Multiple Constants equal one When added.

[GenRabbit]

Example 1;

Code: Select all

#testvalue1 = 98765
#testvalue2 = 98765
Define x.l
    x = #Testvalue1 + 1000
    x = #Testvalue2 + 2000

Example 2;

Code: Select all

#testvalue1 = 99765
#testvalue2 = 100765
Define x.l
    x = #Testvalue1
    x = #Testvalue2

Any chance that future updates will when compiling make those codesnippet generate equal code?

[nco2k]

what?

c ya,
nco2k

[GenRabbit]

In both examples x becomes '99765' in the first line of code. Written 99765 or 98765 + 1000 shouldn't matter. So why different compiled code.

[Bisonte]

I must also say : WHAT ?

I don't understand nothing of your question ...

[Dude]

[Removed my assumption]

[Demivec]

In both examples x becomes '99765' in the first line of code. Written 99765 or 98765 + 1000 shouldn't matter. So why different compiled code.

How do you know the code is compiled differently?

I compiled both and they are identical.

For fun, here is a commented version of the portions that were different (because the source in the comments is different):

Code: Select all

;Example 1
; 
; #testvalue1 = 98765
; #testvalue2 = 98765
; Define x.l
; x = #Testvalue1 + 1000
  MOV    dword [v_x],99765
; x = #Testvalue2 + 2000
  MOV    dword [v_x],100765
_PB_EOP_NoValue:
  PUSH   dword 0
_PB_EOP:
  CALL  _PB_EndFunctions
  PUSH   dword [PB_MemoryBase]
  CALL  _HeapDestroy@4
  CALL  _ExitProcess@4
_PB_EndFunctions:
  RET 

Code: Select all

;Example 2
; 
; #testvalue1 = 99765
; #testvalue2 = 100765
; Define x.l
; x = #Testvalue1
  MOV    dword [v_x],99765
; x = #Testvalue2
  MOV    dword [v_x],100765
_PB_EOP_NoValue:
  PUSH   dword 0
_PB_EOP:
  CALL  _PB_EndFunctions
  PUSH   dword [PB_MemoryBase]
  CALL  _HeapDestroy@4
  CALL  _ExitProcess@4
_PB_EndFunctions:
  RET

You'll notice that PureBasic folded (combined) the constants just as you suggested.

[GenRabbit]

I compiled both and then run a CRC32 check, and the CRC32 values of the two compiled files did not match. So I took it, they where different.
Thanks for taking the time to point it out. That I was wrong was good to hear.

[Dude]

I compiled both and then run a CRC32 check, and the CRC32 values of the two compiled files did not match.

That's just due to the time/date of the compile. To test, I compiled the same source twice, with only 10 seconds between compiles, and both exes had different CRC32 checksums.