@Stewori Stewori released this Sep 21, 2017 · 18 commits to master since this release

Assets 9

JyNI 2.7-alpha5

With JyNI alpha 5, we proudly present the first version of JyNI that runs on Windows.

The main part of this release was developed within a Google Summer of Code project. We thank Google a lot for this wonderful program! For details about the development process including its numerous demanding issues, take a look at the GSoC blog of this project.

New features

The most notable new features include:

  • JyNI now builds and runs on Windows using a makefile-based toolchain and MSVC for Python 2.7
  • ctypes, NumPy and Tkinter were explicitly asserted to work on Windows equally well as on POSIX (i.e. up to known limitations)
  • improved support of PyFile API by Christian Weilbach's contribution
  • improved management of built-in extensions – datetime, _winreg, msvcrt are now built-in
  • support of the _winreg module allows automatic detection of CPython's dll folder on Windows
  • our new module config_util automatically configures paths for CPython's dynamic library folder and lib-tk folder in a platform-independent fashion
  • Tkinter on Windows – not as trivial as it might seem, Tkinter is enabled to properly auto-detect the location of Tcl, Tk and Tix bundled with a CPython installation
  • mbcs support – JyNI now adds the mbcs encoding to Jython (Windows only)
  • JyNI provides putenv implementation in Jython – os.environ now actually calls the system's underlying putenv function on write access
  • JyNI enables sys.winver and sys.dllhandle fields for Jython on Windows
  • various bugfixes

Requirements and platform support

JyNI v2.7-alpha.5 requires Jython 2.7.1.

JyNI v2.7-alpha.5 was tested with Java 7 and 8 on

Binaries are provided for Linux, OS-X and Windows.

Thank you

Thanks go to

Note on NumPy 13.2

NumPy 13.2 is not supported by JyNI alpha 5 due to issues in Cython 0.27.
Please move on directly to NumPy 13.3, where this issue was fixed.

For details see numpy/numpy#9786 and #20.

Note on NumPy 14.x

The NumPy 14 series does not work with this release of JyNI.
Please stick to NumPy 13.3, which is the latest NumPy release known to work with JyNI. We hope to fix issues with NumPy 14 (and 15) in the next release.

For details see #22.