Can I go ahead and land this revision or need to wait for libc++abi group approval?
This is a very small non-functional revision.

LGTM

FYI, the failures reported are unrelated.
They are either timeouts or taking unexpectedly too long (assertion in std/thread/futures/futures.async/async.pass.cpp at line 102).

• Revert "[SystemZ][z/OS] Fix cityhash lit for EBCDIC"

Change to use an array of integers instead of strings.

FYI, the failure ia the timeout on Apple, unrelated to this patch, so consider CI checks as green.

• fix typo

• Fix cityhash lit for EBCDIC
• Switch to array of strings with hard coded values

Addressing comments ...

The above pre-merge checks failures are timeout and not related to this patch. Submitting again to get clean build status.

• Revert 47b49b5f6d40 and 57b747c55d15

• Keep _LIBCPP_EXCEPTION_ABI for classes in experimental namespace.

• Move several exception derived classes to c++abi library
• Remove _LIBCPP_EXCEPTION_ABI from classes those symbols are moved to c++abi.

@fanbo-meng Can you close this review?
I'm going to continue in D138951.

• fix formatting

• trying to fix CI.

Moving several exception derived classes c++abi to address review comments.

@EricWF Thank you for the review.

• correct typo

Addressing comments...

• Remove _LIBCPP_BEGIN_NAMESPACE_EXCEPTION and _LIBCPP_END_NAMESPACE_EXCEPTION

ping

ping @ldionne

LGTM,
ideally we should have macro for __thread which would define to nothing for z/OS.

LGTM

Louis, this has been reviewed by others including yourself and is waiting for your stamp of approval, @ldionne

ping @ldionne

LGTM

Thank you for all reviews which are not addressed.
@ldionne Can you have a look? I just need your stamp of approval.

• consolidate multiple exceptions files into a single file

@ldionne @philnik Please have a look again and let me know if you want to change anything.

• use system notation for #include directive

• switch to use override in place of virtual

• formating changes and make set_target_properties for cxx_exception be consistent with cxx_shared

• config.py should not be part of this change

• Rebase and move recipe to build c++_exception to separate cmake.

Fix CI - Documentation

Fix CI build.

Ready for another round. FYI @ldionne

ping @ldionne

@ldionne I think I addressed all your comments.
Please review it.

Looks like my last commit was skipped. I'm adding it now.

• Add the following options to documentation:

• Make cxx_shared library common for EBCDIC and ASCII build variations.
• Use LIBCXX_CXX_ABI=system-libcxxabi to pick up abi headers and library from previous build.

LGTM

LGTM

Putting rational why I opted out for LIBCXX_CXX_ABI=built-libcxxabi.

• Make cxx_shared library common for EBCDIC and ASCII build variations.
• Introducing built-libcxxabi to pick up abi headers from previous build.
• Adding the following macros:

answering comments and some questions

Addressing Louis's comments to hide z/OS changes as much as possible in cache.

LGTM

LGTM

Louis, thank you for the review and providing some suggestions. However, I'm still not sure how to proceed see my inline comments.

ping, still waiting for the review ...

@ldionne Louis, all comments have been addressed CI is green, can you look at it again? Thx.

LGTM, thank you.

Rebase to pick up 0d7859765260e1dddfdc3490c1fe35537775bb41 which hopefully will fix CL.

The CI is clean now, Can I have the approval or more request, Thx.

• Rebasing and resolving conflicts

No functional changes are introduced since the last patch.

Looking good, I wonder if we can incorporating a small change to set PATH, see inline comment in ibm-zos-libc++-shared.cfg.in.

In D123241#3463208, @zibi wrote:

In D123241#3463076, @zibi wrote:

Have you missed MultiSource/Benchmarks/Prolangs-C/bison/CMakeLists.txt to turn off the warning?

@erichkeane Erich, will you be able to provide the fix for this? Please let me know ETA so I can plan accordingly.

In D123241#3463076, @zibi wrote:

Have you missed MultiSource/Benchmarks/Prolangs-C/bison/CMakeLists.txt to turn off the warning?

I'm seeing this error: