Some of the most common corner cases cross-platform developers for Matlab / GNU Octave run into are over
Rather than wait for 1 / N frustrated users to report such a bug, one can use GNU Octave with popular Windows continuous integration services such as Travis-CI or AppVeyor.
In general, we strongly recommend using Octave ≥ 4.2 as the Matlab compatibility is much better in newer GNU Octave releases.
GNU Octave CI
image: - Visual Studio 2017 - ubuntu1804 build: off init: - cmd: choco install octave.portable - sh: sudo apt-get install -yq --no-install-suggests --no-install-recommends octave > /dev/null test_script: - octave-cli test_myfuncs.m
language: c group: travis_latest addons: apt: packages: - octave matrix: include: - os: linux - os: windows language: sh before_install: - choco install octave.portable - os: osx before_install: - brew install octave > /dev/null script: - octave-cli test_myfuncs.m
For advanced users, consider setting up a Docker image to avoid the repeated GNU Octave installs.
For non-Windows laptops developing for Windows, the typical workaround is to use a Windows virtual machine.