Index
> Compiler Internals > Addressing mode bug |
| Author |
|
|
vid 17 Dec 2006, 17:12
nice one
 |
|||
17 Dec 2006, 17:12 |
|
|
Tomasz Grysztar 17 Dec 2006, 18:12
Indeed, nice one, and about as old as fasm itself. Fixed in 1.67.16.
|
|||
|
17 Dec 2006, 18:12 |
|
|
kohlrak 17 Dec 2006, 20:11
El Tangas wrote: I actually got some productive result out of the xor reg, reg thread runing in are you saying that using the following code to 0 a register is bad? Code: xor eax, eax |
|||
|
17 Dec 2006, 20:11 |
|
|
El Tangas 17 Dec 2006, 20:39
vid, Tomasz: Well, tnx, guys, I do try...
kohlrak: Ah, Good and Bad... It is my personal belief that all forms of zeroing a register have their place in this world. So xor reg, reg is ok for me. However, this is not the right place to discuss this, since there is that thread in the Main forum. |
|||
|
17 Dec 2006, 20:39 |
|
|
LocoDelAssembly 29 Aug 2009, 23:06
A similar issue is still present:
Code: irp m, -1, -2, -3, -4, -5, -8, -9 { alpha = m lea eax, [eax*m] lea ecx, [esi*(m)] lea edx, [ebx*alpha] ; Only to confirm it doesn't fail with literals only } |
|||
|
29 Aug 2009, 23:06 |
|
|
Tomasz Grysztar 30 Aug 2009, 00:11
Wow, this one was really old. The code I had to modify to fix it was not touched since the dark ages in the very beginning of fasm development.
Fixed in 1.69.04. |
|||
|
30 Aug 2009, 00:11 |
|
|
LocoDelAssembly 30 Aug 2009, 00:45
hehe well, this bug is probably really hard to be noticed in real code. Actually I found it because after reading some stuff about formal methods I though of testing this, I wasn't doing any Assembly programming today.
There is something I'm not sure if it is right or wrong. This works: Code: lea eax, [eax-eax*-2] ; Assembled as "lea eax, [eax+eax*2]" lea eax, [0-eax*-2] ; Assembled as "lea eax, [eax+eax]" Code: lea eax, [-eax*-2] ; Invalid expression (should it be assembled as "lea eax, [eax+eax]"?) |
|||
|
30 Aug 2009, 00:45 |
|
< Last Thread | Next Thread > |
Forum Rules:
|