An osu developer found that Windows 27881.1000's libc says that pow(-1, 2) = -1. Poking around in bindiff reveals they tried to add an optimization that skips the main approximation when the base or exponent is 1, but screwed up. Woops!

@jj another instance of "modern CSS being pretty good" i learned about recently: you can specify colours relative to other colours. in this table, the centre square has `background-color: var(--theme)`, and the squares around it have `background-color: oklch(from var(--theme) calc(l + 0.1) calc(c - 0.05) h)`, etc. Oklab generates pretty gradients, being a perceptual colour space

what is going on in EDA land o__O

it's like they're actively trying to make the handwritten version of this as ambiguous as possible

horrible ctf knowledge becoming useful while porting openspades to aarch64: fixed a crash that happened because the script interpreter relied on the x86 representation of pointers-to-member-functions. arm differs from almost every other ABI for C++ because they reserve the lowest address bits!

gnuplot (❤️) one-liner to look for any weird increases in connection attempts due to regreSSHion: journalctl -S 2024-04-01 -u sshd -g "Connection closed" -o short-unix | cut -f 1 -d " " | gnuplot -p -e 'set terminal png size 960,720 small; set xdata time; set format x "%b %d"; binwidth = 60*60*24; set boxwidth binwidth; bin(x,width) = width*floor(x/width); set xtics binwidth rotate out; plot "-" using (bin($1, binwidth)):(1.0) smooth frequency with boxes title "login attempts"'

OpenZFS ASCII art comments, greatest hits volume 1 (according to me)