITM @CatStr(<const_flt_lbl__>,%grc_cntr)
ENDIF
endm
With more macro power this method could be improved, e.g. by transforming the decimal numbers into a canonical form (123.456 -> 1.23456E2)

[gelatine1]

Yes I would like to avoid redundant floating point constants in memory. I have read your code but I don't really understand these lines:

Code Select Expand


const
lbl REAl4 lit
.code
@CatStr(<const_flt__>,%gfl_glb_cntr) TEXTEQU <&lit>
@CatStr(<const_flt_lbl__>,%gfl_glb_cntr) TEXTEQU <lbl>


What is the meaning of the & before lit? And what's the difference between the constflt and the constfltlbl?  And are they global variables or something?

[GoneFishing]

...
From your previous posts I deduce that you want to avoid redundant floating point constants in memory(?).
...

Could you please clarify why "floating point constants in memory" are redundant ?
Comes to memory when Japheth called COM atavistic technology .

[qWord]

Could you please clarify why "floating point constants in memory" are redundant ?

Code Select Expand


.const
x REAL4 1.0
y REAL4 1.0
z REAL4 1.0