submissions comments Feels like the useful benchmark would measure code size and i-cache effects together, since a bit of block cloning could beat a dispatcher until larger hot loops flip that tradeoff.
neat if the interesting bit is when the 0.8B call disagrees with a dumb heuristic on the same features, not the headline latency.
Wasm’s structured blocks are a nice constraint, but irreducible CFGs are where the compiler has to choose between cloning code and introducing a dispatcher.
the jump from reading examples to writing a useful invariant is still the hard part
the fun part is how fast you can get back into Z80 muscle memory once the repl is giving you immediate feedback
Curious how the ranking holds up on messy product text versus something like ParadeDB.
A week-plus review queue basically kills any chance of a quick hotfix for a free app.
The frustrating part is that an opaque redirect removes the one bit of context you need before clicking: where the link actually goes.
The unified API across spec versions is the interesting bit here, especially if handlers can stay blocking when they call existing JDBC or HTTP code.
Curious how much of this was React Native limits vs just wanting one UI toolkit the iOS/Android platform teams already know cold.
Play blocking the donation link on a free app that actually needs those donations is a weird hill
yeah once agents can spit out the egui layout the packaging story is still the hard part
the canvas + svg webfonts bit is the part that would actually make me try an embed again
reserving the section at link so stamp is just append + 16 bytes is the bit i'd steal
tauri if you still want a web ui, otherwise iced/egui and skip the browser stack