Kind of forgot to close the 3.11 and #nuitka series, it's support for 2 releases already ... https://nuitka.net/posts/python311-full-support.html

Testing... #Nuitka + pywebview + Streamlit

Join my #Nuitka live stream, going over changes in recent releases, and working on new optimization. https://www.youtube.com/watch?v=8NBn8S5hmZc

So, this Thursday I will be streaming again some #Nuitka release party explaining changelog stuff, and doing some development: https://nuitka.net/pages/Streaming.html

@pybonacci @computingnature I ran #PyInstaller on demo2.py to compare the output size to #Nuitka. On macOS, Nuitka's demo2.app is 383 MB, PyInstaller's is 687 MB. On Windows, Nuitka's demo2.exe is 140 MB, PyInstaller's is 267 MB. On Linux, Nuitka's demo2.bin is 185 MB, PyInstaller's is 493 MB. So Nuitka is uniformly better. Nevertheless, I would be interested to know what you do to reduce the size of PyInstaller output, especially if you think those tricks might work with Nuitka, too.

@computingnature For a Qt application in #Python built with #Nuitka, I have found that #PySide6 works well on all platforms. When I tried #PyQt6, the Nuitka build crashed on my M1 MacBook. Fortunately, there are very few differences in syntax between PySide6 and PyQt6. See: https://github.com/JaneliaSciComp/python-dist-demo#demo-2-qt

I have added a #linux example to https://github.com/JaneliaSciComp/python-dist-demo. This project explores how to package a #python script and its dependencies into a single executable, making it easy to share. I have found that #nuitka works well so far on all platforms, with the only downside being rather long compilation times.

At long last, #nuitka with Python3.11 support, and double the improvements as usual due to the longer release cycle. https://nuitka.net/posts/nuitka-release-16.html

So, finally #Nuitka 1.5 with some Python3.11 support (1.6 is going to be complete I hope), and lots of general improvements, getting more scalable again: https://nuitka.net/posts/nuitka-release-15.html

Now we officially support https://github.com/Nuitka/Nuitka-Action which has been adopted a while ago and was originally created by @jimkring use that to build with #nuitka on GitHub.

Even experimental scientific software in #Python should be easy to set up, so I feel uncomfortable telling users to make sure Python is installed, then install #Conda, install some packages, maybe use Pip to install some more, etc. As an alternative, I have been testing the bundling of scripts and dependencies into one executable, downloadable from a GitHub release. Doing it with #Nuitka has been working well so far. For details, see https://github.com/JaneliaSciComp/python-dist-demo

Want to make the #nuitka Python compiler creator happy on his birthday? There you go, donate money for the party with real life Nuitka (my wife) here: https://nuitka.net/pages/donations.html

Python 3.11 support is getting there, #Nuitka 1.5 is going to have it...

@dekkzz76 Yeah, it's unfortunate.

There are better runtimes for it, like #PyPy, but barely anyone uses it. The one actual compiler I know for #Python, #Nuitka, needs some love too.

In ideal circumstances I'd say #Guile should replace it. No GIL, proper multithreading, JIT and it's a #Lisp.

The continuation of the Yaml post series for #nuitka is here, this one is also very instructive and even a thing to learn in there: https://nuitka.net/doc/nuitka-package-config-part2.html

If you love #python you love this #nuitka post series, some of the best content in my mind: https://nuitka.net/posts/nuitka-package-config-kickoff.html

@levynoise it's not C++, but Nuitka does translation to C: https://nuitka.net/

#nuitka #transpiling #python

@levynoise The CoPilot interface in VSCode isn't really set up for translating thousands of lines of code, I don't think.

What's the reason for porting to C/C++, vs a safer language (Rust, Go, nim)?

There's a basic Python to Rust transpiler here: https://pypi.org/project/python-rs/

You can also incrementally change to Rust with PyO3.

Something like Nuitka can do automatic translation to C, while maintaining the original source.

#Python #SourceToSourceTranslation #CoPilot #nuitka #transpiling #C

This release of #nuitka improves performance esp. for 3.10, and addresses regressions in performance ever since python 3.7 on the technical levels. Also adds plenty new features and improves esp. macOS support and more AI packages work now. https://nuitka.net/posts/nuitka-release-13.html

I have been chiseling away at one of the oldest technical debts of #Nuitka for many days now, the composite node structure will become key for enhanced scalability, or so I hope.