I had some success using RTools40 and the corresponding R testing
version. I had to do a lot of MSYS shenanigans and couldn't get static
compilation to work, but the binary package seems to work just fine in R
3.6.2.
It makes me very happy to see that R will get an updated toolchain and a
pro
I have used the same setup as in Section 1.2.5 of the manual and have
set CXX_STD=CXX14 in src/Makevars.win. No I also remember that, when I
sent the package to winbuilder, some time ago, it started compiling (the
compile failed, but at least it tried).
On 2/5/20 11:47 AM, Ralf Stubner wrote:
On 5 February 2020 at 11:47, Ralf Stubner wrote:
| On Tue, Feb 4, 2020 at 5:51 PM Guido Kraemer wrote:
| > I have also set `~/.R/Makevars.win` as suggested in the manual "Writing
| > R extensions", Section 1.2.5 "Using C++14 code".
|
| What exactly do you have in src/Makevars.win? Just the simp
On Tue, Feb 4, 2020 at 5:51 PM Guido Kraemer wrote:
> I have also set `~/.R/Makevars.win` as suggested in the manual "Writing
> R extensions", Section 1.2.5 "Using C++14 code".
What exactly do you have in src/Makevars.win? Just the simple CXX_STD
= CXX14 or the more complex set-up from the end of
On 4 February 2020 at 17:51, Guido Kraemer wrote:
| I am testing building a package on Windows. The package requires C++14.
You will have to wait.
The current Rtools is at g++ 4.9.3 which (barely) supports C++11. We all
keep our fingers crossed as the replacement Rtools would bring g++ 8.*.
I am testing building a package on Windows. The package requires C++14.
I have installed the latest versions of R (3.6.2), RTools (3.5), and
RStudio. The package builds fine under Linux.
I get the following error:
> devtools::install_github("gdkrmr/zarr-R")
...
C++14 standard reque
