大幅改善 Python 的 pip 速度

在 Lobsters 上看到「How we made Python's packaging library 3x faster」這個,作者寫下在去年十一月的時候,與 pip 的 maintainer 一起合作,花了不少力氣改善 pip 的速度: Comparing packaging 25.0 and the main branch on Python 3.

Gea-Suan Lin's BLOG

How uv got so fast: https://nesbitt.io/2025/12/26/how-uv-got-so-fast.html

"uv is fast because of what it doesn’t do, not because of what language it’s written in."

#python #PythonPackaging #uv #pip #rust

How uv got so fast

uv’s speed comes from engineering decisions, not just Rust. Static metadata, dropping legacy formats, and standards that didn’t exist five years ago.

Andrew Nesbitt
#Nix is a great replacement for #Pip and the rest of the packaging around #Python, Actually.

DWP update as PIP award changes due from April - what it means for you

https://fed.brid.gy/r/https://www.mirror.co.uk/money/dwp-update-pip-award-changes-36504641

New payment rates for PIP, DLA and Attendance Allowance starting April 2026

https://fed.brid.gy/r/https://www.mirror.co.uk/money/new-payment-rates-pip-dla-36500796

the tool i built to categorize, attribute, describe, and color highlight the various emails, text msgs, and other digital flotsam in the november epstein document dump is now an open source #python package.

i'm particularly pleased to have claimed the "epstein-files" package name on #pypi.

- text messages site: https://michelcrypt4d4mus.github.io/epstein_text_messages/
- all emails/other files site: https://michelcrypt4d4mus.github.io/epstein_text_messages/all_emails_epstein_files_nov_2025.html
- author attributions for redacted counterparties: https://michelcrypt4d4mus.github.io/epstein_text_messages/file_metadata_epstein_files_nov_2025.json
- code on github: https://github.com/michelcrypt4d4mus/epstein_text_messages
- PyPi package: https://pypi.org/project/epstein-files/

#Epstein #JeffreyEpstein #trump #uspol #pip #opensource #EpsteinFiles #dataengineering

un article de @andrewnez expliquant pourquoi uv est plus rapide que #pip pour installer des dépendances #Python : https://nesbitt.io/2025/12/26/how-uv-got-so-fast.html

Notamment :
- uv s'en tient à un formalisme issue de différentes PEP
- ce que uv ne fait pas (.egg, pip.conf, pyc compilation)
- ce que #rust permet d'accélérer (langage exécuté, parallélisation en threads plutôt que par process)
- les optimisations "métier" faites par uv que pip pourrait intégrer

How uv got so fast

uv’s speed comes from engineering decisions, not just Rust. Static metadata, dropping legacy formats, and standards that didn’t exist five years ago.

Andrew Nesbitt
Getting Started — beets