check components

[loriab]

I've checked and your prepare-2-13-0 branch works as-is against Psi4. This includes Erica's mp2-f12 code, so the G12 is working (Psi4 doesn't build the G12DKH). Only two lines of this PR are mildly important (1 and 4 below).

1. It'd be really good to add the check_required_components(${L2}) line so that find_package(... COMPONENTS ...) can reject installations.
2. I noticed that the F12 ints codes for the G12 class aren't completely populating -- I get g12_ssss_d0;g12_ssss_d1 for AM=4. Brief examination of the macro(process_integrals_class_alt function, and I'm not seeing the problem. It doesn't bother me, and I added a line to the INSTALL.md that codes for this ints class are incomplete.
3. tl;dr I added some variables on the target; they can be handy but not too important. Long: At one point I attached a few variables (version, overall AM, ints codes components) onto the cmake target. I was using them in Psi4 to print a build-time warning about Hessians going to finite difference if no 2nd deriv ints available. That's not an important message, and the target variables might be overkill. Same thing could be done with a libint-config.cmake variable but this is a little more even-handed wrt how the target is built/detected. Example print-out at https://github.com/evaleev/libint/actions/runs/21559038451/job/62120277826?pr=391#step:9:211
4. An unexpected consequence of 1. is that the python detection of Libint2 failed since CXX_ho, C, CXX, etc. aren't cmake components now. 🤞 it passes now.

[evaleev]

Look good to me. Could not spot the issue in process_integrals_class_alt either, if this is a deal-breaker eventually let me know, cmake debugger in clion is fairly usable, actually ...