Feature or enhancement
When you build and make install CPython from source, currently the installed bin/ directory contains python3 (symlinked to python3.XX), but it does not contain python. This issue proposes that it should also contain a python symlink.
Pitch
Windows installs (from python.org) do not contain a python3 executable, so there is currently no form that documentation examples can take that will work for all platforms reliably.
#99185 has more discussion of python vs python3 from the Windows perspective.
Venv bin/ dir does contain a python -> python3 symlink.
PEP 394 recommends python over python3 as the canonical way to invoke Python.
The conclusion in #98761 was that python -m ... should be preferred over python3 -m ... in documentation examples.
Of course, most users will not be building and make install ing their own Python. But if our recommendation to redistributors is to provide python -> python3, would it not be best to model that ourselves in make install? Is there any downside?
Feature or enhancement
When you build and
make installCPython from source, currently the installedbin/directory containspython3(symlinked topython3.XX), but it does not containpython. This issue proposes that it should also contain apythonsymlink.Pitch
Windows installs (from python.org) do not contain a
python3executable, so there is currently no form that documentation examples can take that will work for all platforms reliably.#99185 has more discussion of
pythonvspython3from the Windows perspective.Venv
bin/dir does contain apython -> python3symlink.PEP 394 recommends
pythonoverpython3as the canonical way to invoke Python.The conclusion in #98761 was that
python -m ...should be preferred overpython3 -m ...in documentation examples.Of course, most users will not be building and
make installing their own Python. But if our recommendation to redistributors is to providepython -> python3, would it not be best to model that ourselves inmake install? Is there any downside?