1.5.0

• Allow for linking with an external BLAS library by using the HAS_BLAS compiler directive.
• Some minor updates to the CI and Readme.

1.4.1

Minor refactor (remove the original diff function hack and replaced with epsilon() test)

1.4.0

• Incorporated some updates based on the SciPy version:
  • check constraint violation when exiting due to too many resets
  • check for NaNs
  • do not exit with convergence on steps where SQP is inconsistent
  • allow for specifying unbounded variables via an NaN or a specified value (infinite_bound)

For reference, see the SciPy commit history here: https://github.com/scipy/scipy/commits/4d7413563cfa477c08717daa61e5994e968ee79f/scipy/optimize/slsqp

1.3.1

• Fixed a bug in the convergence check (see #36)
• Fixed the logo in the ford-generated documentation
• Fixed the CI

1.3.0

• internal code refactoring
• added option to use the bvls routine to solve the NNLS problem.
• added max iteration optional input for NNLS problem
• added ability to specify real kind with preprocessor directive (e.g, REAL32, REAL64, AND REAL128)
• removed old build script

1.2.3

Full Changelog

• Minor updates to Readme and FPM config file.

1.2.2

Full Changelog

• Added support for the Fortran Package Manager
• Updated CI to GitHub Actions
• Fixed a bug in the finite difference code when m=0

1.2.1

• Minor updates to FORD documentation.

1.2.0

• Added some new stopping conditions.
• Some code refactoring.
• Added a VS Code workspace file.

1.1.0

• New mode to approximate the gradients using finite differences.
• Added a FoBiS config file for building the library and examples.
• Some minor updates to the documentation.