Externals Update, main branch (2024.09.02.)

Updated all projects to atlasexternals-2.1.23.

This MR is meant as a replacement for !72934 (closed). @tsulaia, since a few updates are needed to accommodate the new externals tag, it seemed easier to open a new MR rather than to continue with !72934 (closed). 🤔

The changes wrt. atlasexternals-2.1.20 are (atlasexternals@2.1.20...2.1.23):

• Separated the headers of GeoModel, Acts, Gaudi, Geant4 and VecMem into their separate directories;
  • This is needed to make builds on top of a nightly with a different version of any of these externals possible/robust;
  • As @mbandier let me know today, the simulation group currently accuses these sort of changes for what they describe in ATLASSIM-7250. Which I didn't manage to check out yet. But I can't really imagine that these changes should be able to cause other than compilation problems. 🤔
• Updated CheckerGccPlugins: Improvements with/for GCC 14 (thanks to @ssnyder);
• Fixed a configuration issue in the (standalone) XRootD build (thanks to @elmsheus);
• Updated Findtdaq-common.cmake to pick up more of tdaq-common's Python code (thanks to @fwinkl);
• Made fixes to FindGaudi.cmake, to make it work correctly now that Gaudi has its headers in a unique location.

The header separation updates ended up uncovering an issue in our code that's been there since a very long time, and was hidden by a bunch of externals having their headers in the same place in AthenaExternals. The problem came from Gaudi using the CLHEPConfig.cmake file coming with the CLHEP installation, while we use FindCLHEP.cmake in our own code. The two of them interfere with each other in a really weird way. Now, with an update in FindGaudi.cmake, I was also able to remove a long-standing hack from Athena/CMakeLists.txt. So, finally understanding this issue (a bit more...), had some benefits at least.

Compared to all of that, fixing the build setup mistake in Generators/Photospp_i (also unearthed by the "header separation") was fairly easy. 😉