On Thu, Jun 15, 2017 at 1:59 PM, Nicolas Cellier <nicolas.cellier.aka.nice@
gmail.com> wrote:

> maybe addOnce: would be more expressive than addNew:?
>
addedOnce: to remove the expectation that it would return the value (or
self)?
It is mixing the test result and the command together in one selector -
where else do we do that in core Squeak?

>
>
>
> 2017-06-15 22:52 GMT+02:00 Chris Muller <asqueaker at gmail.com>:
>
>> ^ self select: [ : each | seen addNew: each ]
>>
>> ?
>>
>>
>> On Thu, Jun 15, 2017 at 3:50 PM, Chris Cunningham
>> <cunningham.cb at gmail.com> wrote:
>> >
>> >
>> > On Thu, Jun 15, 2017 at 1:22 PM, Tobias Pape <Das.Linux at gmx.de> wrote:
>> >>
>> >>
>> >> > On 15.06.2017, at 21:40, Eliot Miranda <eliot.miranda at gmail.com>
>> wrote:
>> >> >
>> >> > Hi Tobias,
>> >> >
>> >> > On Thu, Jun 15, 2017 at 12:32 PM, Tobias Pape <Das.Linux at gmx.de>
>> wrote:
>> >> >
>> >> > > On 15.06.2017, at 21:22, Chris Cunningham <cunningham.cb at gmail.com
>> >
>> >> > > wrote:
>> >> > >
>> >> > > #testMissingThenAdd:  ?
>> >> > >
>> >> >
>> >> > #isAdding:
>> >> > #ingests:
>> >> >
>> >> > #wasAbsentButNowIsPresent:
>> >> >
>> >> > (just rambling…)
>> >> >
>> >> > I like added:.  Its concise, accurate and reads nicely:
>> >>
>> >>
>> >> Ok, but then maybe #addedNew: ? So as we know more than just an adding
>> >> operation succeeded… :)
>> >
>> >
>> > So it reads:
>> >
>> > ^self select: [:each| seen addedNew: each]
>> >
>> > I like that how that reads.
>> >
>> > #added: wasn't bad - but I could see myself thinking it meant roughly
>> 'was
>> > this value already in the collection'?  Which misses the add part and is
>> > backwards on the result (true if present, false if added).
>> >
>> > -cbc
>> >>
>> >>
>> >> >
>> >> > withoutDuplicates
>> >> >       "Answer a copy of the receiver that preserves order but
>> eliminates
>> >> > any duplicates."
>> >> >       | seen |
>> >> >       seen := Set new: self size.
>> >> >       ^self select: [:each| seen added: each]
>> >> >
>> >> > messagesDo: aBlock
>> >> >       "Evaluate aBlock exactly once with all the message selectors
>> sent
>> >> > by me."
>> >> >
>> >> >       | scanner aSet |
>> >> >       self isQuick ifTrue: [ ^self ].
>> >> >       scanner := InstructionStream on: self.
>> >> >       scanner scanFor: [ :x |
>> >> >               | selector |
>> >> >               (selector := scanner selectorToSendOrSelf) == scanner
>> >> > ifFalse: [
>> >> >                       ((aSet ifNil: [ aSet := IdentitySet new ])
>> added:
>> >> > selector) ifTrue: [
>> >> >                               aBlock value: selector ] ].
