keyAtValue: -> keysAtValue:

>
> keyAtValue: returns one key of the value specified but if you have more time
> the same value the method return one key.
> An example:
>      t:=Dictionary new.
>      t at: 'toto' put: 4.
>      t at: 'foo'  put: 4.
>      t keyAtValue:4. -> 'foo'
>
> So we can't recuperate all of the keys. I propose to create the methods
> keysAtValue: who return an Array with all the keys for this value.
>
> I'm not in Pharo very long time so I can not say if it can be useful so I am
> asking to you the question.
>
> Tristan Bourgois.
> --
> View this message in context: http://forum.world.st/keyAtValue-keysAtValue-tp2165484p2165484.html
> Sent from the Pharo Smalltalk mailing list archive at Nabble.com.
>
> _______________________________________________
> Pharo-project mailing list
> [hidden email]
> http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project

> Hi,
>
> I think one question one should always ask: Do I have a real need for the new method?
>
> Sometimes it makes sense to add new API methods even outside of direct need.
> Consistency is the most important reason to add something.
>
> So is there a direct need for your method?
>
> Marcus
>
>
>
> On May 10, 2010, at 4:36 PM, Tristan Bourgois wrote:
>
>>
>> keyAtValue: returns one key of the value specified but if you have more time
>> the same value the method return one key.
>> An example:
>>     t:=Dictionary new.
>>     t at: 'toto' put: 4.
>>     t at: 'foo'  put: 4.
>>     t keyAtValue:4. -> 'foo'
>>
>> So we can't recuperate all of the keys. I propose to create the methods
>> keysAtValue: who return an Array with all the keys for this value.
>>
>> I'm not in Pharo very long time so I can not say if it can be useful so I am
>> asking to you the question.
>>
>> Tristan Bourgois.
>> --
>> View this message in context: http://forum.world.st/keyAtValue-keysAtValue-tp2165484p2165484.html
>> Sent from the Pharo Smalltalk mailing list archive at Nabble.com.
>>
>> _______________________________________________
>> Pharo-project mailing list
>> [hidden email]
>> http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project
>
> --
> Marcus Denker  -- http://www.marcusdenker.de
> INRIA Lille -- Nord Europe. Team RMoD.
>
>
> _______________________________________________
> Pharo-project mailing list
> [hidden email]
> http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project

>
> On May 11, 2010, at 9:41 AM, Marcus Denker wrote:
>
>> Hi,
>>
>> I think one question one should always ask: Do I have a real need for the new method?
>>
>> Sometimes it makes sense to add new API methods even outside of direct need.
>> Consistency is the most important reason to add something.
>>
>> So is there a direct need for your method?
>>
>> Marcus
>>
>>
>>
>> On May 10, 2010, at 4:36 PM, Tristan Bourgois wrote:
>>
>>>
>>> keyAtValue: returns one key of the value specified but if you have more time
>>> the same value the method return one key.
>>> An example:
>>>    t:=Dictionary new.
>>>    t at: 'toto' put: 4.
>>>    t at: 'foo'  put: 4.
>>>    t keyAtValue:4. -> 'foo'
>>>
>>> So we can't recuperate all of the keys. I propose to create the methods
>>> keysAtValue: who return an Array with all the keys for this value.
>>>
>>> I'm not in Pharo very long time so I can not say if it can be useful so I am
>>> asking to you the question.
>>>
>>> Tristan Bourgois.
>>> --
>>> View this message in context: http://forum.world.st/keyAtValue-keysAtValue-tp2165484p2165484.html
>>> Sent from the Pharo Smalltalk mailing list archive at Nabble.com.
>>>
>>> _______________________________________________
>>> Pharo-project mailing list
>>> [hidden email]
>>> http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project
>>
>> --
>> Marcus Denker  -- http://www.marcusdenker.de
>> INRIA Lille -- Nord Europe. Team RMoD.
>>
>>
>> _______________________________________________
>> Pharo-project mailing list
>> [hidden email]
>> http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project
>
>
> _______________________________________________
> Pharo-project mailing list
> [hidden email]
> http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project

>
> On May 11, 2010, at 10:35 AM, Stéphane Ducasse wrote:
>
>> But marcus the problem is right now keyAtValue: is bogus. Did you see it?
>>
>>
>
> It does work when ypu know that there is only one value that is identical.
>
> Than you can use it to map back and forth. (slowly one of them, of course).
>
> If there are two values that are the same, the message does not make that much sense.
>
> I guess this method is used for just the cases when the client needs a table for maping in
> both directions.
>
> The only client in the Core image is ServerDirectory:
>

>
> On May 11, 2010, at 11:20 AM, Tristan Bourgois wrote:
>
>>
>> I understand what you said.
>> But I think keyAtValue: is not pertinent. In the case where there are more
>> than one same values, you can't know which value is selected.
>>
>> I should say that replace keyAtValue: by keysAtValue: because, I think, it's
>> more pertinent.
>>
>
> And clients that want to use the back-mapping can just do
>
> (myDict keysAtValue: 'hello') first.
>
> This would make the problem explicit.
> *and* they could even than add an assert to make sure that the mapping does not degenerate.
>
> I cleaned ServerDirectory a little:
> http://code.google.com/p/pharo/issues/detail?id=2408
>
> So there is just Encoder left. Which will disappear in 1.2 anyway.
>
> So yes, we could do that.
>
> Marcus
>
> --
> Marcus Denker  -- http://www.marcusdenker.de
> INRIA Lille -- Nord Europe. Team RMoD.
>
>
> _______________________________________________
> Pharo-project mailing list
> [hidden email]
> http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project