yes, we need composition to avoid API explosion (things are already very bad in this respect)
> On 22 Jan 2018, at 10:38, Guillermo Polito <guillermopol...@gmail.com> wrote: > > In most collection libraries I know there is the "take" function coming from > the functional world that does that: > > #(1 2 3 4 5 6 7 8) take: 5 > => #(1 2 3 4 5) > > #(1 2 3 4) take: 5 > => #(1 2 3 4) > > Now, I understand there are two different things in here. > - First is the fact that we would like to apply this iterator to more than > sequenceable collections. I think that the name "take" is order agnostic in > that sense. > - The other thing is that people would like to not calculate the > sub-collection... The problem is that we will end up with ugly combinations > of selectors like > > take:thenDo: > take:thenCollect: > take:thenSelect: > ... > > And that pattern already exists also for: > > select:thenCollect: > select:thenDo: > > collect:as: > select:as: > > ... > > I think I would like to have a more composable kind of iterators/collections. > There are Transducers that make "iterators" kind of command objects that you > can apply on a collection. Personally I don't like the API choice but the > idea behind is nice: > > https://github.com/Pharophile/Transducers > > sum := #+ init: 0. > sum1 := #(1 1 1) reduce: sum. > sum2 := (1 to: 1000) transduce: #odd filter reduce: sum. > > > And there are XTreams that allows one to work on a collection as a stream and > then compose decorators on top. This means however that we need to create a > stream and "close" it to obtain the resulting collection on every > manipulation. > > https://github.com/mkobetic/Xtreams > https://code.google.com/archive/p/xtreams/wikis/Core.wiki > > ((aCollection reading collect: [ ... ]) select: [ ... ]) rest. > > Guille > > On Mon, Jan 22, 2018 at 4:11 AM, Ben Coman <b...@openinworld.com> wrote: > On 22 January 2018 at 00:47, Stephane Ducasse <stepharo.s...@gmail.com> wrote: > > Hi Ben and Clement > > > > I have a collection (a dictionary in my case) and I want to get > > maximum 5 bindings out of it and iterate on them. > > I want keysAndValuesDo: or do: but only up to 5 elements. > > > > aDict atMax: 5 do: [:each | ] > > "atMax" sound a bit like comparison rather than counting. Perhaps > better would be... > aDict atMost: 5 do: [:each | ] "or" > aDict for: 5 do: [:each | ] > > but rather than introduce three or more messages... > aDict atMost: 5 do: [:each| ... ] > aDict atMost: 5 select: [:each| ... ] > aDict atMost: 5 collect: [:each| ... ] > > why not introduce just one method?.... > (aDict any: 5) do: [:each| ... ] > (aDict any: 5) select: [:each| ... ] > (aDict any: 5) collect: [:each| ... ] > > > cheers -ben > > > > So I learned from:to:do: > > > > aCollection atMax: 5 do: [:each | ] > > > > Does it make sense? > > > > Stef > > > > On Sun, Jan 21, 2018 at 1:16 PM, Clément Bera <bera.clem...@gmail.com> > > wrote: > >> I don't think we do. Do you need it on SequenceableCollection or > >> HashedCollection too ? > >> > >> Recently I was trying to iterate over the first N elements of a collection > >> and since there was no #first:do: I used #from:to:do:. I guess you could > >> use > >> that too: > >> > >> aCollection from: 1 to: (aCollection size min: 1000) do: aBlock > >> > >> Which guarantees you iterate at max over 1000 elements. But that API is > >> SequenceableCollection specific. > >> > >> On Sun, Jan 21, 2018 at 11:44 AM, Ben Coman <b...@openinworld.com> wrote: > >>> > >>> On 21 January 2018 at 18:36, Stephane Ducasse <stepharo.s...@gmail.com> > >>> wrote: > >>> > Hi > >>> > > >>> > I would like to iterate at max on a certain amount of elements in a > >>> > collection. > >>> > And I was wondering if we have such iterator. > >>> > >>> I'm not clear what functionality your asking for. Could you present > >>> it as code & result if you assumed the iterator you want was > >>> available? > >>> > >>> cheers -ben > >>> > >> > >> > >> > >> -- > >> Clément Béra > >> Pharo consortium engineer > >> https://clementbera.wordpress.com/ > >> Bâtiment B 40, avenue Halley 59650 Villeneuve d'Ascq > > > > > > > -- > > Guille Polito > Research Engineer > > Centre de Recherche en Informatique, Signal et Automatique de Lille > CRIStAL - UMR 9189 > French National Center for Scientific Research - http://www.cnrs.fr > > Web: http://guillep.github.io > Phone: +33 06 52 70 66 13
