Fortran good first issues

This issue proposes a significant enhancement to the CP2K documentation, specifically the Input Reference. The goal is to improve usability for new users by providing more detailed explanations of keywords, including practical usage, typical use cases, parameter sensitivity, and common pitfalls, rather than relying solely on brief descriptions and external literature. The suggestion is to expand theoretical background and computational methodology sections, cross-linking them with keyword descriptions, and potentially encouraging community contributions.

A bug in the FATES model causes liquid water boundary conditions to be incorrectly set to total soil water during the first timestep of each day. This occurs because the `dynamics_driv()` subroutine overwrites correct liquid water values set in `wrap_btran()`. The issue affects specific plant functional types and certain FATES diagnostics when soil ice is present.

This issue highlights a missing documentation entry for two specific files, `seedname.vdw` and `seedname_band_proj.dat`, within the user guide's file description chapter. The user guide currently references these files but fails to explain their purpose or content.

This issue requests the addition of solutions for specific examples (15 and 23-28) to a tutorial's solution booklet. It's an enhancement to the existing documentation, aiming to provide more complete learning resources for users.

The user observed that the order of specified atomic projections in the nnkpt file is not preserved when listed on a single line, with orbitals being sorted by their angular momentum quantum number (l). This behavior is contrasted with specifying projections on separate lines, which does maintain the order. The user is seeking clarification on whether this is intended behavior and if it's documented, as the ordering is important for certain operations like symmetrization.

The user is reporting a potential error in Figure 14f of the solution booklet. They were able to reproduce figures 14a-e but not 14f, and observed that figures 14e and 14f appear identical in the booklet. They have attached their own reproduction of 14f with specific disentanglement parameters for review.

This issue is a comprehensive collection of requested features and enhancements related to k-point calculations in CP2K. It covers a wide range of functionalities including various DFT+U methods, symmetry reduction, Wannier functions, alternative diagonalization schemes, and post-SCF methods. Many of these features are already implemented, but several are still pending or require further discussion.

A user has successfully compiled CP2K with Intel oneAPI 2025.2.1, but is experiencing significantly slower performance and a memory leak during cell optimization tasks compared to GCC/OpenMPI builds. This issue persists even with an older version of oneAPI, suggesting potential compatibility problems between CP2K and oneAPI, beyond just compilation.

A user is reporting that the exchange-correlation energy is calculated as exactly zero for an r2SCAN + rVV10 functional combination in CP2K version 2026.2. This unexpected result occurs despite the calculation completing successfully, and the user has provided input and output files for debugging.

A latent out-of-bounds read occurs in the `s_compute_viscous_stress_tensor` function when a finite-difference stencil for an immersed-boundary body is centered on a cell adjacent to a domain boundary. This happens because the velocity-gradient stencil indexes coefficient arrays that are only allocated for interior cells, leading to an invalid read when accessing boundary ghost cells.

This issue reports a recurring problem with incorrect runoff mapping between the CTSM (Community Terrestrial System Model) and mizuRoute components in CESM. The problem appears to be exacerbated when lake polygons are included, leading to incorrect spatial patterns and values in the mapped runoff, potentially due to issues with normalization by land fraction during the regridding process.

The user is questioning the implementation of Hirshfeld charge calculations in CP2K, specifically regarding the SHAPE_FUNCTION GAUSSIAN and SHAPE_FUNCTION DENSITY with SELF_CONSISTENT T methods. They believe these methods may not adhere to standard schemes and produce chemically unintuitive charges, seeking clarification and potential options to revert to older definitions.

The `scalapack.pc` file installed by the CP2K toolchain incorrectly specifies `libdir=/lib` and requires system BLAS/LAPACK, causing linking errors when building MPI-enabled versions of CP2K. This leads to the system's BLAS/LAPACK being picked up instead of the toolchain-built OpenBLAS, resulting in undefined reference errors.

This issue highlights a potential inaccuracy in how constrained molecular dynamics (cAIMD) is used to calculate potential of mean force (PMF) in some reported works. The user points out that these works incorrectly interpret the Lagrange multiplier from the SHAKE algorithm as the ensemble-averaged constrained force, neglecting important corrections (Z and G). While the official CP2K example is deemed acceptable for simple distance constraints, the user argues that for more complex collective variables (CVs), these corrections are crucial and their omission can lead to significant errors in derived activation energies and reaction rates.

This issue proposes renaming the `fill_threshold` parameter in EAMxx to `time_sampling_threshold` or a similar alternative. The goal is to avoid potential confusion with `horiz_remap_fill_threshold` and ensure clarity in the codebase.

The `no_create` parameter for GPU macros is implemented for OpenACC but causes an unconditional abort in the OpenMP directive generation. This breaks OpenACC-only builds because the OpenMP code is evaluated even when not needed, leading to a build failure.

This issue proposes to refactor the frequency change tolerance check in `need_check.f90` and `realloc_array`. Currently, the tolerance is hard-coded in two separate locations, leading to potential inconsistencies. The fix involves moving this tolerance value to a shared constants module to ensure it's updated uniformly across both subroutines.

This issue identifies that the `query_testlists` tool incorrectly handles multiple comments for a single test in `testlist_clm.xml`, only displaying the last one. The task is to merge these multiple comments into a single, consolidated comment for each affected test to ensure all relevant information is displayed.

This issue proposes to enhance an existing error message related to changing history file configurations mid-run. The current error message is unhelpful, and the suggested improvement is to inform users that they can avoid this by performing a "branch run" instead. This change aims to improve usability by providing clearer guidance to users encountering this specific error.

The GCC Linux build test is failing due to a cache miss in the `cache-env` step, leading to the `install-dependencies-with-spack` step failing because the `SPACK_STACK_DIR` environment variable is not defined. The user attempted a fix by modifying `gcc.yml` but was unsuccessful, and the test has been temporarily disabled.

This issue proposes refactoring CP2K's Libint wrapper to address code duplication and improve extensibility. The current integration is spread across multiple files, making it challenging to add new electronic-integral APIs or alternative libraries. The discussion aims to explore a more unified interface and its organizational structure.

This issue proposes evaluating the migration of CP2K's internal Gaussian integral implementations to an external library like Libint. The goal is to potentially leverage existing Libint wrappers and improve efficiency or maintainability by consolidating integral calculations.

This issue proposes several small cleanup tasks across five tooling files. The majority involve removing duplicated code, such as identical result dictionaries, getter functions, and emoji maps. One task addresses hardcoded scheduling thresholds, and another identifies a redundant configuration dictionary.

The build system incorrectly assumes a target is fully configured if `CMakeCache.txt` exists, leading to issues when builds are interrupted. This can cause configure steps to be skipped permanently or dependencies to be incorrectly marked as built and installed, resulting in confusing build errors.

The `toolchain/modules` script silently truncates environment variable values containing spaces to their first word. This occurs because the `export` command is executed with an unquoted value, leading to word-splitting and subsequent errors that are hidden by output redirection. This can cause critical build configurations to be incomplete without any visible indication.