R devel - Nov 2017 - check does not check that package examples remove tempdir()

I was looking at the CRAN package 'bfork-0.1.2', which exposes the Unix
fork() and waitpid() calls at the R code level, and noticed that the help
file example for bfork::fork removes R's temporary directory, the value of
tempdir().   I think it happens because the forked process shares the value
of tempdir() with the parent process and removes it when it exits.

This seems like a serious problem - should 'check' make sure that
running
code in a package's examples, vignettes, etc. leaves tempdir() intact?
> dir.exists(tempdir())
[1] TRUE
> library(bfork)
> example(fork)
fork>     ## create a function to be run as a separate process
fork>     fn <- function() {
fork+         Sys.sleep(4)
fork+         print("World!")
fork+     }

fork>     ## fork the process
fork>     pid <- fork(fn)

fork>     ## do work in the parent process
fork>     print("Hello")
[1] "Hello"

fork>     ## wait for the child process
fork>     waitpid(pid)
[1] "World!"
[1] 7063
> dir.exists(tempdir())
[1] FALSE


Bill Dunlap
TIBCO Software
wdunlap tibco.com

	[[alternative HTML version deleted]]

Related to this problem - from R-devel NEWS
(https://cran.r-project.org/doc/manuals/r-devel/NEWS.html):

* tempdir(check = TRUE) recreates the tmpdir() if it is no longer
valid (e.g. because some other process has cleaned up the ?/tmp?
directory).

Not sure if there's a plan to make check = TRUE the default though.

/Henrik

On Wed, Nov 8, 2017 at 4:43 PM, William Dunlap via R-devel
<r-devel at r-project.org> wrote:
> I was looking at the CRAN package 'bfork-0.1.2', which exposes the
Unix
> fork() and waitpid() calls at the R code level, and noticed that the help
> file example for bfork::fork removes R's temporary directory, the value
of
> tempdir().   I think it happens because the forked process shares the value
> of tempdir() with the parent process and removes it when it exits.
>
> This seems like a serious problem - should 'check' make sure that
running
> code in a package's examples, vignettes, etc. leaves tempdir() intact?
>
>> dir.exists(tempdir())
> [1] TRUE
>> library(bfork)
>> example(fork)
>
> fork>     ## create a function to be run as a separate process
> fork>     fn <- function() {
> fork+         Sys.sleep(4)
> fork+         print("World!")
> fork+     }
>
> fork>     ## fork the process
> fork>     pid <- fork(fn)
>
> fork>     ## do work in the parent process
> fork>     print("Hello")
> [1] "Hello"
>
> fork>     ## wait for the child process
> fork>     waitpid(pid)
> [1] "World!"
> [1] 7063
>> dir.exists(tempdir())
> [1] FALSE
>
>
> Bill Dunlap
> TIBCO Software
> wdunlap tibco.com
>
>         [[alternative HTML version deleted]]
>
> ______________________________________________
> R-devel at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-devel

I think recreating tempdir() is ok in an emergency situation, but package
code
should not be removing tempdir() - it may contain important information.

Bill Dunlap
TIBCO Software
wdunlap tibco.com

On Wed, Nov 8, 2017 at 4:55 PM, Henrik Bengtsson <henrik.bengtsson at
gmail.com
> wrote:
> Related to this problem - from R-devel NEWS
> (https://cran.r-project.org/doc/manuals/r-devel/NEWS.html):
>
> * tempdir(check = TRUE) recreates the tmpdir() if it is no longer
> valid (e.g. because some other process has cleaned up the ?/tmp?
> directory).
>
> Not sure if there's a plan to make check = TRUE the default though.
>
> /Henrik
>
> On Wed, Nov 8, 2017 at 4:43 PM, William Dunlap via R-devel
> <r-devel at r-project.org> wrote:
> > I was looking at the CRAN package 'bfork-0.1.2', which exposes
the Unix
> > fork() and waitpid() calls at the R code level, and noticed that the
help
> > file example for bfork::fork removes R's temporary directory, the
value
> of
> > tempdir().   I think it happens because the forked process shares the
> value
> > of tempdir() with the parent process and removes it when it exits.
> >
> > This seems like a serious problem - should 'check' make sure
that running
> > code in a package's examples, vignettes, etc. leaves tempdir()
intact?
> >
> >> dir.exists(tempdir())
> > [1] TRUE
> >> library(bfork)
> >> example(fork)
> >
> > fork>     ## create a function to be run as a separate process
> > fork>     fn <- function() {
> > fork+         Sys.sleep(4)
> > fork+         print("World!")
> > fork+     }
> >
> > fork>     ## fork the process
> > fork>     pid <- fork(fn)
> >
> > fork>     ## do work in the parent process
> > fork>     print("Hello")
> > [1] "Hello"
> >
> > fork>     ## wait for the child process
> > fork>     waitpid(pid)
> > [1] "World!"
> > [1] 7063
> >> dir.exists(tempdir())
> > [1] FALSE
> >
> >
> > Bill Dunlap
> > TIBCO Software
> > wdunlap tibco.com
> >
> >         [[alternative HTML version deleted]]
> >
> > ______________________________________________
> > R-devel at r-project.org mailing list
> > https://stat.ethz.ch/mailman/listinfo/r-devel
>
	[[alternative HTML version deleted]]

On Thu, Nov 9, 2017 at 1:43 AM, William Dunlap via R-devel
<r-devel at r-project.org> wrote:
> I was looking at the CRAN package 'bfork-0.1.2', which exposes the
Unix
> fork() and waitpid() calls at the R code level, and noticed that the help
> file example for bfork::fork removes R's temporary directory, the value
of
> tempdir().   I think it happens because the forked process shares the value
> of tempdir() with the parent process and removes it when it exits.
This has come up a few times:

 - https://stat.ethz.ch/pipermail/r-devel/2017-February/073748.html
 - https://stat.ethz.ch/pipermail/r-devel/2017-April/074149.html

You may want to read up on the latter discussion.