Is using Ubound in For To Next declarations loops bad practice

[guinness]

czardas,

Read this

[guinness]

Proof that arrays are only passed when 'copy-on-write'.

Note: Requires WinAPIEx or AutoIt Beta, you choose.

expandcollapse popup

#include <WinAPIEx.au3>

SomeFunc()

Func SomeFunc()
    Local Const $iCount = 999999
    Local $aArray[$iCount]
    For $i = 0 To UBound($iCount) - 1
        $aArray[$i] = $i
    Next

    MemoryCheck('Array before passing to normal function')
    ArrayFuncNormal($aArray)
    MemoryCheck('Array after passing to normal function')

    ConsoleWrite(@CRLF)

    MemoryCheck('Array before passing to normal function with write access')
    ArrayFuncNormalWrite($aArray)
    MemoryCheck('Array after passing to normal function with write access')

    $aArray = 0
    Return True
EndFunc   ;==>SomeFunc

Func ArrayFuncNormal($aArray)
    MemoryCheck('Array inside the normal normal function')
EndFunc   ;==>ArrayFuncNormal

Func ArrayFuncNormalWrite($aArray)
    $aArray[0] = 0
    MemoryCheck('Array inside the normal normal function with writing')
EndFunc   ;==>ArrayFuncNormal

Func MemoryCheck($sData)
    Local $aMem = _WinAPI_GetProcessMemoryInfo(@AutoItPID)
    ConsoleWrite($sData & ': ' & _ByteSuffix($aMem[7]) & @CRLF)
    $aMem = 0
EndFunc   ;==>MemoryCheck

Func _ByteSuffix($iBytes, $iRound = 2) ; By Spiff59
    Local $iIndex, $aArray[9] = [' bytes', ' KB', ' MB', ' GB', ' TB', ' PB', ' EB', ' ZB', ' YB']
    While $iBytes > 1023
        $iIndex += 1
        $iBytes /= 1024
    WEnd
    Return Round($iBytes, $iRound) & $aArray[$iIndex]
EndFunc   ;==>_ByteSuffix

[czardas]

czardas,

Read this

Now I feel like going back to only using Dim. Only joking. Thanks guinness!

[guinness]

Now I feel like going back to only using Dim. Only joking. Thanks guinness!

You're welcome and I hope you are.

[BrewManNH]

Irrespective of anyone's thoughts on the process, if the local variable is declared outside of a function, it's a Global variabe regardless of what you put in front of it.

[czardas]

You're welcome and I hope you are.

It's a promise.

The dim days are now a thing of the past.

Edited January 3, 2013 by czardas

[guinness]

Irrespective of anyone's thoughts on the process, if the local variable is declared outside of a function, it's a Global variabe regardless of what you put in front of it.

Though this might be true regarding how AutoIt handles Local variables outside of a function, it's still best practice to do this. As to the reason why I presume is obvious.

[BrewManNH]

It's never been obvious to me, but I just go with the flow.

[guinness]

I would say readability is a major one. You can easily tell which is truly a Global variable and which isn't.

Today has been a major day of teaching and learning, wow! I'm truly knackered right now.

Edited January 3, 2013 by guinness

[czardas]

I concur with you BrewmanNH. I thought I had this scope thing nailed a while ago. I did understand the distinction by Valik after a couple of reads. It's all in the last sentence.

Edited January 3, 2013 by czardas

[guinness]

I'm always learning, even now I question if it's beneficial to use a Global variable over passing it as a parameter, as sometimes this is the case. Which is why I don't fully agree that 'All Global variables are bad' theory.

Edited January 3, 2013 by guinness

[kylomas]

@guinness -

'All Global variables are bad'

This would indeed be foolish. Assuming that the code is functionally correct it now comes down to coding style and maintainability. Hence the importance of good naming conventions, in my opinion.

What the hell does "knackered" mean?

kylomas

edit: incidentally, nice demo with arrays!!

Edited January 3, 2013 by kylomas

[czardas]

I think there will nearly always be exceptional cases which are less obvious. Yeah it's a good learning experience talking to you guys. I know I can be stubborn at times, but you have a lot of patience fortunately.

[czardas]

What the hell does "knackered" mean?

Tired or broken - ready for the knacker's yard.

Edited January 3, 2013 by czardas

[guinness]

I think there will nearly always be exceptional cases which are less obvious. Yeah it's a good learning experience talking to you guys. I know I can be stubborn at times, but you have a lot of patience fortunately.

Well what I say isn't gospel and I'm sure I'm wrong on many occasion.

I love discussions of this nature, because sooner or later when I do learn C/C++ it should be easier than going in head first.

