#3 pmars assembler bug

[Anonymous]

I have a redcode source file that does not
assemble correctly with pmars
0.9.2. The problem is with a line that reads:

fnd2:  mov.i  qbomb, >fnd-(gap*rep)/2+1

Pmars 0.9.2 assembles it as if the second
operand were:

>fnd-((gap*rep)/2+1)

which is not at all the same thing.

[Pavel Šavara]

Logged In: YES
user_id=1569256
Originator: NO

Simplified version
;problem of operators precedence
mov 0, 0-2/2+11 ; this statement produce (-12)
mov 0, 0-(2/2+11) ; same as this (-12)
mov 0, 0+(-2/2)+11 ; but should produce this (10)

[Björn Günzel]

I've uploaded a tentative patch for this in the patches section.