Using of "deprecated" c++-17 class(MacOs)

[haotianmichael]

• [x] I have searched open and closed issues for duplicates
• [x] I made sure that I am not using an old project version (DO: pull Phasar, update git submodules, rebuild the project and check if the bug is still there)

---

Bug description

I try to build phasar. And include/phasar/PhasarLLVM/Pointer/PointsToSetOwner.h and include/phasar/PhasarLLVM/Pointer/LLVMPointsToSet.h use #include <memory_resource>, but when I used brew install llvm@13.0.1 to compile, it emitted errors that filememory_resourcenot found. I replaced this name with experimental/memory_resoruce, but it still emitted that unsynchronized_pool_resource not found. I did some research as follows:

Steps to reproduce

• TestFile.cpp with experimental/memory_resource included use class&operations in namespace pmr etc. respectively.
• Try LLVM 10.0.0/12.0.1/13.0.1 respectively.
• Use -std=c++17
• make sure PATH well

Actual result: class&operations commented in memory_resource still not found.

Expected result: Worked well.

Context (Environment)

• phasar: [324302ae]
• googletest: [e2239ee6]
• json: [4f8fba140]
• WALi-OpenNWA: [ 2bb4aca0]

Operating System:

• [ ] Linux
• [ ] Windows
• [x] macOS

Build Type:

• [x] cmake
• [ ] custom build

Possible solution

I Googled libc++ supported c++17's implementations on memory_resource

• for libc++ And I saw same situation with mine in stackoverflow
• fliesystem not found

But still not familiar well with c++ and libc++, So Please tell me what to do, thanks~

Example files

/usr/local/opt/llvm/include/c++/v1/experimental/memory_resource

[haotianmichael]

Roollbacking to 83b519e6 and rebuilding, everything works well. It seems that libc++'s implementation doesn't support unsynchronized_pool_resource? Any help would be grateful.

[janniclas]

Clang and specifically apple-clang don't support polymorphic memory resources as seen here. Depending on your setup you can try to use the gcc compiler to compile phasar. If you use a MacBook with the new M1 processor then this will be troublesome too, since the gcc support on the M1 is not soo good in my experience. At least I also still struggle to get it running. If you have any progress in this area please let me know.

[janniclas]

Also, feel free to check out this branch. This fixed the issues for me.

[MMory]

@janniclas can we close this issue?