Drop gga_x_herman

Closes #399

See merge request libxc/libxc!601

Drop gga_x_herman: the real functional described by Herman is a meta-GGA, it's real form is unclear, and the real functional seems pathologically unstable.

Draft: Implemented actinide fitted version of PBE by Mitrofanov and coworkers.

Closes #405

See merge request libxc/libxc!571

r2scan hybrids

See merge request libxc/libxc!599

Implemented hyb_gga_xc_relpbe0, a refitted version of PBE0 for actinides by Mitrofanov and coworkers. The implementation matches with the authors' reference input to nEh level precision for N and Ne. Note that the parameter values are more accurate than in the paper, and the functional contains scaling factors that were not reported in the paper!

Resolve "Missing comma at end of line in mgga_xc_b97mv.c"

Closes #437

See merge request libxc/libxc!598

Implement hyb_mgga_xc_gas22, the Google Advanced Science 2022 hybrid meta-GGA functional.

Closes #419

See merge request libxc/libxc!589

Implement hyb_mgga_xc_gas22, the Google Advanced Science 2022 hybrid meta-GGA functional. Note that this implementation uses lda_c_pw_mod parameters, not lda_c_pw parameters like the authors' reference implementation using JAX. The slightly different LDA correlation parameters cause uEh level total energy differences for the N atom; if one uses the lda_c_pw_mod parameters in the Jupyter notebook the agreement becomes essentially exact.

Fix mgga_c_revscan

Closes #429

See merge request libxc/libxc!588

Add XC_FLAGS_NEEDS_TAU to Fortran interface

Closes #432

See merge request libxc/libxc!590

Fix the incorrect implementation of mgga_c_revscan. This implementation is in excellent agreement with the original authors' reference implementation.

Resolve "gga_pow compilation with g++ is failing 2 tests"

Closes #434

See merge request libxc/libxc!596

This reverts commit f213cf3e.

Fix memory allocation in xc-regression.c

See merge request libxc/libxc!591

Port changelog from 6.0.0 and bump version of master branch

See merge request libxc/libxc!586

Fix FindStandardMathLibraryC.cmake reusing variable names

See merge request libxc/libxc!585

The original version of the FindStandardMathLibrary test has been copied from elsewhere, but that version has
been intended to check the need of linking the math lib from C++.
libxc has adapted this test to instead check whether it is needed from C.

The problem here is that during this change, the variable names for the result
of the performed compile tests have been left untouched.
This becomes a problem when building libxc alongside another project that has
copied the same original test but leaves that test to work for C++ (as e.g. Eigen does).
Since CHECK_C_SOURCE_COMPILES does not re-perform the test when its result
variable is already available from cache, this can lead to situations where another
version of FindStandardMathLibrary that works for C++ is performed first (almost
certainly indicating that no additional linker flag is required) and once it's libxc's 
turn, the cached result will be reused. This, however, is nonesense as the cached
result is valid only for C++, but libxc interprets it as being valid for C. Thus, this can
lead to omitting the `-lm` flag wrongfully.

Implement gga_x_ncapr

Closes #421

See merge request libxc/libxc!581

reference implementation. Also use more precise values in gga_x_ncap.

xc-threshold should not be installed

See merge request libxc/libxc!584

Fix Fortran compiler detection in automake

Closes #415

See merge request libxc/libxc!583

Fix limit of TASK (#404)

Closes #404

See merge request libxc/libxc!582

CASE21 journal reference

See merge request libxc/libxc!577

Implement B97-3c

See merge request libxc/libxc!580

Rename erratum references to same convention as everything
else. Cleanup BibTeX entries so that capitalizations in output are
correct.

See merge request libxc/libxc!579