Charles Oliver Nutter a écrit : > Rémi Forax wrote: > >> If you take a close look to the current API: >> http://download.java.net/jdk7/docs/api/java/dyn/MethodHandles.html >> there are not lot of primitive adapters, i.e adapters that can not be >> written from >> the others. >> > > Are these all functional right now? I could obviously use guardWithTest > for my guarded handles without having custom Java code for it. >
I think guardWithTest is functional but based on spread/collect if there is more than one element. See sun.dyn.MethodHandleImpl so in my opinion, custom Java will be faster. > >> The purpose of primitive adapters is to hep the language developer to fight >> the inherent complexity of dealing with polymorphic signature. >> >> In case of Jython or JRuby (correct me if I'm wrong) but current >> implementation >> massively use method with object as parameter type. >> >> In that case, guards etc can be implemented using only convert, spread >> and collect >> and the JavaMethodHandle. >> >> But, for my pet project, for Groovy or for Neal's closure, you have to >> deal with >> signatures including primitive types. >> In that case, insert/drop, permute and combine/guardWithTest are very >> useful. >> > > Actually we have as much need for the primitive support as Groovy, since > we integrate almost as well with Java libraries and have a need to > eliminate all the intermediate layers from our call protocols to and > from Java code. > > The current collection of primitive MHs already makes me giddy. > Ooups, sorry, I forget that we don't live in a closed world :) > - Charlie > Rémi _______________________________________________ mlvm-dev mailing list mlvm-dev@openjdk.java.net http://mail.openjdk.java.net/mailman/listinfo/mlvm-dev
