Wrong error message

[rbotafogo]

The following code has a bug... in 'mutate', 'gain_per_hour' uses 'hour' in the formula instead of 'hours'. Using 'hours' gives the right result.

library('dplyr')
library('nycflights13')

flights_sml =
  flights %>%
    select(year: day,
           ends_with("delay"),
           distance,
           air_time)

print(flights_sml)

mu = flights_sml %>%
  mutate(gain = arr_delay - dep_delay,
              hours = air_time / 60,
              gain_per_hour = gain / hour)

print(mu)

and this is the stack trance:

RuntimeError:
java.lang.ArrayIndexOutOfBoundsException: -1 (RInternalError)
from com.oracle.truffle.r.nodes.function.FunctionDefinitionNode.execute(FunctionDefinitionNode.java:325)
from org.graalvm.compiler.truffle.runtime.OptimizedCallTarget.callProxy(OptimizedCallTarget.java:275)
from org.graalvm.compiler.truffle.runtime.OptimizedCallTarget.callRoot(OptimizedCallTarget.java:264)
from org.graalvm.compiler.truffle.runtime.OptimizedCallTarget.callBoundary(OptimizedCallTarget.java:251)
from org.graalvm.compiler.truffle.runtime.OptimizedCallTarget.doInvoke(OptimizedCallTarget.java:233)
from org.graalvm.compiler.truffle.runtime.OptimizedCallTarget.callDirect(OptimizedCallTarget.java:216)
from org.graalvm.compiler.truffle.runtime.OptimizedDirectCallNode.callProxy(OptimizedDirectCallNode.java:84)
from org.graalvm.compiler.truffle.runtime.OptimizedDirectCallNode.call(OptimizedDirectCallNode.java:68)
from com.oracle.truffle.r.nodes.function.call.CallRFunctionNode.execute(CallRFunctionNode.java:59)
from com.oracle.truffle.r.nodes.function.RCallNode$DispatchedCallNode.execute(RCallNode.java:1169)
Caused by:
-1 (ArrayIndexOutOfBoundsException)
from java.util.ArrayList.elementData(ArrayList.java:422)
from java.util.ArrayList.remove(ArrayList.java:499)
from com.oracle.truffle.r.ffi.impl.common.JavaUpCallsRFFIImpl.Rf_unprotect(JavaUpCallsRFFIImpl.java:1572)
from com.oracle.truffle.r.ffi.impl.upcalls.Rf_unprotectCall$Rf_unprotectCallFactory$1.execute(Rf_unprotectCall.java:54)
from org.graalvm.compiler.truffle.runtime.OptimizedCallTarget.callProxy(OptimizedCallTarget.java:275)
from org.graalvm.compiler.truffle.runtime.OptimizedCallTarget.callRoot(OptimizedCallTarget.java:264)
from org.graalvm.compiler.truffle.runtime.OptimizedCallTarget.callBoundary(OptimizedCallTarget.java:251)
from org.graalvm.compiler.truffle.runtime.OptimizedCallTarget.doInvoke(OptimizedCallTarget.java:233)
from org.graalvm.compiler.truffle.runtime.OptimizedCallTarget.callDirect(OptimizedCallTarget.java:216)
from org.graalvm.compiler.truffle.runtime.OptimizedDirectCallNode.callProxy(OptimizedDirectCallNode.java:84)

[steve-s]

Hi Rodrigo,

thank you for the report. We can reproduce the issue and are working on a fix.

[steve-s]

Should be fixed in d8bb267

Edit: sorry to mislead here a bit, the other dplyr issue (#51) is fixed in that commit, but in this case we believe that the problem may be actually on the side of the Rcpp package used by dplyr. We've reported that issue RcppCore/Rcpp#935. In the meantime as a workaround you can patch Rcpp sources:

--- a/inst/include/Rcpp/protection/Shelter.h
+++ b/inst/include/Rcpp/protection/Shelter.h
@@ -26,7 +26,7 @@ namespace Rcpp {
         Shelter() : nprotected(0){}
 
         inline SEXP operator()(SEXP x){
-            nprotected++;
+           if ( x != R_NilValue ) nprotected++;
             return Rcpp_protect(x) ;
         }