Open barracuda156 opened 2 months ago
Thank you. I changed the Makevars
file and will run the necessary tests (local, rhub, CRAN). An update should be ready within a week or so. Let me know if it solves your problem then.
Thank you for addressing this. I think f0209d9a158fd097629d3ee7a7bdb829a89e0c67 should work fine (reasonable package manager is supposed to prefer its own prefix, so this should use correct headers and libs by default).
Update 1.2.7 has been approved by CRAN and can be installed via R's install.packages()
function, which (for now) builds it from source. Pre-compiled binaries should be ready within the next few days (you can check at https://cran.r-project.org/web/checks/check_results_PoissonBinomial.html). Please let me know if your problem is fixed.
Update 1.2.7 has been approved by CRAN and can be installed via R's
install.packages()
function, which (for now) builds it from source. Pre-compiled binaries should be ready within the next few days (you can check at https://cran.r-project.org/web/checks/check_results_PoissonBinomial.html). Please let me know if your problem is fixed.
I will update the port once the new release is available either on your GH or CRAN. https://ports.macports.org/port/R-PoissonBinomial/details (For now I added a patch to fix prefix in Makevars, but this was indeed excessive, I just wanted to maintain code style. Now that can be dropped.)
I see this happening when I build this in MacPorts:
This happens because
Makevars
hardcode/usr/local
. This is a wrong behavior prone to producing breakages (consider that for some reason I actually had a second copy oflibfftw3
in/usr/local
). While MacPorts uses/opt/local
, other package managers have different defaults. It should be left to a user or picked from environment.