A forked process in mcparallel will wipe the (shared) tempdir when it
quits. I think this is almost always undesirable:
file.exists(tempdir())
parallel::mcparallel(q("no"))
file.exists(tempdir()) # false
file.create(tempfile()) # error
A simple solution might be to skip over R_CleanTempDi
Thanks for these explanations - it all makes sense, that is, the
default behavior for a process that does not capture SIGQUIT is to
quit and perform a core dump
(https://en.wikipedia.org/wiki/Unix_signal#SIGQUIT).
Then the remaining question, as Luke says, is: should R handle this
signal? For ins
Good day,
> In general, the device-limits info is not on the help page because we do not
> know it.
It seems like it would be near-impossible to do in a cross-platform way.
> Normally you will get warning(s) accompanying that Error ...
> Warning: unable to allocate bitmap
> Warning: opening dev
Hi,
Posting here as bugzilla is closed to registration.
The zlib version checking code does not handle double digits for the
patch version in the semantic versioning scheme. Consequently, a
./configure fails when using a zlib version ≥ 1.5.10. I suggest
something like the following patch:
--- a/
This has been fixed (https://cran.r-project.org/doc/manuals/r-devel/NEWS.html):
CHANGES IN R 3.3.2 patched:
INSTALLATION on a UNIX-ALIKE
* The configure check for the zlib version is now robust to versions
longer than 5 characters, including 1.2.10.
in SVN r71889 (2017-01-03):
https://github.c
Henrik Bengtsson writes:
> This has been fixed
> (https://cran.r-project.org/doc/manuals/r-devel/NEWS.html):
>
> CHANGES IN R 3.3.2 patched:
>
> INSTALLATION on a UNIX-ALIKE
>
> * The configure check for the zlib version is now robust to versions
> longer than 5 characters, including 1.2.10.
>
