[rchk] Stack imbalance in RcppExports.cpp

[romainfrancois]

I'm getting many variations of :

Function _dplyr_grouped_df_impl
  [UP] ignoring variable <unnamed var:   %22 = alloca %struct.SEXPREC*, align 8> as it has address taken, results will be incomplete 
  [PB] has possible protection stack imbalance /opt/R-svn/packages/build/6XQsgOpC/dplyr/src/RcppExports.cpp:254

from the rchk results. I believe this is because of the VOID_END_RCPP macro:

#ifndef VOID_END_RCPP
#define VOID_END_RCPP                                                                            \
    }                                                                                            \
    catch( Rcpp::internal::InterruptedException &__ex__) {                                       \
        rcpp_output_type = 1 ;                                                                   \
    }                                                                                            \
    catch (Rcpp::LongjumpException& __ex__) {                                                    \
        rcpp_output_type = 3 ;                                                                   \
        rcpp_output_condition = __ex__.token;                                                    \
    }                                                                                            \
    catch(Rcpp::exception& __ex__) {                                                             \
       rcpp_output_type = 2 ;                                                                    \
       rcpp_output_condition = PROTECT(rcpp_exception_to_r_condition(__ex__)) ;                  \
    }                                                                                            \
    catch( std::exception& __ex__ ){                                                             \
       rcpp_output_type = 2 ;                                                                    \
       rcpp_output_condition = PROTECT(exception_to_r_condition(__ex__)) ;                       \
    }                                                                                            \
    catch( ... ){                                                                                \
       rcpp_output_type = 2 ;                                                                    \
       rcpp_output_condition = PROTECT(string_to_try_error("c++ exception (unknown reason)")) ;  \
    }                                                                                            \
    if( rcpp_output_type == 1 ){                                                                 \
       Rf_onintr() ;                                                                             \
    }                                                                                            \
    if( rcpp_output_type == 2 ){                                                                 \
       SEXP stop_sym  = Rf_install( "stop" ) ;                                                   \
       SEXP expr = PROTECT( Rf_lang2( stop_sym , rcpp_output_condition ) ) ;                     \
       Rf_eval( expr, R_GlobalEnv ) ;                                                            \
    }                                                                                            \
    if (rcpp_output_type == 3) {                                                                 \
        Rcpp::internal::resumeJump(rcpp_output_condition);                                       \
    }
#endif

which PROTECT() but never UNPROTECT()

I understand that there is a stop call, however rchk does not seem to understand that.

ping @lionel-

[lionel-]

I guess it makes sense to unprotect just to help static analysis?

Rf_eval( expr, R_GlobalEnv ) ;

This should also be R_BaseEnv.

[romainfrancois]

That's what I'm thinking too, I'm working on a branch.

[eddelbuettel]

Done in #949. Thanks.