On Mon, Apr 18, 2016 at 5:18 PM, Levente Uzonyi <leves at caesar.elte.hu>
wrote:

> On Mon, 18 Apr 2016, Eliot Miranda wrote:
>
>
>>
>> On Mon, Apr 18, 2016 at 3:43 PM, Levente Uzonyi <leves at caesar.elte.hu>
>> wrote:
>>       Hi Eliot,
>>
>>       On Mon, 18 Apr 2016, Eliot Miranda wrote:
>>
>>             Hi All,
>>                 this used to work:
>>
>>                 (Set with: #module) = (IdentitySet with: #module)
>>
>>             when did this change and why?
>>
>>
>>       I changed it in Collections-ul.669 last November, because it didn't
>> work well. Let's see a different example:
>>
>>       | s i |
>>       s := Set with: 'module' copy.
>>       i := IdentitySet with: 'module' copy.
>>       {
>>               s = i.
>>               i = s
>>       }
>>
>>       The old implementation would return #(false true), while the new
>> one evaluates to #(false false).
>>
>>
>> Right.  That's a pain, but fair enough.  This is clumsy: (aSet size = 1
>> and: [aSet anyOne = #module]).  This slightly less so:  aSet = (aSet
>> species with: #module).
>>
>
> You could also write it as
>
> aSet asArray = #(module).
>
> or as
>
> (aSet as: Set) = (Set with: #module).


Expensive when aSet is large.  I've gone with the anyOne form.



> Levente
>
>
>
>>       Levente
>>
>>             _,,,^..^,,,_
>>             best, Eliot
>>
>>
>> _,,,^..^,,,_
>> best, Eliot
>>
>
_,,,^..^,,,_
best, Eliot
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.squeakfoundation.org/pipermail/squeak-dev/attachments/20160419/89748714/attachment.htm