Index
> Compiler Internals > fasm '~defined' only detects var if there is more than one. |
| Author |
|
|
BAiC 31 Dec 2012, 22:53
I am using the 'defined' operator/command (with the ~ operator) to conditionally initialize the first of potentially many variable-style labels.
when only one variable is defined then FASM doesn't register its' presence. this macro starts the code: .init.firstCallb is a unique, forward-referenced variable that is defined in the second macro (indeed all 4 labels are defined in the second macro) Code: macro ORB.Fx.Header dot
{
ali 12, 16
dw (dot # .end.code - dot)/8
dw dot # .init.firstCallb - dot
dw dot # .lastCall - dot
dw dot # .function - dot
} this is the segment of code that declares the only '.init' and therefore the only '.init.firstCall': Code:
mov r9d, Color.white
cal m.base.2to10l, .init the following is the macro that reads the non-existant '.init.firstCall' and generates the unique '.init.firstCallb' variable. Code: macro ORB.Fx.Footer dot
{
if ~defined dot # .init.firstCall
dot # .init.firstCallb = dot
dot # .lastCall = dot
else
dot # .init.firstCallb = dot # .init.firstCall
dot # .lastCall = dot # .init
end if
dot # .function = dot # .f
virtual at $
rali dot, line
dot # .end.code:
end virtual
} finally; this is the definition of the described "cal" macro. Code:
macro cal dst, lab;lab is the '.init' in 'cal m.base.2to10l, .init'
{
if lab eq
call dst
else
if ~ defined lab
db 0xE8 ; the op code for CALL rel32
lab = $ ; .init
lab # .firstCall = $ ; generating .init.firstCall
dd ((dst) and (1 shl (32-bits.ll) - 1))
} I omitted the rest of the "cal" macro in the post. the full macro for "cal" begins on line 419 of 'PseudoInstructions.asm' in the set of files I'll upload next.
_________________ byte me. |
|||||||||||
31 Dec 2012, 22:53 |
|
|
revolution 31 Dec 2012, 23:00
|
|||
|
31 Dec 2012, 23:00 |
|
|
revolution 31 Dec 2012, 23:37
BAiC wrote: revolution:both of your examples involve FASM being incapable of generating the code at all... |
|||
|
31 Dec 2012, 23:37 |
|
|
BAiC 01 Jan 2013, 00:39
revolution: FASM generated the code I gave it but the constants were wrong.
but anyway... nevermind; I've switched my code over to a token method which is working as needed. only took a few minutes to replace "if ~ defined lab" with "if lab = -1" .. and adding a "lab = -1" to the first macro (the header section). |
|||
|
01 Jan 2013, 00:39 |
|
|
l_inc 01 Jan 2013, 16:50
BAiC
Quote:
You should clearly understand, that the only situation this condition can be true is when the label lab is defined somewhere below this condition. And that happens exactly when you use the macro cal more than once. |
|||
|
01 Jan 2013, 16:50 |
|
|
BAiC 02 Jan 2013, 16:33
l_inc: it's easy to say something is obvious after the result is clearly visible.
|
|||
|
02 Jan 2013, 16:33 |
|
|
l_inc 02 Jan 2013, 18:29
BAiC
I answered, because you asked a question and you didn't show, that you got an answer and understood your mistake. Your question was still open and there was no "after". |
|||
|
02 Jan 2013, 18:29 |
|
|
BAiC 03 Jan 2013, 19:26
l_inc: regardless of your intentions the words you used did not describe anything new.
|
|||
|
03 Jan 2013, 19:26 |
|
|
l_inc 03 Jan 2013, 20:24
BAiC
You asked for an explanation of a situation you experienced. I gave it. If the explanation wasn't new for you, you shouldn't have asked. |
|||
|
03 Jan 2013, 20:24 |
|
< Last Thread | Next Thread > |
Forum Rules:
|