Skip to content

Releases: rjfarmer/gfort2py

Release v.2.6.1

07 Mar 20:09
Compare
Choose a tag to compare

Add support for for settings strings in derived types

Release v2.6.0

21 Jun 16:55
Compare
Choose a tag to compare

This release primarily updates support for numpy 2.0

Release v2.5.0

29 Feb 22:14
Compare
Choose a tag to compare
  • Fix bug when returning a derived type containing an allocatable array
  • Raise AttributeError instead of IndexError when trying to slice scalar derived type
  • Better error message when passing wrong number of arguments to a procedure
  • Fotran names are now treated as case insensitive
  • Debug output now uses Python's logging module

Release: v2.4.2

04 Feb 17:40
Compare
Choose a tag to compare
  • Add support for new platforms
  • Switch compile function to be mostly keyword only arguments
  • Various packages needed for building have had their minimum version bumped
  • mod data can now be exported as JSON

Release: v2.4.1

02 Dec 14:51
Compare
Choose a tag to compare
  • Add support for callback functions
  • Allows access to bind(C) variables and functions

Release: v2.4.0

16 Nov 11:08
Compare
Choose a tag to compare
  • Adds the ability to load Fortran code from either a string or file. This avoids the user needing to have to compile and build a shared library themselves.
  • Update to support Python 3.12
  • Update to support gfortran 13

Release: v2.3.0

16 Oct 14:53
Compare
Choose a tag to compare
  • Initial support for quad precision variables
  • Switch from pyparsing to cpyparsing for faster loading of the mod files
  • Add caching option of mod files for faster startups
  • Fix bug in optional character dummy arguments
  • Fix bug when we need to re-resolve runtime arguments

Release: v2.2.1

29 Aug 10:56
Compare
Choose a tag to compare

Fix issue with assumed shape character arrays

Release: v2.2.0

10 Jul 14:33
Compare
Choose a tag to compare

Add support for complex parameter arrays

Expand the kinds supported for integers, reals, and complex numbers

Allow functions to return an array as a result

Release: v2.1.0

02 Jul 17:53
Compare
Choose a tag to compare

Fixes bugs in logical parameters

Enables support for runtime-defined explicit arrays where the dimension requires evaluating a unary op i.e dimension(n+1) or dimension((n*2)+1)

Be more careful about using np.double for double-precision variables