FORLAB

FORLAB is a Fortran module that provides some functions for scientific computing. It's more like a small toolbox.
FORLAB uses stdlib as an upstream package. FORLAB hopes to be a small scaffolding tool. Compared with stdlib, FORLAB is less formal.

Getting Started (中文文档)

Get the code

git clone https://github.com/fortran-fans/forlab.git
cd forlab

Supported Compilers

The following combinations are tested on the default branch of forlab:

Build with fortran-lang/fpm

Fortran Package Manager (fpm) is a great package manager and build system for Fortran. You can build using provided fpm.toml:

fpm build
fpm test --list
fpm test <test_name, see `fpm.toml` or list>

To use forlab within your fpm project, add the following to fpm.toml file:

[dependencies] # or [dev-dependencies] for tests.
forlab = { git="https://github.com/fortran-fans/forlab.git", branch="forlab-fpm" }

API-Doc

ford API-doc-FORD-file.md  # todo

see forlab-API-doc.

Some examples are prepared in the ./example folder, and you can use fpm to run them.

fpm run --example --list
fpm run --example <demo_name, see `fpm.toml` or list>

More informations

Links

1. keurfonluu/Forlab
   Forlab is mainly developed by Keurfon Luu originally.
2. stdlib
   Fortran standard library.
3. Fortran Generics

Fypp

The original intention of developing the multi-precision library(forlab) is to facilitate the user to switch the program accuracy requirements in a timely manner, which is challenging. We use fypp to build a multi-precision forlab. I have to say that fypp has helped us a lot. I learned that the use of code to generate code is called meta-programming. I also think that metaprogramming has great potential, especially for some low-level polymorphic functions and improving the dynamics of statically compiled languages, which is very helpful.
I hope that fypp will get better and better, and that fortran will natively support meta-programming technology in the future.

The problems we encountered

1. The adaptability of fortran metaprogramming ability is not strong;
2. Modular development module and setting submodule should best be combined effectively to improve development efficiency.
3. We don't want forlab to increase its volume unlimitedly. We hope that it can be used in areas where it can achieve value, such as rapid development of fortran automation applets. So we will keep the forlab lightweight, and update and repair it from time to time.
4. Fpm currently has some problems and pain points when compiling the program (But we are very optimistic about the potential of fpm):
   • Slow compilation speed. (Improvements in this PR: optimize file listing)
   • Cannot manage and distribute fpm packages well now.
5. Fortran Generics: Due to the lack of more complete generics, certain functions such as multiple precision and multiple array dimensions cannot be implemented now.