Ternary and AU3check

[iamtheky]

Why do I have to turn AU3check off for this  (If I leave it on I have to unnecessarily set the ternary to a variable)?

#include <Inet.au3>
#AutoIt3Wrapper_Run_AU3Check=n

$sISBN = InputBox("ISBN" , "ISBN" , "1416524525")

$sURL = "http://www.amazon.com/gp/search/ref=sr_adv_b/?search-alias=stripbooks&unfiltered=1&field-keywords=&field-author=&field-title=&field-isbn=" & $sISBN & "&field-publisher=&node=&field-p_n_condition-type=&field-feature_browse-bin=&field-subject=&field-language=&field-dateop=During&field-datemod=&field-dateyear=&sort=relevanceexprank&Adv-Srch-Books-Submit.x=0&Adv-Srch-Books-Submit.y=0"

$sBook = _INetGetSource($sURL)

$aLink = stringregexp($sBook , '\"http://www.amazon.com/gp/offer-listing/\d+"' , 3)

(IsArray($aLink)) ? (ShellExecute($aLink[0])) : (msgbox(0, '' , "No books found for " & $sISBN))

Edited December 17, 2015 by iamtheky

[SmOke_N]

My guess is that there's no assignment.

Edit:

Local $vDummy = (IsArray($aLink)) ? (ShellExecute($aLink[0])) : (msgbox(0, '' , "No books found for " & $sISBN))

Edit2:

I should state I find it annoying as well though, I've just been using 1 dummy var to do it for all of them.

Edited January 12, 2015 by SmOke_N

[kylomas]

Maybe the real question is "should AU3Check validate this construct?"

[iamtheky]

Agreed, but that seems to only be a requirement of AU3check (there is no actual error).  Is that all this is, or am I missing something larger?

@kylomas - that is indeed the question, I am not nearly as well spoken

Edited January 12, 2015 by boththose

[guinness]

This is the same in C#, only for assignment. Some people are on the fence with this one as Au3Check does one thing and the interpreter another. I would say Au3Check is correctly showing an error. The code above should ideally be using an If...Else...EndIf.

[czardas]

I see this as saying VALUE... but value what? Assigning the VALUE to a variable is saying something at least... $VAR = VALUE. Of course processeses can be run according to the condition, but (return) VALUE by itself is not quite a full statement.

I'm not saying this should be an error - I don't really know. Perhaps it's simply not the cleanest solution. One thing you can say for sure is that if this (type of) construct is by design, then the problem is with AU3 check. In relation to intended usage in AutoIt, Jon or trancexx (who introduced the ternary operator) are definately the best people to ask.

Edited January 12, 2015 by czardas

[iamtheky]

So a criteria of "an If/else that must be assigned" is common usage in other languages?  I thought I was reading "an If/else that can be assigned".   My only criteria were:

1)  must have an 'else', in that it must not be doable with just an If.

2)  must have a simple condition and simple values

edit:  my terrible grammar

Edited January 12, 2015 by boththose

[czardas]

My criteria are:

1)  must have an 'else', in that it must not be doable with just an If.

2)  the return value must actually be used - regardless of the condition.

I don't agree with the 'condition or values need to be simple' argument - I think it's totally irrelevant. Using another construct would not simplify the condition. This doesn't mean you should go out of your way to write complicated expressions - quite the opposite.

Edited January 12, 2015 by czardas

[guinness]

In C# and most other languages (I don't think C++ is one) it's used as means of assignment.

 

I don't agree with the 'condition or values need to be simple' argument

But then something like this isn't "simple" as it introduces redundancy, nor correct usage of the ternary operator.

 

var example = true ? "Some long string with a value of 200" : "Some long string with a value of 300"; // Bad

// can be shortened to...

var example = "Some long string with a value of " + (true ? "200" : "300"); // Good

Edited January 12, 2015 by guinness

[czardas]

I don't see that as a complicated condition or value. Clearly the second version is preferable, for space preservation, although there is an extra step involved (concatenation). When I say complicated, I mean the expression can't easily be simplified, if at all.

;

Func OneComplexLine()
        Local $a = Random(1, 2, 1), $b = Random(1, 2, 1)
        Return String( ( ($a ^ $b >= 3 - $b) ? Ceiling($b / $a) : Floor(3 - $a / $b) ) )
EndFunc

;

I haven't run the example.

Edit: Writing out several lines of code doesn't necessarily make the code easier to understand - if you have some inline comments to add then splitting the expression is often a good idea. I think each case is different.

Edited January 12, 2015 by czardas

[guinness]

I would say that can be considered valid usage as it's still easy to understand.

[JohnOne]

Is it not just because AU3Check has not been updated to allow for new syntax?

[guinness]

Well it was updated to accept ternary, but there was a misconception that not being used as an assignment operator was considered an error in the AutoIt interpreter, which it isn't, though it should be in my humble opinion.

Edited January 12, 2015 by guinness

[czardas]

It doesn't seem to cause any issues as it is right now. The funny thing is, I often use the ternary operator instead of using an assignment. Declaring a variable, querying a condition and then assigning a value seems an awful lot of messing about when the variable can be immediately disposed of. Why bother to create a variable if it isn't needed?

This post:

Edited January 13, 2015 by czardas

[guinness]

I am agreeing with you, what I don't agree with is this is valid In AutoIt (not Au3Check) even those most high level languages class it as invalid.
 

True ? MsgBox(4096, 'Ex', '1') : MsgBox(4096, 'Ex', '2')

[czardas]

Very good! I'm happy we agree. That's a classic bad example. If people don't get it by now, at least we tried.

Edit: Code like that should at least be labelled bad practice; and while it may be easy for me or you to see this, it may be rather difficult to clearly define the rules of usage so that the syntax can easily be checked.

Edited January 13, 2015 by czardas

[guinness]

I guess the problem really stems from saying it's an If...Else replacement, which trancexx pointed out too. So I can understand why people would see the above code (post #15) as being perfectly acceptable.

I think this is the first time we have agreed on something -_0

[czardas]

No it's not the first time I've agreed with you guinness. Sometimes I've misunderstood what you said. There are many contested arguments in all kinds of fields and sometimes I make mistakes.