[interop] Fix build errors with -Dtesting=off

[aaronj0]

This was due to a bug in CppInterOp where option(CPPINTEROP_ENABLE_TESTING "Enables the testing infrastructure." BOOL) was defined twice conditionally. This is incorrect and does not allow users to override the option, leading to testing always being on. Fixed with compiler-research/CppInterOp#506

This also requires ROOT to set the same flag in the interpreter CMake (which did not work earlier due to the underlying bug in InterOp)

cc @bellenot

[github-actions]

Test Results

    18 files      18 suites   4d 10h 40m 28s ⏱️
 2 725 tests  2 722 ✅ 0 💤 3 ❌
47 351 runs  47 348 ✅ 0 💤 3 ❌

For more details on these failures, see this check.

Results for commit 0f2b341.

♻️ This comment has been updated with latest results.

[vgvassilev]

Why do we need this PR I thought everything was already in CppInterOp?

[aaronj0]

Why do we need this PR I thought everything was already in CppInterOp?

Yes the fix from compiler-research/CppInterOp#506 is pulled in by using the later commit, but the testing flags are not correctly propagated and CppInterOp does not configure or check the flags when it is not built standalone. For eg: currently with ROOT you can set both USE_CLING and USE_REPL ON. This is fixed by the patch here which will be opened upstream. The changes in ROOT's interpreter cmake is to set the flags in a better way.

[guitargeek]

LGTM! Thanks for the PR, and it also fixed the usecase of building ROOT with -Dfail-on-missin=ON -Dtesting=OFF -Dbuiltin_gtest=OFF without internet connection, which was broken by the initial interop PR. Building ROOT without internet connection is important for some experiments need to build ROOT on network-isolated nodes (see also #11603). I also tested that it works again with this PR.

@aaronj0, please merge when you're ready.

Comments addressed.

[aaronj0]

The failures look unrelated to this PR, merging