You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
jruby: using Graal Truffle Runtime
[truffle] opt fail block in each:./core/kabench.rb:6 <split-0-U> |Reason com.oracle.graal.nodes.util.GraphUtil$2: Found illegal recursive call to HotSpotMethod<Class.getCanonicalName()>, must annotate such calls with @TruffleBoundary!
[truffle] opt fail m:./core/kabench.rb:1 <split-0-U> |Reason com.oracle.graal.nodes.util.GraphUtil$2: This code path should never be part of a compilation. []
[truffle] opt fail Range#each(core):core: Range#each <split-0-U> |Reason com.oracle.graal.nodes.util.GraphUtil$2: This code path should never be part of a compilation. []
I assume that there's just a TruffleBoundary missing somewhere as suggested by the error message. But unfortunately, I'm unable to tell if that's really the case and if yes how to fix it.
Any ideas?
Regards,
Fabio
The text was updated successfully, but these errors were encountered:
Our current implementation of keyword arguments doesn't compile - it only works in the interpreter. You're supposed to get a nice error message telling you this, due to the notDesignedForCompilation directives (e.g.
) but for some reason the compiler here is failing before it gets a chance to report that error. That happens sometimes.
If you change m to have a keyword-rest argument, as in m(**i), you do get the nice compiler error.
The reason the keyword arguments don't compile is that we've taken shortcuts and used slow operations such as verySlowToKeyValues and Java's equals rather than setting up proper Ruby call sites yet.
Hi folks,
If you run the following code with JRuby+Truffle:
you'll get something similar to this:
I assume that there's just a TruffleBoundary missing somewhere as suggested by the error message. But unfortunately, I'm unable to tell if that's really the case and if yes how to fix it.
Any ideas?
Regards,
Fabio
The text was updated successfully, but these errors were encountered: