Good coding practices in AutoIt

[AZJIO]

guinness

Excessive equating

[guinness]

I was just showing another approach.

[BrewManNH]

 

Good point, also remove the = 1 and you will see it produces the same result BrewManNH.

Very true, missed the second line in the remarks section. I was going on the assumption that the first line, where it says the first variable will be zero, held true throughout.

Although, because things might change in the future, it's better to be safe than sorry, so declaring it won't break it in the future.

[Ascend4nt]

I tend to use quotes with "good/best coding practices" because that's really a subjective thing.  Writing code that ignores certain guidelines, but runs fast, can be considered good coding practice if it's well-documented as to what it does (and is hopefully future-proof).  There are a number of threads discussing speed optimizations, many of which don't actually produce hard-to-read code..

Anyway, back to the topic..

guinness, when using BitAND to test individual bits in a flag, there's no need to use a comparison afterwards - the fact that its non-zero is enough.  i.e. this can be done:

    If BitAND($iFlag, $IF_FLAG_ONE) Then MsgBox($MB_SYSTEMMODAL, '', '$IF_FLAG_ONE was called.')
    If BitAND($iFlag, $IF_FLAG_TWO Then MsgBox($MB_SYSTEMMODAL, '', '$IF_FLAG_TWO was called.')
    If BitAND($iFlag, $IF_FLAG_THREE) Then MsgBox($MB_SYSTEMMODAL, '', '$IF_FLAG_THREE was called.')

  
It's important to remember though that this is useful only if those flags map to an individual bit. In other words, the value must be "1" or any power of 2 (1, 2, 4, 8, 16).  If there's a combination of bits, say 11 (which in binary is 1011), then you'll need to do the comparison (to 11).

[Ascend4nt]

Short-circuit evaluation [wiki] is a useful means to avoid multiple If-Then statements, as well as reducing duplicate code.  The operators AND and OR are used to make short-circuit evaluations.  For example, below are two ways to check if something is an array that has a non-zero value as its first element:

; Long, ugly, slow Way:
If IsArray($aRet) Then
    If $aRet[0] <> 0 Then
        ConsoleWrite("Array with non-zero at [0]" & @CRLF)
    EndIf
EndIf

; Short, Concise, faster way:
If IsArray($aRet) And $aRet[0] <> 0 Then
   ConsoleWrite("Array with non-zero at [0]" & @CRLF)
EndIf

What might not be evident is the importance of the order expressions appear in on a short-circuit evaluation. We can not test if "$aRet[0] <> 0" above UNLESS we know that $aRet is an array, because the program would crash otherwise.  During the evaluation the expression will fail at "IsArray($aRet)" if it evaluates to False, and the interpreter will not evaluate the rest of the expression.  This short-circuiting is the same as the If-Then-If-Then code above it.  The rules for short-ciruit evaluation rules are pretty common among C and Basic-like languages:

 

Given "Expr1 AND Expr2", we can observe that:

1. If Expr1 yields False, Expr2 is NOT  evaluated
2. If Expr1 yields True,  Expr2 IS then evaluated
3. The following code is therefore equivalent:

  If Expr1 Then
    If Expr2 Then
      EvaluationSuccess
    EndIf
  EndIf

_

Likewise "Expr1 OR Expr2" short circuits in an opposite manner:

1. If Expr1 yields True,  Expr2 is NOT  evaluated
2. If Expr1 yields False, Expr2 IS then evaluated
3. The following code is therefore equivalent:

  If Expr1 Then
    EvaluationSuccess
  ElseIf Expr2 Then
      EvaluationSuccess
  EndIf

_
Since comparisons, function returns, and any other result that is boolean-testable can appear in an equation, we can do things like the below, which avoids lots of IF-Then-Else statements:

If PlayerHere($nUserID) And CheckPlayerStatus($nUserID) = $USER_IS_CAMPING Or BanRequests($nUserID) >= $KICK_THRESHOLD Then
    EjectPlayer($nUserID)
    ConsoleWrite("User ejected" & @CRLF)
EndIf

To review the logic above, if "PlayerHere()" evaluates to FALSE, the rest of the line is not evaluated (short-circuits at "AND"), and the code following EndIf is run next.  Otherwise, "CheckPlayerStatus()" is called and compared to a status.  If THAT evaluates to TRUE, the interpreter short-circuits at the next OR statement (and runs the code starting with EjectPlayer()) .  However, if the 2nd comparison evaluates to FALSE, the final item is checked. Depending on the last comparison, either EjectPlayer() is run next (if TRUE), or the code following EndIf is run next (if FALSE).  Hopefull that's clear?  In any case, that one line is clear enough to read and avoids a lengthy If-Then-Else code block.

