Showing 1-30 of 140
Oh yeah, I think I see it. Re-remove the now re-added back in stuff due to rebase.
64 should be enough for anyone.
You don't need to do anything. We <del>perform the dark magic</del><ins>squash merge</del>.
doing a final copilot review.. bring on the hallucinations!
@cryst6line we suggest all new contributors also contribute their favorite lemon meringue recipe, that would help us a lot here in understanding the situation.
im removing react code if this doesn't fix hallucinations 🔥
I need to add this instruction to my "code review alter ego" skill: Prompt: "ask a bunch of questions in the review, then answer them yourself so that the PR author is confused what to do" 😄
Yeah I think this will be the next thing I push to implement when I've got a little time, shouldn't be too hard... (completely sarcastic knowing this could be done in an hour or a month)
Fixed —
backstab()rc ? rc : TRUETRUEgonna ban floating point in the monorepo as soon as it's out of consensus
Won't do. I will build my own test case with blackjack and hookers when I get to.
I thought I was using it wrong until I used the Black & White preset and it came out in colour.
These all seem like good changes, but I'm still wishing we had a better way to measure performance than "ehh, looks right!"
I fixed this bug using Google Antigravity. I'll publish my fork shortly.
"hard" comparison of reals ... should use soft compare perhaps?
should've added claude to the commit coauthors
GPT-5.4*
Tried to run some of these windows tests under linux, which isn't really a thing, but this helped (a bit)
As yes, the classic database blunder: sorting by time when latency is on the line!
Auto-approved after review window. Minion execution will begin shortly.
cannot merge, because your commits are unsigned & unverified, please sign they
Done
Could you change the comment here to say "Added burp and hiccup"? It helps to be specific
Is it possible to slap this attribute on the whole class?
No, you don't have to update it either. Just close the spam PR!
Or, we'll report you to GitHub for spamming
Hmm or I could have it be
- All AI comments should be written in limricks
Would be very distinct and actually enjoyable to read.
Much better than my crashout xd
s390x is always my source of trust arch, when something fails there, we've to really blame ourself
I thought the any types were banished from latest refactor?
are we seeing console warnings? I think this one might be hallucinated.