this post was submitted on 19 Jun 2025
366 points (100.0% liked)
Programmer Humor
24829 readers
808 users here now
Welcome to Programmer Humor!
This is a place where you can post jokes, memes, humor, etc. related to programming!
For sharing awful code theres also Programming Horror.
Rules
- Keep content in english
- No advertisements
- Posts must be related to programming or programmer topics
founded 2 years ago
MODERATORS
you are viewing a single comment's thread
view the rest of the comments
view the rest of the comments
python:
a: str = 1
And then assign an int to a string just to mess with the interpreter.
only the linter gives a hoot - the interpreter will happily leave that footgun for later
Dude, even just a "FY,I, you sure about this?" would be nice. I gladly embrace python's by-all-means-shotgun-your-leg-off philosophy, but the noobs could use the help.
the problem is that the language doesn't and can't support one single way to use type annotations without changing fundamental functionality. you can absolutely hook up mypy to your editor for newbies, but once you get on the intermediate level, fighting with mypy takes more code than actually solving the problem.
also there was that proposed update to mypy that was put on held when it turned out that the maintainers didn't know how annotations are used in the wild.
Oh I'm well aware. Took me a solid year to appreciate type annotations for what they are and yeah I'm happy using what we have in stdlib now and not messing with mypy tyvm. The problem is that history is lost to newcomers who have very different expectations. Modern IDE's mostly solve it though, so for all my Java peeps dipping their toes into the snake waters, listen to your ide
i mean, i'm all for rejiggering the internals. i've personally written at least two libraries that uses type annotations in reverse to force arguments into the correct type, and i feel like that should probably be a separate mechanism to "just call the annotation"
dataclasses do this for you at the class level. They enforce type annotations at instantiation.
https://docs.python.org/3/library/dataclasses.html
i think we're talking about different things. you use enforce to mean "validate", i used it to mean "coerce". one of the cases was a command line argument parser that consisted of a single decorator, so you could write
and call it with
$ myfile.py foo --bar 3 --baz 2.2and it would print 13.4another was about creating working protocol buffers from an excel sheet, nested types and enums and oneofs and everything. we used it to parameterize tests of our bluetooth protocol.