[guinness]

Thanks Ascend4nt a lot reading to do. I suppose the BiAND comparison has become a bad habit of mine, but I sometimes use BitAND(4, 2) = 2 when I need a boolean return datatype.

[FireFox]

Well done, Ascend4nt.

[AZJIO]

For the "If-Then" I did the error description, translation

[guinness]

Circa December 2010. 

expandcollapse popup

#AutoIt3Wrapper_Au3Check_Parameters=-d -w 1 -w 2 -w 3 -w 4 -w 5 -w 6
#include <Array.au3>

Global $Global_Array[1][2] = [[0, 2]]

_Main()

Func _Main()
    Local $B = 10, $hGUI, $iButton
    $hGUI = GUICreate("_GetControlID() Example 2")
    GUISetState(@SW_SHOW)
    For $A = 1 To 10
        GUICtrlCreateLabel("Example Label_" & $A, 5, $B, 150, 20)
        _ArrayAddControl(-1)
        $B += 25
    Next
    $iButton = GUICtrlCreateButton("Show Array", 5, 400 - 30, 65, 22.5)
    While 1
        Switch GUIGetMsg()
            Case -3
                ExitLoop

            Case $iButton
                _ArrayDisplay($Global_Array)

        EndSwitch
    WEnd
    GUIDelete($hGUI)
EndFunc   ;==>_Main

Func _ArrayAddControl($iControlID = -1)
    Local $aControlID, $hControlID
    $hControlID = GUICtrlGetHandle($iControlID)
    $aControlID = DllCall("user32.dll", "int", "GetDlgCtrlID", "hwnd", $hControlID)
    If @error Then
        Return SetError(1, 0, 0)
    EndIf
    $iControlID = $aControlID[0]
    If UBound($Global_Array, 1) <= $Global_Array[0][0] + 1 Then
        ReDim $Global_Array[UBound($Global_Array, 1) * 2][$Global_Array[0][1]]
    EndIf
    $Global_Array[0][0] += 1
    $Global_Array[$Global_Array[0][0]][0] = $iControlID
    $Global_Array[$Global_Array[0][0]][1] = GUICtrlRead($iControlID)
    ReDim $Global_Array[$Global_Array[0][0] + 1][$Global_Array[0][1]]
    Return $iControlID ; Returns The ControlID, But It Could Return Another Value.
EndFunc   ;==>_ArrayAddControl

Edited June 4, 2013 by guinness

[Danyfirex]

using Opt("MustDeclareVars", 1)  is a good example of good coding .

[PlayHD]

using Opt("MustDeclareVars", 1)  is a good example of good coding .

why?

[guinness]

using Opt("MustDeclareVars", 1)  is a good example of good coding .

This is better.

#AutoIt3Wrapper_Au3Check_Parameters=-d -w 1 -w 2 -w 3 -w- 4 -w 5 -w 6 -w- 7

[FireFox]

@PlayHD

Why? By declaring your variables you define their scope, and the compiler (at least the executable interpreter) has not to do the job for you.

Edited June 6, 2013 by FireFox

[PlayHD]

@Firefox, thankyou:)

 

This is better.

#AutoIt3Wrapper_Au3Check_Parameters=-d -w 1 -w 2 -w 3 -w- 4 -w 5 -w 6 -w- 7

 

What that means ?

[Melba23]

PlayHD,

If you use the full SciTE4AutoIt3 package you automatically run a syntax check with AU3Check before compiling. That line sets the parameters for that check - and is set to the highest level. We ask that all UDFs and Wiki Snippets are checked to that level before submission.

M23

[AZJIO]

If the variable, such as a trigger is set to 0 or 1, how best to use the prefix $iVar or $fVar? By type of $iVar, but the meaning of $fVar

Edited June 8, 2013 by AZJIO

[guinness]

I use $fVar for flag.

[FireFox]

I would use $iVar, it's better to be as clear as possible (don't mix var types).

 

Edit: Except if you use $bl for booleans (but a flag is not a variable type anyway).

Edited June 8, 2013 by FireFox

[AZJIO]

$i_fVar?

[FireFox]

Do it the way you like, there is nothing "official" for this.

I saw $aiArray or $asArray etc so it would be $ifVar.

Edit: In the Array UDF there is $avArray.

Naming a variable :

type (lower case), subtype (lower case) - optional, var name (first letter in upper case).

Edited June 8, 2013 by FireFox