Post

Thanks for sharing, this is definitely a cool idea and quite promising. I wonder what could be done to make lazy imports automatic, so the end user doesn't even have to make it explicit, and the import is only imported once actually used?

"Changing `import` to be lazy by default is outside of the scope of this PEP. From the discussion on PEP 690 it is clear that this is a fairly contentious idea, although perhaps once we have wide-spread use of lazy imports this can be reconsidered." peps.python.org/pep-0810/#ma...

I am sorry Hugo, in your diff I don't see the "lazy" keyword mentioned in the PEP. Why? How does the magic operate, then?

I used the PYTHON_LAZY_IMPORTS env var to make it easy to compare two runs using hyperfine: $ hyperfine --warmup 10 --runs 20 --export-json out.json \ "./python.exe -m pypistats --help" \ "PYTHON_LAZY_IMPORTS=on ./python.exe -m pypistats --help" That then calculates the "faster than" summary.

Thank you very much for highlighting this! 🙏🏼 (my bad) Looking very much forward to this feature! ❤️

In the past I had achieved HUGE boosts on my Django app from switching almost everything to inline imports. Startup time went from like 10s to 100ms. Didn't know about this PEP, it's gonna be fucking amazing