Re: [Rd] How to locate references to an environment?

[Antoine Fabri]

You could try {constructive} : https://github.com/cynkra/constructive/

Here's how it might help:

```
E <- new.env()
fun <- function(x) x
environment(fun) <- E
O <- list(
  a = 1,
  b = fun
)
# remotes::install_github("cynkra/constructive")
library(constructive)

construct(E)
#> constructive::env("0x1075b2380", parents = "global")

construct(O)
#> list(
#>   a = 1,
#>   b = (function(x) x) |>
#> (`environment<-`)(constructive::env("0x1075b2380", parents =
"global"))
#> )
```

By searching for `0x1075b2380` here you'd easily find where it's referenced.
If you have S4 objects in there you might want to try again in a few days
since they're not supported yet.

If the output is too big consider `construct_dump(list(O = O), "out.R")`

Thanks,

Antoine


> Message: 1
> Date: Tue, 23 May 2023 11:29:19 -0700
> From: Peter Meilstrup 
> To: r-devel 
> Subject: [Rd] How to locate references to an environment?
> Message-ID:
>  q...@mail.gmail.com>
> Content-Type: text/plain; charset="utf-8"
>
> R developers,
>
> I am trying to track down a memory leak in my R package.
>
> I have a complex object O which comprises a lot of closures and such.
> Among which, the object uses an environment E to perform computations
> and keep intermediate values in. When O closes/finishes with its task
> it nulls out its reference to E so that that intermediate data can be
> garbage collected; I've verified that it does null the reference.
>
> However, it seems there is another reference to E floating around. I
> can tell because I can ask O to put a large array in E, then tell O to
> close, which nulls the reference to E, but then if I serialize(O,
> ascii=TRUE) I can still see the array in the output.
>
> Dangling references to E could come from a closure created in E, or an
> unforced promise from a function call evaluated in E that created a
> closure I still have a reference to, or, ... my question is how do I
> locate the reference?
>
> Is there a way to scan the workspace for objects that refer to a given
> object?
>
> Or is there a tool that will unpack/explain serialize()'s .rds format
> in a more human-readable way so that I can tell where the reference to
> E occurs?
>
> Peter
>
>
>

[[alternative HTML version deleted]]

__
R-devel@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-devel

Re: [Rd] Heap access across multiple calls from R to C++

[Dirk Eddelbuettel]

On 24 May 2023 at 09:09, Ivan Krylov wrote:
| On Wed, 24 May 2023 02:08:25 +0200
|  wrote:
| > Is there a better way to do this?
| 
| The .Call() interface (where functions take an arbitrary number of
| native R objects and return a native R object) combined with external
| pointers is likely to be a better approach here:
| 
| 
https://cran.r-project.org/doc/manuals/R-exts.html#External-pointers-and-weak-references

Yes. We had numerous threads here and on r-package-devel (better list for the
question IMHO, maybe follow-up there?) stating that new code really should
avoid .C() and stick with .Call().

| On the R side, the returned SEXP will have class 'externalptr', with
| not much to do about it programmatically. C code can additionally
| register a finalizer on the object in order to release the memory
| automatically after the it is discarded.

(And Rcpp::XPtr sets those for you if you define RCPP_USE_FINALIZE_ON_EXIT.)

A simpler approach is of course available too, and I used it too.  Define
your class as you would in C++.  Then define
 - a helper function to initialize a class object via `new`, assign it to a
   static variable or a global pointer
 - helper functions for setters and getters of the values you need
 - a teardown function you can call, you may even tie it to R's on.exit()

Now _you_ control lifetime and allocation. From R. If anything breaks, you
get to keep the pieces.

We can also help on the rcpp-devel list.

Cheers, Dirk

-- 
dirk.eddelbuettel.com | @eddelbuettel | e...@debian.org

__
R-devel@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-devel