>> >> >               false "keep scanning" ]
>> >> >
>> >> >
>> >> >
>> >> > > On Jun 15, 2017 12:19 PM, "Eliot Miranda" <eliot.miranda at gmail.com
>> >
>> >> > > wrote:
>> >> > > Hi Tobias,
>> >> > >
>> >> > > On Thu, Jun 15, 2017 at 12:06 PM, Tobias Pape <Das.Linux at gmx.de>
>> >> > > wrote:
>> >> > >
>> >> > > > On 15.06.2017, at 20:30, Eliot Miranda <eliot.miranda at gmail.com>
>> >> > > > wrote:
>> >> > > >
>> >> > > > Hi Levente, Hi Chris,
>> >> > > >
>> >> > > > On Wed, Jun 14, 2017 at 1:10 PM, Levente Uzonyi
>> >> > > > <leves at caesar.elte.hu> wrote:
>> >> > > > On Wed, 14 Jun 2017, commits at source.squeak.org wrote:
>> >> > > >
>> >> > > > Eliot Miranda uploaded a new version of Collections to project
>> The
>> >> > > > Trunk:
>> >> > > > http://source.squeak.org/trunk/Collections-eem.756.mcz
>> >> > > >
>> >> > > > ==================== Summary ====================
>> >> > > >
>> >> > > > Name: Collections-eem.756
>> >> > > > Author: eem
>> >> > > > Time: 14 June 2017, 11:03:24.917631 am
>> >> > > > UUID: 8d7c03bc-1cdb-44c7-9173-10d50c0dae29
>> >> > > > Ancestors: Collections-eem.755
>> >> > > >
>> >> > > > Add SequenceableCollection>>withoutDuplicates for a more
>> elegant fix
>> >> > > > to MailMessage>>to:
>> >> > > >
>> >> > > > =============== Diff against Collections-eem.755 ===============
>> >> > > >
>> >> > > > Item was added:
>> >> > > > + ----- Method: SequenceableCollection>>withoutDuplicates (in
>> >> > > > category 'copying') -----
>> >> > > > + withoutDuplicates
>> >> > > > +       "Answer a copy of the receiver that preserves order but
>> >> > > > eliminates any duplicates."
>> >> > > > +       | seen |
>> >> > > > +       seen := Set new: self size.
>> >> > > > +       ^self select: [:each|
>> >> > > > +                                 (seen includes: each)
>> >> > > > +                                       ifTrue: [false]
>> >> > > > +                                       ifFalse: [seen add: each.
>> >> > > > true]]!
>> >> > > >
>> >> > > > What a great opportunity to use #addNewElement::
>> >> > > >
>> >> > > >         ^self select: [ :each | seen addNewElement: each ]
>> >> > > >
>> >> > > > I love the functionality but I don't like the selector. It seems
>> to
>> >> > > > imply that one must only add a new element.  So why not call
>> this something
>> >> > > > like addIfAbsent: ?
>> >> > > >
>> >> > > > Here are some suggestions.  Votes?
>> >> > > > - don't change it; stick with addNewElement:
>> >> > > > - addIfAbsent:
>> >> > > > - ifAbsentAdd:
>> >> > > > - ifMissingAdd:
>> >> > > >
>> >> > > > I think I prefer ifAbsentAdd: cuz addIfAbsent: looks too much
>> like a
>> >> > > > potential spelling error, and conflicts with typical ifAbsent:
>> arguments
>> >> > > > supplying exception blocks.  But I could go with ifMissingAdd:
>> because it is
>> >> > > > more distinctive.
>> >> > >
>> >> > > Well, we do have
>> >> > >
>> >> > >         Collection>>addIfNotPresent:
>> >> > >
>> >> > > So why invent a new one?
>> >> > >
>> >> > > Because addIfNotPresent: answers its argument and we need one that
>> >> > > answers whether the element was absent.  So alas addIfNotPresent:
>> is not a
>> >> > > suitable candidate.
>> >> > >
>> >> > >
>> >> > > I think the important thing with #addNewElement: is that it returns
>> >> > > *whether* it added a new element
>> >> > > but then again it breaks the tradition of #add* returning its
>> >> > > argument…
>> >> > >
>> >> > > a very clear and very strange one that would reveal that a boolean
>> is
>> >> > > returned would be #isAbsentAndIfSoAdd: …
>> >> > >
>> >> > > wasAbsentAdding: or ifWasAbsentAdding: would be less cumbersome
>> but I
>> >> > > like something snappier that people will remember.  ifMissingAdd:
>> looks good
>> >> > > because it doesn't conflict with the add*: methods answering their
>> argument,
>> >> > > and the ifMissing implies the answer is true if the element wasn't
>> already
>> >> > > present.
>> >> > >
>> >> > > Stéphane, can you live with ifMissingAdd: ?  Chris?
>> >> > >
>> >> > > Best regards
>> >> > >         -Tobias
>> >> > >
>> >> > > > _,,,^..^,,,_
>> >> > > > best, Eliot
>> >> > >
>> >> > > _,,,^..^,,,_
>> >> > > best, Eliot
>> >> >
>> >> > _,,,^..^,,,_
>> >> > best, Eliot
>> >> >
>> >>
>> >>
>> >
>> >
>> >
>> >
>>
>>
>
>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.squeakfoundation.org/pipermail/squeak-dev/attachments/20170615/66cc2f9c/attachment-0001.html>