AcademySoftwareFoundation
Fix NanoVDB CI
This is an attempt to fix NanoVDB CI. For Windows, this boils down to downgrading the CMake to cmake 3.28. I follow an example given on this page to downgrade the CMake version. Perhaps I should also update the documentation here: https://www.openvdb.org/documentation/doxygen/dependencies.html#depKnownIssues.
I tried the fix that @Idclip suggested for MacOS, but the CI is still failing when the homebrew installs the dependencies.
@apradhana thanks for trying to look into this - but are you saying we only support CMake version up to 3.28 with this change? That doesn't sounds appropriate to me, can you link me to any filed issues with the newer versions of CMake and zlib through vcpkg please
@apradhana thanks for trying to look into this - but are you saying we only support CMake version up to 3.28 with this change? That doesn't sounds appropriate to me, can you link me to any filed issues with the newer versions of CMake and zlib through vcpkg please
Hi @Idclip , this is the issue I was looking at: https://gitlab.kitware.com/cmake/cmake/-/issues/25355.
@Idclip : I also got an email from github explaining "Jobs using the macos-11 YAML workflow label should be updated to macos-14, macos-13, macos-12, or macos-latest.". So I'm trying out the macos-14 runner in case it fixes the macos problem.
@apradhana the issue you linked is talking about a regression in CMake 3.27. Your PR seems to be back-porting CMake from 3.29 to 3.28. I'm not sure why your change is related to this.
It also looks like more recent Windows builds have been succeeding so I believe this has been fixed on vcpkg mainline.
Currently all our CI runners are offline (I have no idea why). When they are back up I will investigate this