Not an issue but Gratification From A User

[thearchivalone]

Hey there. I just wanted to say thank you for creating and actively maintaining this library and port work. When I found it, I was working on building a custom Python-based build system for my current reverse engineering project due to there not being anything close to what I wanted. I'm very much a simple developer and have a basic system but it works like I wanted.

There were a few main requirements and your library has been integral to my needs. I found you through these steps and needs:

1. Portable, self contained Python build that could be used on platforms beyond Windows, Mac and Linux: https://github.com/bjia56/portable-python
2. The python build didn't have a Pathlib implementation built-in, so I searched for a compatibility framework or library
3. Found yours pretty quickly

• NOTE: I haven't tested your library beyond the 3.11 branch yet, but at least 3.11.13 works phenomenally as a drop in replacement

[bitranox]

Hey, thanks a lot for your message, that really means a lot to hear! 😊
I’m glad the library has been helpful for your setup, especially since you’re working with a portable Python build, which can be a tricky environment to support.

I have to admit, I haven’t touched this code in a while, since I moved my own projects to Python 3.12+, and I no longer needed the compatibility workarounds for older versions of pathlib. Do you happen to use any of my extra utilities, like append_suffix or the shutil wrappers?

Also: one thing I wanted to mention - I’ve noticed that with pipx, distributing or running standalone Python-based applications has gotten easier. pipx lets you install application-style packages into isolated environments but expose their executables globally (without polluting your main environment)

For your portable-Python scenarios, that might simplify distributing or bootstrapping the tools your build system relies on.

If you ever want me to update parts of the library (to better support newer Python versions, new features, etc.), or just review changes you make, I’d be more than happy to help. Thanks again for reaching out — it’s always encouraging to hear the work is actually useful! 🙌

[thearchivalone]

I am still fairly new to the Python ecosystem and knew nothing of pipx but am very familiar with its inspirations. A quick cursory glance (and correct me if I'm wrong): it seems to rely on system python to be installed to work out of the box but also may actually be something I implement if it can work within my current structure. Using the portable python version I'm running was mostly due to how many platforms it actively supported and how much of it automatically stayed within the directory it was installed to. It omitted some features (especially venv) but also made sense for why they were omitted.

Do you know if pipx only contains the python runtimes inside of its virtual environments or if one could contain native c code within it too if called and built by Python? That's a big consideration I have that I'm looking for options on. Keeping everything contained without implementing containers (like Docker) is appealing, especially when most of my core audience are not tech people and would not touch Docker or WSL for any reason.

[bitranox]

here a small pipx overview : https://pipx.pypa.io/stable/comparisons/
for advanced usecases You might consider poetry or conda :
https://python-poetry.org/
https://docs.conda.io/projects/conda/en/stable/user-guide/getting-started.html

of course I dont know Your usecase, but it feels a bit like pissing against the wind. i would rather invest my time to create an installscript to install docker, and then run my programs in a defined 3.13/3.14 environment - that saves a lot of headache for you. Alone the forking/spawning differences between windows and linux, and the upcoming versions without GIL ....

[thearchivalone]

Yeah, pipx wouldn’t work. It and PDM (an alternative that has a similar purpose but is a modified npm style system) rely on venv.

Yeah, Docker / containerization would work wonders for me but none of my core audience. Using git is too much for non developers as is and WSL for docker is a no go for most of them. 😂

[bitranox]

I strongly disagree ;-). You can make an installscript that hides all the complexity from the user - installing WSL2 and/or docker it can be really automated. However - if that can not be done, I guess conda will solve all Your problems - it can package anything. Python, Rust, whatever ....
And what has docker to do with git ? You host Your docker containers at the dockerhub, and the users pull the container, thats it. Very easy. Maybe I miss something out ;-)

[thearchivalone]

Either way, getting way off original topic. Thanks again.

For features: I haven’t ran into anything particular that was a major issue with palthlib3x. There are some slight off spec parts compared to what’s in the official implementation but the comments in the code are excellent and those parts are not required or used by me.