[kylomas]

@guinness - then I guess it means "junk yard", cool.

I love discussions of this nature, because sooner or later when I do learn C/C++ it should be easier than going in head first.

Yes, and in some ways "strongly typed" languages are easier to work with as the ambiguity of what a variable is/does is reduced.

[trancexx]

Proof that arrays are only passed when 'copy-on-write'.

Note: Requires WinAPIEx or AutoIt Beta, you choose.

expandcollapse popup

#include <WinAPIEx.au3>

SomeFunc()

Func SomeFunc()
    Local Const $iCount = 999999
    Local $aArray[$iCount]
    For $i = 0 To UBound($iCount) - 1
        $aArray[$i] = $i
    Next

    MemoryCheck('Array before passing to normal function')
    ArrayFuncNormal($aArray)
    MemoryCheck('Array after passing to normal function')

    ConsoleWrite(@CRLF)

    MemoryCheck('Array before passing to normal function with write access')
    ArrayFuncNormalWrite($aArray)
    MemoryCheck('Array after passing to normal function with write access')

    $aArray = 0
    Return True
EndFunc   ;==>SomeFunc

Func ArrayFuncNormal($aArray)
    MemoryCheck('Array inside the normal normal function')
EndFunc   ;==>ArrayFuncNormal

Func ArrayFuncNormalWrite($aArray)
    $aArray[0] = 0
    MemoryCheck('Array inside the normal normal function with writing')
EndFunc   ;==>ArrayFuncNormal

Func MemoryCheck($sData)
    Local $aMem = _WinAPI_GetProcessMemoryInfo(@AutoItPID)
    ConsoleWrite($sData & ': ' & _ByteSuffix($aMem[7]) & @CRLF)
    $aMem = 0
EndFunc   ;==>MemoryCheck

Func _ByteSuffix($iBytes, $iRound = 2) ; By Spiff59
    Local $iIndex, $aArray[9] = [' bytes', ' KB', ' MB', ' GB', ' TB', ' PB', ' EB', ' ZB', ' YB']
    While $iBytes > 1023
        $iIndex += 1
        $iBytes /= 1024
    WEnd
    Return Round($iBytes, $iRound) & $aArray[$iIndex]
EndFunc   ;==>_ByteSuffix

That code proofs nothing considering it has major logical error.

guinness you have good programming skills and I very much appreciate that, but what you lack is higher degree of critical thinking that would make you much much better. Please don't read this as any form of bashing, rather as constructive criticism.

To prove something correctly, the correctness of the method used for proving is essential. When there is an error in method then even if the result is correct it cannot be classified as valid because of fallacious middle-step. That's why I can say "your proof is incorrect" and I can prove it. See, you are actually not filling the array. Your code is based on (unintentional) assumption that every array element contains data - occupies memory if it exists. That is something you haven't proven and therefore can't be used to proof anything furher (before it is actually proven). This is rather common error fundamentalis logical fallacy.

Another thing is that you have choosen to prove copy-on-wright, but omit copy-on-read. The absence of that particular proof (regardless of the conclusion) makes your final conclusion prone to misinterpretations because of simple bifurcation.

You mistake is clearly unintentional, but unfortunatly like you said for something else before, things like this get to stuck with people and that's why it's important to get them right.

[guinness]

That code proofs nothing considering it has major logical error.

guinness you have good programming skills and I very much appreciate that, but what you lack is higher degree of critical thinking that would make you much much better.

I don't see it as bashing, how does one develop further without criticism? Plus you gave me a compliment which is much appreciated.

See, you are actually not filling the array. Your code is based on (unintentional) assumption that every array element contains data - occupies memory if it exists. That is something you haven't proven and therefore can't be used to proof anything furher (before it is actually proven). This is rather common error fundamentalis logical fallacy.

OK, as I thought filling the array with an integer was achieving this.

Another thing is that you have choosen to prove copy-on-wright, but omit copy-on-read. The absence of that particular proof (regardless of the conclusion) makes your final conclusion prone to misinterpretations because of simple bifurcation.

You mistake is clearly unintentional, but unfortunatly like you said for something else before, things like this get to stuck with people and that's why it's important to get them right.

That's fair, I didn't show copy-on-read. I would rather know the correct approach than teach people the wrong, which has often been the case. Edited January 3, 2013 by guinness

[trancexx]

OK, as I thought filling the array with an integer was achieving this.

It would, but check your code again. You aren't filling anything.

[kylomas]

In the for loop you are using the wrong variable for the TO value. Should be

for $i = 0 to ubound($aArray) -1