[Squeakfoundation]ContextCleanupPlus-ajh (was: Re: KCP & 3.6)

[Stephane Ducasse]

Thanks anthony.
So this sounds good then from the simple naive superficial review I did.

By the way have you thought about having a test for the equalLiteral and
add in the method the comment you gave me?

>> Anthony Hannan wrote:
>>> Class>>becomeClass: otherClass - Become forward the other class and
>>> fix up the global dictionary and system organization as well.
>> Do you have some test for these one because I would like to know what
>> is the consequence limit of becomeClass: especially in terms of the
>> metaclass? Does it work on metaclass too?
>
> Yes, but I have taken this change out as well as several others  
> because I don't
> use them anymore.  If someone wants them I can post them in a separate  
> package.

Please do.
Because this may be something quite interesting to have

>
>> By the way what is the difference between literal equal and ==. I'm
>> sure there is one but your example does not bring that up front.
>> 'foo' = #foo, but ('foo' literalEqual: #foo) not.
>
> When compiling a method, multiple references to the same literal are  
> combined
> together into a single literal and held in the compiled method.   
> Suppose you
> have both 'foo' and #foo in your method.  If the compiler uses #= it  
> would
> incorrectly combine 'foo' and #foo.  Hence it uses #literalEqual:  
> instead.
> Previously this test was coded in long form in the compiler and other  
> places
> like CompiledMethod>>= that tested for literals..
>
>> I want to know if these protocols work well in subclasses
>>> SequenceableCollection>>allButFirstDo:,allButLastDo:,atLast:,atLast:p 
>>> ut
>>> :
>>> ,copyGrowBy:copyWithFirst:,detectIndex:,reverseDetect:,slide:by:,with 
>>> :w
>>> i
>>> thIndexDo:
>
> As far as I know they work in subclasses.  Some of been removed like
> becomeClass: because they are no longer used.
>
>> I really think that we should not introduce this that way.
>>> Collection>>collectArray: - Same as collect: except always return an
>>> array.
>> Normally collect returns the same kind of collection that the one that
>> is iterated, so what are the other alternatives.
>> Anthony why do you need it and why (mycol collect: []) asArray or
>> myColl asArray collect: []
>> does not work. I imagine that this is too costly. So why don't you
>> manipulate directly array or why do you need array at the end?
>
> Your right.  I removed this.
>
>>> OrderedCollection>>bottom,top,pop,push,etc. - stack protocol.
>> I really have a problem with this one. because I will get an
>> orderedCollection sends it but I do not know what push means in
>> terms of OrderedCollection because I will be able to send other  
>> message
>> of orderedCollection to the collection and I can mess up everything.
>> so why don't you implement a Stack if you need one and we can this
>> class to the collection hierarchy.
>
> Your right again.  I removed this stack protocol from Collection.  And  
> changed
> my code to use normal OrderedCollection protocol.
>
>>> Number>>extend: - Like #to: except designate the length instead of  
>>> the
> last.
>> DNU (but did not get the time to look inside)
>
> I removed this one as well.
>
>> and also to the point of Smacc. Using if using the output of a
>> tool forces us to have the same license that the tool, I stop CS.
>> Because I use latex and it does not mean that my articles are GNU
>> license.
>
> Goran answered this.  We want everything changeable within Squeak.
>
> Cheers,
> Anthony
>
> __________________________________
> Do you Yahoo!?
> SBC Yahoo! DSL - Now only $29.95 per month!
> http://sbc.yahoo.com
> _______________________________________________
> Squeakfoundation mailing list
> Squeakfoundation at lists.squeakfoundation.org
> http://lists.squeakfoundation.org/listinfo/squeakfoundation
>