Neoteroi
Investigate the possibility of removing python-dateutil
Yeah,python-dateutil can provide useful functions,but it is enough essential? because it depends a old package 'six',a python version2 package. a pure asgi framework(needs python version 3.5+) depends a python version 2 package, that is funny.
It should be possible, such as dateparser.
But in fact, six is just a compatibility layer between python2 and python3, not a pure py2 library, and python-dateutil may still be used by py2 users? I don't know.
Considering that dateutil is only used in blacksheep.server.bindings it looks like a very good idea to remove this dependency, for the sake of reducing the number of dependencies. After all, I only wanted to improve the user experience when parsing input parameters that should be dates. This is a good recommendation.
But I don't understand the notion of "depends on a Python version 2 package" being funny. What is the problem? Just that some package has a classifier such as Programming Language :: Python :: 2 and is still compatible with that version of the programming language?
My GitHub Workflow has a matrix and the framework is tested with the following:
matrix:
python-version: [3.9, "3.10", "3.11", "3.12", "3.13"]
os: [ubuntu-latest, macos-latest, windows-latest]
And I am publishing wheels for those. So, clearly everything works well with Python 3. I bet most of the source code in CPython 3 itself was already present in CPython 2.
Why I suggest reduding the number of dependency package reduding the number of dependency package can make the maintenance of the framework is easier,because of no need to trace more many extenal packages,and not restricted by the extenal packages.when the extenal packages change the api,the framework is forced to follow it.if the extenal packages is discontinued,the framework is affected.