cmake error when build the source #230
-
|
Dear phoebe team: I try to install phoebe on the centos7.9 when I using The compelete information is below: does not contain a CMakeLists.txt file. CMake Error at lib/CMakeLists.txt:2 (add_subdirectory): does not contain a CMakeLists.txt file. CMake Deprecation Warning at googletest/CMakeLists.txt:4 (cmake_minimum_required): Update the VERSION argument value or use a ... suffix to tell CMake Deprecation Warning at googletest/googlemock/CMakeLists.txt:45 (cmake_minimum_required): Update the VERSION argument value or use a ... suffix to tell CMake Deprecation Warning at googletest/googletest/CMakeLists.txt:56 (cmake_minimum_required): Update the VERSION argument value or use a ... suffix to tell -- Found Python: /opt/hhpc/pub/apps/miniconda3/bin/python3.10 (found version "3.10.10") found components: Interpreter The modul e loaded when build the source as the following: |
Beta Was this translation helpful? Give feedback.
Replies: 2 comments 9 replies
-
|
Hi @zengyoga, I have a feeling you have missed a step in the installation -- if you look here, you can see you probably skipped the "recurse submodules" part of the git clone line: You can fix this in an already established clone by running: Let me know if that doesn't solve the issue, |
Beta Was this translation helpful? Give feedback.
-
|
Yes, I did skip the git clone step because my computer server is on a local area network and can only compile from source. If I using the git clone, I will receive the following message: Are there other ways to solve this problem? |
Beta Was this translation helpful? Give feedback.
-
|
Can you clone it on a local machine, run the submodules line, and then copy the directory up to the cluster? Perhaps that will work. This never used to be an issue, but my understanding is that national lab computers in some countries now have this restriction. In the longer term, perhaps I will find a more permanent solution to this, as submodules are currently a very commonly used git feature. Let me know if it's not working still, |
Beta Was this translation helpful? Give feedback.
-
|
Hi @zengyoga, Thanks for being patient -- I think you have an old copy of a linker/ld/binutils on your system that Phoebe is picking up -- I've seen this before (and if you google the error message, you can see it's not a code problem, but a problem with out-of-date toolchains). I think one way you could get around it, other than seeing if you can update to a newer binutils, is to try removing Best, |
Beta Was this translation helpful? Give feedback.
-
|
Hi. Jenny, Thanks for you suggestion, the error message is still there when I remove |
Beta Was this translation helpful? Give feedback.
-
|
Ok, and one more try to downgrade the -- instead of where Also, when you make this change, be sure you are deleting CMakeCache (or simply, the whole build directory) and rebuilding. Otherwise, you will be building with old CMake files, and the build will not register the changes you make to CMakeLists.txt. Let me know what happens, |
Beta Was this translation helpful? Give feedback.
-
|
Hi. Jenny, The error message is still there, update /usr/bin/ld is so dangerous for me, I'll try install 'phoebe' in a brand new local GPU machine. Thanks a lot for your suggestions. |
Beta Was this translation helpful? Give feedback.
-
|
Hi @zengyoga, It seems like you might be on a cluster, in which case, I meant that you should just have done something like "module load binutils" to get a better version. Another alternative might be to try newer compiler version, or to use something other than GCC. You can see that our test suite passes (the code builds and runs in github actions) with GCC 14. , so I know at least there is not an inherent issue with the code when using up-to-date GCC and linking utilities. Jenny |
Beta Was this translation helpful? Give feedback.
Can you clone it on a local machine, run the submodules line, and then copy the directory up to the cluster? Perhaps that will work.
This never used to be an issue, but my understanding is that national lab computers in some countries now have this restriction. In the longer term, perhaps I will find a more permanent solution to this, as submodules are currently a very commonly used git feature.
Let me know if it's not working still,
Jenny