New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Support for R-4.2.0 #4763
Support for R-4.2.0 #4763
Conversation
For the most part, most of the CMake works, but R-Interface cannot be complied. I guess this is due to the changes to the UCRT64, and it can be resolved by setting one of the compilations flags somewhat stricter. I'm going to investigate that next.
Instead of asking CMake to find Boost when building R-Interface, I'm now only check if `libboost_nowide` exists, and only links that. This also removes a warning that we had before. I'm not sure, but this might have even solved a silent bug where there was a mixup linking to Boost::nowide. In addition, I made the R-Interface build more verbose.
As I was expecting, some of the changes here are fixing the issue that Koen ran into. While I'm still not 100% sure, I suspect that part of the issue was that when passing the build to the R-Interface, CMake was getting confused a bit on what's what, and was picking up the wrong In this situation, I basically don't trust CMake to pick up the right Boost, and only check if |
- They are still broken, but I've removed some of the old configs, and set it up with Conan in mind. - For some reason, Conan cannot build the `libiconv` on macOS! 🤷🏻♂️ - and on Windows, MERGE_MODULES cannot be found because of different PATHing for VS
5d5da1d
to
4bb64d1
Compare
165d5f2
to
17dc4e8
Compare
For the most part, I expect things to work.
but on Windows, I'm running to some issues compilingR-Interface
. I need to investigate this a bit more, but I feel we need to add a few extra flags for the compiler to be more specific, and I've not found them so far.In addition to upgrading R-4.2.0, this PR includes the following:
gfortran
from statics.jasp-stats.org as wellMINGW
are (will be) replaced byRTOOLS
libboost_nowide
directly instead of relying on CMake to find it.Know Issues,
jaspBase
requirements need building then theinstall-jaspBase.R
will fail, and therefore the build fails. This is due to the strict signing requirement on macOS; and it is most likely going to be resolved by more packages having their build on ARM sort out.jaspBase::installJaspModule
where we sign them using theJASPEngine
right away, here,install-jaspBase.R
cannot do that, and therefore we have an unsigned library which R cannot load.installJaspModules
and make it a standalone script that we can just use ininstall-*.R
files. I think this is a good idea, and will resolve the issue but it might need to wait for now.Things to do,
Boost::nowide
or not.