For the record, a note by Guillaume on a desirable improvement of ocaml-r
Well, something noteworthy: the current implementation frees n values in n^2 quadratic time in the worst case. This is due to the particularities of R API itself, which is designed to call foreign code, not to be called by foreign code (this matters as the GC-related primitives for R's GC are supposed to work on the stack frame of the stub calling C code from the R side). This iss the major technical issue with OCaml-R, as it makes it somewhat difficult to do full-fledged programming in OCaml that holds references to R code. I believe you need an intermediary memory management layer between OCaml's GC and R's GC to have a design compatible with a GC based on Cheneyŝ algorithm, which is the case for OCaml.
_______________________________________________ Ocaml-r-devel mailing list [email protected] https://mail.gna.org/listinfo/ocaml-r-devel
