26 comments

[ 0.21 ms ] story [ 69.1 ms ] thread
Nice to finally have TOML support in the stdlib, especially considering pip's dependence on it.
Now if only Azure Functions would support Python versions newer than 3.9!

Come on Microsoft!

> and it will experience repulsive gravity that will push it back to the original region, but then it will experience the pull from the singularity again and will repeat this process forever.

If the observer had a charge, wouldn't this create radio waves, bleeding off energy?

Hasn't radiation from black holes been observed? It's called Hawking radiation I think.
I've been really looking forward to the variadic generics. ML and data code is full of types like "FloatTensor" and "DataFrame" which really don't narrow things down. Hopefully this will make its way into the ecosystem asap.
Those new error messages look divine compared to what we've had up until now! They seem like they might have been inspired by Rust's error message format to me, but I'm not sure. I'm just happy that I won't have to split chained operations onto individual lines anymore just to have any ability to debug them with the regular old stacktrace errors.
(comment deleted)
I thought not all features were available in 3.10 yet
I believe you may be thinking about a different language? Python only releases a feature release yearly. Features which don’t make it into a release like 3.9, 3.10, and 3.11 are usually agreed upon before the feature freeze window to avoid specifically this.
Looking forward to the 1.22x speedup! Been using Nim for a lot of those types of use cases personally.
I sure do hope Arbitrary Literal String Type (https://peps.python.org/pep-0675/) solves the problem they think it does and doesn't drive every typechecker in the world stark-raving mad

    literal_name: LiteralString
    expect_literal_string(f"hello {literal_name}")
    # OK because it is composed from literal strings.
Maybe I'm being too cynical and this is laying the groundwork for py4000 where they will type check at runtime
I was looking forward this release. The performance improvements in this release are insane! I just created a ticket for my team to upgrade all our production Pythons to this release.
Brave going onto a .0 release ;)
Engineers just postponed the upgrade because of this very reason. I agree. We’ll wait for .1 or .2.
I am curious what kind of work are you guys doing that you need a sudden upgrade? Some DS/ML Work?
This release sent me down a rabbit hole (it started with the sidenote on physics), which made me wonder what software https://discuss.python.org is running on. I couldn't find a link to GitHub or elsewhere or any project page which could enable me to host such a discussion board on a server. Does it have a name?
I'm fairly sure it's Discourse[1]. They switched to it quite recently.

[1] https://www.discourse.org

From the HTML source code:

    <meta name="generator" content="Discourse 2.9.0.beta10 - https://github.com/discourse/discourse version 0ffd40867466b0f6b6fbcbc08e01d1a62633d915">