[core] deprecate compiledata.h

[ferdymercury]

This Pull request:

Changes or fixes:

Header was kind of broken, since header guards were missing.
Luckily it was only used by cxx files once.
Plus naming problems.

In a later step we can privatize it, and later skip this header all the way and go towards a CMake definition target #23291

[pcanal]

Besides the naming problem. I find this file useful for debugging purposes of user installation. Unless we also publish the information in a 'easy/clear' to find way in a move to CMake based, I don't see that move as a beneficial move.

I.e. I would not mind a renamed and improvement but I would not deprecate it (although there is a point to be made that it is an internal details and should not rely upon/used in user code).

[ferdymercury]

Besides the naming problem. I find this file useful for debugging purposes of user installation. Unless we also publish the information in a 'easy/clear' to find way in a move to CMake based, I don't see that move as a beneficial move.

I.e. I would not mind a renamed and improvement but I would not deprecate it (although there is a point to be made that it is an internal details and should not rely upon/used in user code).

This could maybe go into etc/ subfolder as a .txt file, or alternatively into info by RootTargets.cmake

But we could decide that later, after telling people to not include it in their downstream code?

[pcanal]

This could maybe go into etc/ subfolder as a .txt file, or alternatively into info by RootTargets.cmake

With any solution, this file (or its successor) must be the source of truth (as used to build up the ACLiC command line), if it is "only" a duplicata the likelihood that over time it drift to be inaccurate is extremely high.

[github-actions]

Test Results

    22 files      22 suites   3d 18h 47m 3s ⏱️
 3 879 tests  3 877 ✅ 0 💤 2 ❌
76 265 runs  76 263 ✅ 0 💤 2 ❌

For more details on these failures, see this check.

Results for commit 0d5a5ef.