@Julian finally made me interact on X by posting his GSR work there.
My original varops budget proposal simply tried to constrain the re-added opcodes, and those which now had to handle larger data.
His more empirical approach allowed him to expand the budget concept to cover script interpretation overhead, allowing to a very simple and elegant macro system (inspired, I believe, by Robin Linus?).
This maintains the "no loops" simplicity of Script but compresses the inevitable "now do this 32 times" cut & paste that happens in complex scripts.
It does this by adding a base cost to every opcode, which prevents abuse. A very slight change which makes it more general. Impressive!
Rusty Russell
rusty@rusty.ozlabs.org
npub179e9...lz4s
Lead Core Lightning, Standards Wrangler, Bitcoin Script Restoration ponderer, coder. Full time employed on Free and Open Source Software since 1998. Joyous hacking with others for over 25 years.
I disagree with nvk on several things, but I like his uncompromising approach to hardware and shipping. Testing for weak entropy is hard, this is absolutely a mistake I could have made, and I would also be furious if this lost my funds.
I've been trying to come up with a witty name though. Stolecard? Goxbox? "MicroSiphon inside"?