pytest
SoftwareWikipedia
11
MENTIONS
7
EPISODES
7
PODCASTS
Search complete. 11 mentions across 7 episodes found for "pytest".
Sep 8, 2026
HackerNoon AINARRATOR
But be honest with your team about what tests pass actually proves, which on current evidence is less than most people assume.
HackerNoon AINARRATOR
METR published an analysis in March 2026, where maintainers from three real open-source projects, Scikit-learn, Sphinx, and Pytest, reviewed two hundred and ninety-six AI-generated pull requests thought that already passed SWE-bench verified, a standard benchmark for coding agents.
HackerNoon AINARRATOR
Roughly half of those test-passing PRs would not have been merged into the real codebase.
HackerNoon AINARRATOR
The maintainers' reasons weren't nitpicks: core functionality failures, code that broke other parts of the system, and straightforward code quality problems the test suite had no way to catch.
Calvin Hendryx-ParkerHOST
Verification matters more.
Calvin Hendryx-ParkerHOST
You see right here, PyTest, 45%.
Calvin Hendryx-ParkerHOST
So we see lots more testing happening, which I think is also a very good sign.
Calvin Hendryx-ParkerHOST
And I'll talk a little bit about that in my second article as well.
Lachlan ReidHOST
tax.
Lachlan ReidHOST
I cannot tell you how many times I've stared at my terminal in the shed watching Claude run PyTest, hit that old 4,000 character wall, write a temp file, and then burn a whole extra turn just going, oh, let me read that file for you.
James TurnerHOST
It was so frustrating.
James TurnerHOST
Every single build error or stack trace over 4,000 characters forced that extra read tool call round trip.
Thomas MachlerGUEST
They will always be there.
Thomas MachlerGUEST
And then here, for example, I have my development environment where I need additional libraries like PyTest and PyTest coverage and whatever.
Thomas MachlerGUEST
And I can then say, okay, this is my default environment and then also add in the test environment or the dependencies that I specify in the test feature.
Thomas MachlerGUEST
And here you see this is the solve group.
Thomas MachlerGUEST
You can just say, I always set log level to debug.
Thomas MachlerGUEST
It's just a random thing I just picked.
Thomas MachlerGUEST
Whenever I'm checking out the the deaf environment, and then I want to define my tasks, which are tied to the deaf environment, and this would be test, where I'm running PyTest with certain arguments, and then the test CI, which I want to run only on my CI runner, and then I want to use different arguments for this, for example.
Thomas MachlerGUEST
And then I can run pixi run eDevTest or eDevTestCI and it will then execute these different command line commands for me.
Herman PoppleberryHOST
Inside that, you'd have markdown files for persistent preferences.
Herman PoppleberryHOST
I use PyTest for testing.
Herman PoppleberryHOST
I prefer type hints everywhere.
Herman PoppleberryHOST
Here are the libraries I reach for by default, that kind of thing.
AmitHOST
He did.
AmitHOST
He explicitly told it to use red-green test-driven development and PyTest.
AmitHOST
That means the agent has to write a failing test first, then write the code to make it pass.
YuvalHOST
That seems like a very deliberate choice to keep the agent grounded.
YuvalHOST
Are those tests running against all three database engines?
speaker_2UNKNOWN
They are.
speaker_2UNKNOWN
He told it to use PyTest fixtures, which is a standard pattern in the Python ecosystem for running the same test against multiple backends.
speaker_2UNKNOWN
Because PyTest is so well established, the agents understand exactly how to implement those fixtures.