Channels
#stdlib
- v
voddan
01/11/2016, 4:44 AM@orangy: question: where on github i can see the files generated by Dokka for stdlib? - y
yole
01/11/2016, 8:48 AMnowhere. why would we put then on Github? - y
yole
01/11/2016, 8:49 AMwe have a TeamCity build configuration that builds the stdlib docs: https://teamcity.jetbrains.com/viewType.html?buildTypeId=Kotlin_StandardLibraryDocumentation - v
voddan
01/11/2016, 9:40 AMthanks! - v
voddan
01/11/2016, 9:42 AMi think that would be a good idea to publish it in html or pdf so that devs could use it offline - e
Eugenio
01/12/2016, 1:01 PMWould you find useful a
property delegate in the sdlib? I can make a PRThreadLocal - m
mplatvoet
01/12/2016, 1:09 PMI’m a bit dual about ThreadLocal. The property delegate is easily created but any (required) cleanup is easily forgotten - m
mplatvoet
01/12/2016, 1:11 PMthus adding something like
delegates to the stdlib also introduces the risk of being used while not fully understood, hence leading to potential memory leaksThreadLocal - e
Eugenio
01/12/2016, 1:45 PMfair point - h
h0tk3y
01/12/2016, 11:06 PMHi everyone. I've posted an answer on SO recently (attached) and now I'm wondering why
and its variations forArray(n) { ... }
,IntArray
,DoubleArray
are named in upper case while these are just functions? Why notBooleanArray
,array(n) { ... }
,intArray
etc.? Does anyone has an explanation for this decision?booleanArray - h
h0tk3y
01/12/2016, 11:09 PMBy the way, as there are
andarrayOf
, lowercase names would be consistent withsomePrimitiveArrayOf
/sequence
andsequenceOf
/lazy
.lazyOf - o
orangy
01/13/2016, 8:17 AMThey used to be constructors of appropriate types. - d
damian
01/19/2016, 4:17 PMWhat is the reason for
not being inlined? Is it just the length of the function?List.binarySearch - a
abreslav
01/19/2016, 4:21 PM@abreslav has left the channel - i
ilya.gorbunov
01/20/2016, 2:30 PM@damian: What overload do you asking about? This one
?fun <T> List<T>.binarySearch(fromIndex: Int = 0, toIndex: Int = size(), comparison: (T) -> Int): Int - i
ilya.gorbunov
01/20/2016, 2:30 PMYes, in that case the length was the reason. - d
damian
01/20/2016, 2:34 PM@ilya.gorbunov: Ah, alright. That's good to know. It's too bad; I find myself using binary search in some performance critical situations. I just ended up writing an inlined version though ¯\_(ツ)_/¯ - v
voddan
01/21/2016, 12:54 PM@ilya.gorbunov: I just checked:
results intosetOf(1, 2, 3).toString()
, same as[1, 2, 3]
. Is it intentional? Because before that sets were printedlistOf(1, 2, 3)
(as everywhere else, ei Python){1, 2, 3} - i
ilya.gorbunov
01/21/2016, 12:56 PM@voddan: This is how
implemented for sets in JDK.toString - v
voddan
01/21/2016, 1:45 PM@ilya.gorbunov: I strongly feel that the current way of representing sets is not the right one. Because I just misinterpreted the debug output of a program. Is there a way to fix that? (those classes are mapped anyway) - y
yole
01/21/2016, 1:47 PMmapping means that Kotlin substitutes a different interface for the Java implementation - y
yole
01/21/2016, 1:47 PMso it allows us to change the interface, but not the implementation - v
voddan
01/21/2016, 1:49 PMoh, so it would be possible to change the behaviour for
, not forSetHashSet - v
voddan
01/21/2016, 1:49 PMpity - y
yole
01/21/2016, 2:00 PMno, we cannot change any behavior. only the API. - j
jasoet-gdp
01/21/2016, 11:46 PMWhy
in Kotlin, doesn’t behave likeNothing
in scala?Nothing - j
jasoet-gdp
01/21/2016, 11:46 PMNothing is a subtype of every other type (including scala.Null); there exist no instances of this type
- n
natpryce
01/22/2016, 9:07 AMWhat's the difference? - j
jasoet-gdp
01/22/2016, 9:20 AMNothing in scala is as subtype of every other type - j
jasoet-gdp
01/22/2016, 9:20 AMso if i have method that still pending