Release 2.5.3

This new (minor) version introduces support for ARM64, improvements for the dynamics simulator and support for multivalued functions, among others.

Legenda:

  • N: new feature
  • A: addition to a feature
  • C: change to the behavior of a feature
  • F: fixed feature
  • R: removed feature

Changes:

  • #34 (N) Support multivalued functions (for Interval Taylor models support)
  • #328 (N) Add atan support for NormedAlgebra
  • #796 (N) Add Foundations module for (topo)logical notions
  • #569 (A) Adding multiple points simulator
  • #685 (A) Support ARM by making it properly fall into C99 rounding
  • #686 (A) Support GCC 12 on all platforms
  • #334 (C) Clean-up function models, notably template parameters
  • #337 (C) Rename partial functions to FunctionPatch
  • #665 (C) Remove Value, and use plain FloatXX as exact object
  • #666 (C) Template Float classes
  • #681 (C) Refactor the logger as a ConcLog logger submodule
  • #688 (C) Modify atan to avoid using long double, to support ARM architecture
  • #702 (C) Refactor the concurrency modules as a BetterThreads concurrency submodule
  • #788 (C) Add explicit suggest(…) for step-size suggestion to integrator classes
  • #698 (C) Use snprintf instead of the deprecated sprintf
  • #198 (C,F) Modify differential inclusion code to conform to other evolvers, fix reach set returned to be rigorous
  • #678 (F) Address warnings using AppleClang, which is now usable on C++20 Concepts code
  • #710 (F) Change use of char to short to address issue that ARM GCC char is unsigned
  • #735 (F) Fix bug in multiplication of Taylor models causing under-approximation of error
  • #741 (F) Fix fault in VectorFieldSimulator by using Approximation instead of Real step size