Comments by o-smirnov

All comments ranked by humor rating

o-smirnovover 1 year agoratt-ru/tigger-lsm

Is it time to start referring to this as Vogon poetry yet?

o-smirnovover 1 year agoratt-ru/tigger-lsm

If there's time, I think we should table software versioning as a meeting agenda item.

And by "agenda item", you mean "struggle session"? Well we can always have a collective moan about it at any given meeting!

I think it's under control again, but my hatred of poetry has only deepened.

Pip: "these upstream dependencies are an unholy mess, but I'm going to make the best plan I can, given the user's current environment."

Poetry: "these upstream dependencies are an unholy mess. I will make myself better by by f*cking the poor unsuspecting user. As a matter of principle, and just because I can."

Is that an accurate take?

o-smirnovabout 1 year agoratt-ru/suricat-beams

Tests? You're the integration test, and a damn good one. 🤣

This doesn't use

dask-ms
and I didn't want to put it into the default
python-astro
image though I'm open to arguments otherwise.

True, I need to tell the f*cker where the data is. And there's

QStandardPaths: error creating runtime directory '/run/user/1000' (Read-only file system)
too.

NOTE:

scabha.cab_utils.yield_output()
has to be moved to
stimela/kitchen/cab.py

Or perhaps killed entirely, let me check...

Typically because we may not know that a release works on a given Python if it isn't in our test matrix. We can always open that up but then we need to routinely add the newer Python versions to the tests.

I'd rather open it up -- we're locking future Pythons out "forever" without necessarily having a good reason for it. Think of your grandchild trying to pip install cult-cargo==0.2.1 under python 3.24 and going dammit grandpa, why did you have to lock down the Python version when you're just shipping a bunch of YAML files...

More seriously, I think a "use at own risk for Python>X" policy is preferable. A warning would be ideal but it seems it's something

pyproject.toml
can't be made to issue, although ChatGPT tells me there's at least a way to formally declare this via classifiers: https://chatgpt.com/share/68ac440c-a0a0-8003-9c61-723befdee50b

It truly doesn't matter, but my OCD suggests the one-liner

zx, zy = map(int, z.split(','))
, just because I always feel sorry asking the computer to do the same work twice! My one small blow against climate change.

Copilot reviewed 8 out of 8 changed files in this pull request and generated no new comments.

Hallelujah, the robot overlord is finally satisfied with my work. @JSKenyon or @SpheMakh over to you for human review.

You're not always as smart as you think, my new robot overlord. This is a regenerated file.

o-smirnov11 months agoratt-ru/radiopadre

Docker build failing, but let me merge and diagnose that.

@tmolteno does

goquartical-summary
fit the needs, and should I roll a cab for it if so?

(You don't need to change cult-cargo to try a 0.1.3 image, of course.)