safer Function<>::operator(), Rcpp_list*(), Rcpp_lang*() #1014
Conversation
| } | ||
|
|
||
| */ | ||
| template <typename T1> | ||
| SEXP operator()(const T1& t1) const { | ||
| return Rcpp_fast_eval(Rcpp_lcons(Storage::get__(), pairlist(t1)), R_GlobalEnv); |
There was a problem hiding this comment.
the result of pairlist() was not protected, and Rcpp_lcons() allocates, so it might gc() it.
the result of Rcpp_lcons() is not protected, so it might be claimed when Rcpp_fast_eval() allocates ...
all changes in that file are the same.
| inline SEXP Rcpp_list7( SEXP x0, SEXP x1, SEXP x2, SEXP x3, SEXP x4, SEXP x5, SEXP x6 ) | ||
| { | ||
| Shield<SEXP> out( Rf_cons(x0, Rcpp_list6(x1, x2, x3, x4, x5, x6)) ); |
There was a problem hiding this comment.
Rf_cons() allocates so it might claim x0 and the result of Rcpp_list6().
| @@ -111,6 +111,14 @@ namespace Rcpp{ | |||
| Shield<SEXP> x( Rf_findFun( nameSym, env ) ) ; | |||
| Storage::set__(x) ; | |||
| } | |||
|
|
|||
| SEXP invoke(SEXP args_, SEXP env) const { | |||
There was a problem hiding this comment.
Just a private support function to make operator() easier to write.
|
Thanks for the PR. Next you consider a new PR, please also consider taking another look at the file Contributing. We really prefer issue tickets before PRs. Thank you. |
|
Sure. I opened #915. Edit by DE Removed uncalled-for tone. |
|
LGTM! If I understand correctly, this basically mirrors what R itself does when constructing pairlists as well. Do you think it would be worthwhile to implement a variadic C++11 version that avoids code bloat for these as well? |
|
Definitely. |
eddelbuettel
force-pushed
the
Function_operator
branch
from
930827d
to
3af5556
Compare
|
C.f. RcppCore/rcpp-logs@0e28b87 Rebased to |
comments in context.