bool operator== (const _CREDIT_LINE &/*src*/)	{ return false; }
		bool operator!= (const _CREDIT_LINE &/*src*/)	{ return true; }

That's the same combination I have on my luggage!

Closing this test issue before it starts mining fiat. 🪙⚡

This is the hot path -> return null.

Closed by #177 which became (ironically) possible thanks to addition of "impossible" values to the language

Regarding the failing test: not sure what causes it. It actually works fine when running it locally - aka. "works on my machine" 💀 I executed both

make test

make test-e2e

PR made by AI. PR reviewed by AI. PR approved by AI

How do we get Prusa to actually do something to fix thier bug, thier CS gaslights everything I bring to them.

I don't want to tell a maintainer "trust me, it's green".

tell me you use a mac without telling me you use a mac

I was going to do a follow-up PR, keeping the changes atomic, or at least atomic-ish.

theres only so much you can do in C to prevent a user from shooting themselves in the foot 😿

if only we had smart

constexpr

Closing in favor of issue #133. Good to know these issues, but let's just iterate with the marmot we have.

so this is "just" min-heap right?

(it's fine I just didn't see tree of losers)

We can't use a loser tree because Feldera is a win-win for everyone

ell... DOOM Ex Machina

More catalogs: More problems?

@copilot also note in instructions that if this particular model is working on a pr, it's in clown mode and likely to be full of problems

This is civilized app, we use trash. Throwing garbage on the floor or through the window isn't currently supported. It will not be on priority list for now.

These tests seem LLM generated with how busy they are. 😉

I would prefer to just one additional test for one of the affected actions to test the behavior.

These tests seem LLM generated with how busy they are. 😉

I would prefer to just one additional test for one of the affected actions to test the behavior.

@copilot agents are SO HORNY to add this in, add very explicit and severe documentation to prevent this and also make them use a clown emoji when they inevitably fail to read it

@copilot why did you add a nuget.exe?

just gonna give this one a cheeky self-merge

drug and drop

Ввиду большого количества багов у отражений в зеркале, зеркала более не отражают

I doesn't crash anymore but returns a null DataFrame instead

ghci> df <- D.readParquet "/home/max/Stats/Parquet/0_stock_moves.parquet" 
ghci> D.dimensions df
(0,0)

You have opened a can of worms. I shall now hve to sort through the worms, and fix.

what if user password is null?

We will not have any screen readers.

If we decide to be compatible with them, we have bigger problems