https://kotlinlang.org logo
Channels
Powered by
#stdlib
  • y

    yole

    02/09/2016, 3:27 PM
    I’d do it differently
  • y

    yole

    02/09/2016, 3:27 PM
    Copy code
    fun String.countMatches(sub: String) = (0..length-1).count { substring(it).startsWith(sub) }
  • y

    yole

    02/09/2016, 3:28 PM
    the question is which semantics you want
  • y

    yole

    02/09/2016, 3:28 PM
    what is “aaaa”.countMatches(“aa”)?
  • k

    kirillrakhman

    02/09/2016, 3:28 PM
    what would RegEx' answer be?
  • k

    kirillrakhman

    02/09/2016, 3:28 PM
    3?
  • y

    yole

    02/09/2016, 3:29 PM
    the apache commons javadoc is silent on this
  • c

    cbruegg

    02/09/2016, 8:12 PM
    A sliding window isn't nearly as efficient as Knuth-Morris-Pratt (which is really simple to implement), so you could also consider that.
  • c

    cedric

    02/09/2016, 8:18 PM
    It’s the algorithm that searches from the end?
  • c

    cbruegg

    02/09/2016, 8:25 PM
    No, it's the one that uses a prefix function for the pattern to look for to avoid checking matches already known
  • h

    hannesstruss

    02/10/2016, 2:40 PM
    @hannesstruss has left the channel
  • i

    ilya.gorbunov

    02/11/2016, 4:34 PM
    https://kotlinlang.slack.com/archives/general/p1455207914001721 That would break Map's contract that 
    get(K)
    returns 
    null
    when 
    contains(K)
    returns false.
  • m

    mg6maciej

    02/11/2016, 4:46 PM
    @ilya.gorbunov: I mean 
    get(key: K) = map[key] ?: default.invoke(key)
    and 
    contains(key: K) = get(key) != null
  • m

    mg6maciej

    02/11/2016, 4:49 PM
    So contains returns true when and only when either original map contains the value or default function returns non-null value.
  • m

    mg6maciej

    02/11/2016, 4:49 PM
    Then you don't need 
    MapWithDefault
    interface at all; it's all in 
    Map
    interface.
  • i

    ilya.gorbunov

    02/11/2016, 4:55 PM
    Interesting proposal, though I'm afraid it won't play well with property delegation, when the property has the nullable type.
  • i

    ilya.gorbunov

    02/11/2016, 4:58 PM
    Regarding to https://kotlinlang.slack.com/archives/general/p1455208341001726, if this is a function, wouldn't 
    emptyMap<String, String>().getOrElse { "blabla" }
    fit?
  • m

    mg6maciej

    02/12/2016, 6:34 AM
    @ilya.gorbunov: Not with my current design. I'm storing the map itself in a data structure, but I'll think about changing that to expose a function instead with a default implementation just getting values from the map.
  • n

    natpryce

    02/12/2016, 12:06 PM
    I notice that 
    Iterable<T>
    has methods like 
    mapIndexed
    , 
    filterIndexed
    etc. but not 
    indexed
    that returns an 
    Iterable<Pair<Int,T>>
  • n

    natpryce

    02/12/2016, 12:06 PM
    Would that be a good idea?
  • n

    natpryce

    02/12/2016, 12:07 PM
    E.g. 
    fun Iterable<T>.indexed() = mapIndexed { i, v -> i to v }
  • i

    ilya.gorbunov

    02/12/2016, 12:08 PM
  • n

    natpryce

    02/12/2016, 12:08 PM
    Ah… didn’t think of auto-completing starting with 
    with
    !
  • n

    natpryce

    02/12/2016, 12:08 PM
    Thanks
  • m

    mikehearn

    02/12/2016, 5:20 PM
    Iterable.distinct() says " The elements in the resulting list are in the same order as they were in the source collection."
  • m

    mikehearn

    02/12/2016, 5:20 PM
    However it works by doing toMutableSet().toList()
  • m

    mikehearn

    02/12/2016, 5:20 PM
    doesn't stuffing things into a set make them unordered?
  • i

    ilya.gorbunov

    02/12/2016, 5:22 PM
    Not in case if that set is 
    LinkedHashSet
    .
  • m

    mikehearn

    02/12/2016, 5:24 PM
    ah, i see. isn't that kind of high overhead? i'd have thought toMutableSet() would return a hashset
  • m

    mikehearn

    02/12/2016, 5:25 PM
    did this change when the method was renamed?
1...474849...149Latest