Channels
#random
- i
irus
04/29/2016, 11:27 AMI planned to use asciidoctor(or markdown) as article language - k
kirillrakhman
04/29/2016, 11:27 AM@martyglaubitz: a kotlin server would still need html/css/js to produce any useful website - k
kirillrakhman
04/29/2016, 11:28 AMthe actual problem is probably the inline php all over the html - m
martyglaubitz
04/29/2016, 11:29 AM@kirillrakhman: yes and that is actually my biggest problem with PHP: application code embedded in markup - c
cbruegg
04/29/2016, 11:30 AM@martyglaubitz: It's fine for templating and web frameworks for C# and Java allow the same. - m
martyglaubitz
04/29/2016, 11:32 AM@cbruegg: yeah but there it at least feels somewhat ugly (at least that's my experience with JSP) in PHP it felt like the most natural thing in the world - c
cbruegg
04/29/2016, 11:33 AM@martyglaubitz: That sounds like an advantage for PHP. Just know how to use your tools simple smile - m
martyglaubitz
04/29/2016, 11:36 AMsomehow i think there should be a language + dedicated runtime which does nothing but renders data to markup. which in turn is then embedded into JVM or .NET webapps to keep app logic clean whilst keeping .jsp & .aspx at bay - k
kirillrakhman
04/29/2016, 1:46 PMam I right in assuming thatCopy code
doesn't produce a single allocation?val a: String? = null val b = a?.isBlank() == true - d
dmitry.petrov
04/29/2016, 1:49 PMNo, since
can be null.a?.isBlank()Copy code
We might be able to avoid that boxing. Do you find it more natural to writeval a: String? = null val b = a?.isBlank() ?: false // no allocation
instead ofnullableBoolean == true
?nullableBoolean ?: false - k
kirillrakhman
04/29/2016, 1:50 PMI don't understand, what exactly can be allocated in my example? Regarding your question, yes, that's my preferred style - m
michaelrocks
04/29/2016, 1:51 PMFrom what I observed such constructions do not produce boxing related allocations. - k
kirillrakhman
04/29/2016, 1:51 PMthe bool will be boxed, but
always returns a cached instanceBoolean.valueOf - m
michaelrocks
04/29/2016, 1:52 PMOh, sorry. I meant the code @dmitry.petrov posted. - k
kirillrakhman
04/29/2016, 1:53 PMI'm surprised that
produces different bytecode than== true?: false - c
cedric
08/11/2016, 11:23 AMI mean at some point, there is diminishing return in obfuscating the code just for nice error messages - c
cedric
08/11/2016, 11:23 AM(which hardly ever happen) - c
christophsturm
08/11/2016, 11:23 AMthat assertion asserts much more - n
natpryce
08/11/2016, 11:23 AMCopy codeassertThat({ throw ExecutionException(RuntimeException("cause")) }, throws<ExecutionException>(has(Throwable::message, contains("cause"))))))) - c
christophsturm
08/11/2016, 11:24 AMfwiw hamkrest > assertj-core - n
natpryce
08/11/2016, 11:24 AMthat would be the same as the assert4j exception - c
christophsturm
08/11/2016, 11:25 AMassertj has many specialized matchers - n
natpryce
08/11/2016, 11:26 AMThe assert4j example is also checking a different thing … that an exception object’s message contains some text, not that a call to some code throws an exception object that meets some criteria - c
christophsturm
08/11/2016, 11:29 AMfor me the main problem here is the error message that the kotlin compiler throws - n
natpryce
08/11/2016, 11:33 AMYes. The Rust team have done some wonderful things with compiler error messages recently - n
natpryce
08/11/2016, 11:34 AMtype inference is notorious for resulting in cryptic errors - n
natpryce
08/11/2016, 11:35 AM - n
natpryce
08/11/2016, 11:35 AMWould be nice to see Kotlin do something similar - m
mbickel
08/11/2016, 11:37 AM@cedric To continue from #C0922A726, Windows for development always had this "second class" feel to me. It was engineered towards end-users/consumers and it showed everywhere. - c
christophsturm
08/11/2016, 11:37 AMpretty nice