Filippo Valsorda @filippo.abyssdomain.expert · 29/08/2026> the scrapers are evil and stupid (and most of them are!) I get where you're coming from, but I think it makes sense for me to keep sending patches where I'm sending them (e.g. avoiding 10x cost jumps on PQ handshakes), but that doesn't make my observation wrong. 1150 15
Filippo Valsorda @filippo.abyssdomain.expert · 29/08/2026Someone should definitely make an anonymous cgit renderer that's up to 2010s qps standards, though, agreed. It's actually a very LLM friendly job, because you can even target HTML byte-equivalence on—ironically—a large example dataset.10:23 PM · Aug 29, 2026 · 18 views 4181
Xe @xeiaso.net · 29/08/2026↪ Replying to @filippo.abyssdomain.expertThe core problem is that git packfiles aren't seekable, so figuring out where an object is in a packfile requires seeking the entire packfile 3160 16
adg @nf.id.au · 30/08/2026↪ Replying to @filippo.abyssdomain.expertThere is some stuff to be optimised in cgit itself, I think. I set the robot on it and it was able to make some improvements: github.com/nf/cgit/blob... The change to reduce the work in finding decorations on the commit page seems particularly promising. (github.com/nf/cgit/comm...) 060 6
Josh Bleecher Snyder @commaok.xyz · 30/08/2026↪ Replying to @filippo.abyssdomain.expertI wonder whether you could kill two birds with one stone by making the client compute the diff (and do the downstream rendering). 020 2