[Rcpp-devel] ColDatum constructors memory safety

[Dirk Eddelbuettel]

Note to self, no more coding when tired...

On 18 March 2010 at 21:49, Dirk Eddelbuettel wrote:
| So I became a little more ambitious and made the runit file a fuller test of
| ColDatum and RcppFrame (both of the old API):
| 
| test.ColDatum.vector <- function() {
|     src <- 'std::vector<ColDatum> colDatumVector(3);
|             colDatumVector[0].setDoubleValue(1.23);
|             colDatumVector[1].setIntValue(42);
|             colDatumVector[2].setLogicalValue(0);
|             std::vector<std::string> names;
|             names.push_back("A");
|             names.push_back("B");
|             names.push_back("C");
|             RcppFrame fr(names);
|             fr.addRow(colDatumVector);
|             RcppResultSet rs;
|             rs.add("data.frame", fr);
|             return R_NilValue;';
|     funx <- cfunction(signature(), src, Rcpp=TRUE)
|     checkEquals(funx(), NULL, msg = "RcppColDatum.vector")
| }

That had an extra return R_NilValue; in there we didn't want. I fixed that and renamed it:

R> library(RUnit)
R> library(inline)
R> test.RcppFrame <- function() {
+     src <- 'std::vector<ColDatum> colDatumVector(3);
+             colDatumVector[0].setDoubleValue(1.23);
+             colDatumVector[1].setIntValue(42);
+             colDatumVector[2].setLogicalValue(0);
+             std::vector<std::string> names;
+             names.push_back("A");
+             names.push_back("B");
+             names.push_back("C");
+             RcppFrame fr(names);
+             fr.addRow(colDatumVector);
+             RcppResultSet rs;
+             rs.add("data.frame", fr);
+      return rs.getReturnList();';
+     funx <- cfunction(signature(), src, Rcpp=TRUE)
+     dframe <- data.frame(funx()[[1]]) ## needs a data.frame() call on first list elem
+     checkEquals(dframe, data.frame(A=1.23, B=42, C=FALSE), msg = "RcppColDatum.vector")
+ }
R> test.RcppFrame()
Loading required package: Rcpp
[1] TRUE
R> 

Now -- this works fine __with or without your patch__ so at this point, I am
still not sure what error your patch is fixing.  Can you help me out?

Many thanks,  Dirk

-- 
  Registration is open for the 2nd International conference R / Finance 2010
  See http://www.RinFinance.com for details, and see you in Chicago in April!