Paul Ganssle @pganssle@qoto.org
Programmer working at Google. Python core developer and general FOSS contributor. I also post some parenting content.
Joined Nov 2019
Paul Ganssle
boosted
@pganssle you might have to save your profile after adding the links to the sites
@Natris1979 Ah that did it! Thanks!
Does it take a while for the little verification check marks to show up next to your website on your profile? Can't tell if I did something wrong here, since I added rel="me" links to ganssle.io and blog.ganssle.io, but my profile doesn't have the check marks.
I thought maybe my instance doesn't do it, but I found @true_mxp's profile and that seems to have one via the same process...
Paul Ganssle
boosted
"Horrible edge cases to consider when dealing with music"
https://dustri.org/b/horrible-edge-cases-to-consider-when-dealing-with-music.html #music #names #programming
Paul Ganssle
boosted
Okay, so, I made a thing.
https://www.patreon.com/posts/welcome-to-my-75389720
I do not actually expect anyone to sign up for this in its current state, but the point is that it's up, it's live, and I can now maybe begin the gradual transition from "internet vagrant" to "internet busker".
The main thing this gets you is the ability to create a vague sense of inadequacy and dread that I'm not giving you your money's worth, so, please go ahead and stress me out.
Time to do my #PyConUS2023 proposals.
I think I'm actually going to pitch a time zone talk again this year (I realized that this one: https://ganssle.io/talks/#working-with-timezones-pycon is from before `zoneinfo` got into the standard library…)
Paul Ganssle
boosted
Has anyone ever heard of software that takes an epub, divides it up into chunks and then creates an RSS feed of those chunks dated at regular intervals? That way you could read a book alongside your blogs in your new reader. If not, should I make a thing like this?
I think I read about something similar that turns audiobooks into podcast feeds.
Boosts welcome
Paul Ganssle
boosted
On December 1 at 15:00 UTC, as part of #PyData Global 2022
2022, I am leading a tutorial on #Bayesian Decision Analysis.
Learn more and register here: https://buff.ly/3gDgFLh
PyData Global uses pay-what-you-can pricing, with donations based on location, so it is accessible to all!
@jacob @ambv FWIW I've been using KeepassXC (and before that KeepassX and Keepass) for over a decade. I sync it to all my devices over WAN-only using syncthing, but it would be pretty easy to sync using any other file syncing service. Very happy with it, and I'm also happy with Keepass2Android Offline as well.
Paul Ganssle
boosted
@pganssle @jugmac00 @hynek Officially announced today as well:
https://github.com/actions/setup-python/issues/544#issuecomment-1332535877
@mattjohnsonpint Yeah, Hynek figured it out in a parallel reply: https://mastodon.social/@hynek/109434203897348469
So now the Windows + Python users of northern Mexico can have more accurate time zones today. If they update their Python packages. 😅
@mattjohnsonpint Hmm... Allegedly you can use it with anything in this list: https://github.com/actions/python-versions/blob/main/versions-manifest.json
But it doesn't seem to be working, at least for 2.7 and 3.6: https://github.com/python/tzdata/actions/runs/3585907944/jobs/6034467045
They aren't very clear about how to specify it in this README: https://github.com/actions/setup-python/blob/main/docs/advanced-usage.md#available-versions-of-python-and-pypy
Anyone have a good example of setting up GitHub Actions for old, out-of-support versions of Python (e.g. 2.7, 3.6)?
Paul Ganssle
boosted
Inspired by @mkennedy, and the work I'm doing on profiling for Python data processing jobs, some initial scattered thoughts on how performance differs between web applications and data processing, and why they therefore require different tools.
1. Web sites are latency-focused. Web applications typically require very low latency (milliseconds!) from a _user_ perspective. Throughput matters from website operator perspective, but it's more about cost.
Paul Ganssle
boosted
Just did first pass of fork()-based multiprocessing profiling for Sciagraph (https://sciagraph.com), a profiler for #Python #datascience pipelines.
First test passed, now to polish it up.
Notes:
1. In case you are not aware, #Python's `multiprocessing` on Linux is BROKEN BY DEFAULT (https://pythonspeed.com/articles/python-multiprocessing/).
2. As a result, this code is quite evil.
3. I am so so happy I am writing software in #Rust. Writing a robust profiler of this sort in C++ would've been way beyond my abilities.
Paul Ganssle
boosted
The release candidate for tox 4 - a complete rewrite of the project - is now out; see https://tox.wiki/en/rewrite/changelog.html#v4-0-0rc1-2022-11-29. Please try it because if no show-stoppers are reported, will be a stable release on the 6th of December 2022. I'd hate to break your CI, so test it beforehand. 😀Thanks! https://pypi.org/project/tox/4.0.0rc1/
@btskinn That just seems weird to me. I mainly put all my requirements in `tox`, and occasionally I'll create a virtualenv where I do `pip install -e . && pip install ipython` just to have a little environment where I can play around with my library or whatever.
You can also just, like, not include `.` in your `requirements-dev.txt` file: `pip install -e . && pip install -r requirements-dev.txt`.
Paul Ganssle
boosted
Who wants to randomly chat with me at re:Invent?
Programmer working at Google. Python core developer and general FOSS contributor. I also post some parenting content.
Joined Nov 2019
