Once again, I am so frustrated by Python static typing https://github.com/pallets/click/issues/2558 If any typing expert out there wants to help, that would be wonderful. #Python
WTF is wrong with the Apple ecosystem? I've tried 3 different podcast apps now (Podcasts, @overcastfm and PocketCasts) and it seems like none of them are capable of easily like... opening a link to an RSS feed and turning that into a podcast if the feed is only available on the LAN. It's infuriating that every time I deal with some Apple thing it's something like this.
This weekend I’ve spent more time in the tox rabbit hole than I expected, but check out how I’ve cut the runtime of full tox runs of the attrs test suite by 75%:
Two Ways to Turbo-Charge tox
https://hynek.me/articles/turbo-charge-tox/
I've been tempted to get an Oura ring, but I'm a bit worried that the device would become useless if the company went under and also, more importantly, it seems like they have a proprietary charger that costs $60!
Why don't people build their stuff to work with standard wireless charging pads 😭
wow TIL that as of #tox 4, you can absolutely SLAY the environment build times if your project builds universal wheels (= virtually all pure-#Python ones)
Just add
```
[testenv]
package = wheel
wheel_build_env = .pkg
```
and it won’t build your package for every Python version anew: https://tox.wiki/en/latest/upgrading.html#universal-wheels
For my service-identity which has a very quick test suite but a lot of factor combinations, it’s 3m vs 30s. structlog that has fewer but slower envs is still 2:35 vs 1:47.
Me: "I should get this shirt."
Wife: "... Why?"
Me: "To look extremely cool, obviously."
Nailed it.
(Not shown: the matching one my 5 year old was wearing.)
Calling all #Python library maintainers! 🐍
Python 3.12 is in beta! 🎉
This means no new features are allowed in and it's now time to start testing your code. You might find things in your code to fix, or even better, you might find things to fix in Python itself!
Here's how to do it:
https://dev.to/hugovk/help-test-python-312-beta-1508 #python312 #beta #test #CI #GitHubActions
Note: I will eventually try `fish` or `zsh`, but my schedule and to-do list don't currently permit me to rework all my [dotfiles](https://github.com/pganssle/dotfiles) to work with a new shell.
I've tried `atuin` so far, and I don't like it. I think I can only use it with the full screen history search and somehow despite the fact that it's automatically merging commands from all shells, it still loses commands, sometimes *from the shell I'm currently using*.
Does anyone have a good solution for handling bash history?
What (I think) I want:
- History search for each shell searches the history of that shell first, then all other shells' history.
- I don't want to lose history from windows opened in tmux or when I'm running a bunch of shells in parallel.
What I don't care about:
- Syncing history
- Fuzzy search
What I don't want:
- Mandatory full-screen history search
This post basically describes my approach to understanding computers (and to doing physics):
https://rachelbythebay.com/w/2023/05/30/eng/
- watch / poke / use system
- develop mental model of why it is
- find where that model breaks and you can not predict what the system does correctly
The places where your mental model and reality are wrong are the most interesting and productive things to look at!
Do you use #linux and #keepassxc?
I am wondering if you also see this behavior: https://github.com/keepassxreboot/keepassxc/issues/9438#issuecomment-1564883094
If you don't, what OS and desktop environment are you using?
Next, GitHub will be removing support for Python 2.7 from actions/setup-python in under a month, on 19th June. 🪦
"Breaking changes
"Hello everyone. The Python 2.7.x will be removed from python-versions and it won't be possible to set up Python 2.7.x by setup-python.
"The motivation for the changes
"Python2.7 is not supported since January 1, 2020."
This is the one place where ChatGPT has seemed useful to me, by the way. When I try and use it for anything Python-related, it's worse than useless because I know Python stuff *way* better than ChatGPT, and it's faster for me to just do it myself anyway.
When I'm trying to get a handle on super basic questions, ChatGPT seems to be kind of OK at explaining things in such a way as to unblock me.
Not that I think that there's something we or they could be doing a ton better. I'm sure there are reasons for why it's all confusing and disorienting.
Programmer working at Google. Python core developer and general FOSS contributor. I also post some parenting content.