oh, well... but it's also doable this way:
guessTypeOf: aNumber
| range |
self configureSelectorsOnly.
recurseCount > 10 ifTrue: [ ^ nil ].
recurseCount := recurseCount + 1.
range := ranges at: aNumber.
range isSelf ifTrue: [ ^theClass ].
range isSuper ifTrue: [ ^theClass superclass ].
range isConstant ifTrue: [ ^range asType ].
range isArgument ifTrue: [ ^self guessArgument: range ].
range isTemporaryVariable ifTrue: [ ^self guessTempVarClass: range ].
range isInstanceVariable ifTrue: [ ^self guessInstVarClass: range ].
range isClassVariable ifTrue: [ ^self guessClassVarClass: range ].
range isGlobal ifTrue: [ ^self guessGlobal: aNumber ].
which of course has it's drawbacks (like multiple returns, which I also was
told it is usually a bad thing), but I think is much clearer... and the usual
way to do case statements on smalltalk.
btw... this implementation reveals a bug: there is no default behavior if none
of the conditions is fulfilled. In that case, it will answer "self", and I
think that's wrong (I think it should answer nil). Not sure, because never dug
into this code before, but looks like :)
On Apr 23, 2012, at 12:49 PM, Nicolas Cellier wrote:
> These are not really nested if, but rather a kind of caseOf:
> implemented with elseif.
> It's not to say that the style is ideal...
> I did not read the code, and don't know what a range is, nor if it has
> subclass specializations.
> If so, maybe some actions could simply be dispatched to range.
>
> Nicolas
>
> Le 23 avril 2012 10:58, Esteban Lorenzano <[email protected]> a écrit :
>> he... is not related with the error, but... did you see this method?
>>
>> guessTypeOf: aNumber
>> | range |
>> self configureSelectorsOnly.
>> recurseCount > 10 ifTrue: [ ^ nil ].
>> recurseCount := recurseCount + 1.
>> range := ranges at: aNumber.
>> ^ range isSelf
>> ifTrue: [ theClass ]
>> ifFalse:
>> [ range isSuper
>> ifTrue: [ theClass superclass ]
>> ifFalse:
>> [ range isConstant
>> ifTrue: [ range asType ]
>> ifFalse:
>> [ range isArgument
>> ifTrue: [
>> self guessArgument: range ]
>> ifFalse:
>> [
>> range isTemporaryVariable
>>
>> ifTrue: [ self guessTempVarClass: range ]
>>
>> ifFalse:
>>
>> [ range isInstanceVariable
>>
>> ifTrue: [ self guessInstVarClass: range ]
>>
>> ifFalse:
>>
>> [ range isClassVariable
>>
>> ifTrue: [ self guessClassVarClass: range ]
>>
>> ifFalse: [ range isGlobal ifTrue: [ self
>> guessGlobal: aNumber ] ] ] ] ] ] ] ]
>>
>>
>> when I was younger, at informatics school, they told me that "3 nested ifs
>> means bad design, with no doubt"... what about 7?
>>
>>
>> On Apr 23, 2012, at 10:30 AM, Esteban Lorenzano wrote:
>>
>>> yes, I can reproduce this one... taking care of it now :)
>>>
>>> On Apr 23, 2012, at 10:11 AM, Stéphane Ducasse wrote:
>>>
>>>> Note that the bug shows up if you edit the code in the result of the
>>>> implementors.
>>>> Esteban can you reproduce the bug?
>>>>
>>>> Stef
>>>>
>>>> On Apr 23, 2012, at 9:18 AM, Stéphane Ducasse wrote:
>>>>
>>>>> Hi
>>>>>
>>>>> I get a bug with OCompletion/Ecompletion
>>>>>
>>>>> TextEditor>>changeEmphasis
>>>>>
>>>>> try to type
>>>>>
>>>>> menuList nextPut:
>>>>>
>>>>> and you get a bug.
>>>>>
>>>>> Now in addition I do not understand what I see in the debugger:
>>>>>
>>>>> As you see model returns a debugger and the debugger title show
>>>>> FlatMessageListBrowser
>>>>>
>>>>> Stef
>>>>>
>>>>>
>>>>>
>>>>> <Screen Shot 2012-04-23 at 9.08.45 AM.pdf>
>>>>
>>>>
>>>
>>
>>
