git history examples
| import 'ses'; | |
| lockdown(); | |
| console.log('lockdown successful'); |
| packages/swingset-runner$ node -r esm bin/runner --init --config demo/exchangeBenchmark/swingset.json --benchmark 1 run -- --prime | |
| alice moola before: balance {"brand":{},"value":3} | |
| alice simoleans before: balance {"brand":{},"value":0} | |
| bob moola before: balance {"brand":{},"value":0} | |
| bob simoleans before: balance {"brand":{},"value":3} | |
| bob moola after: balance {"brand":{},"value":1} | |
| bob simoleans after: balance {"brand":{},"value":2} | |
| alice moola after: balance {"brand":{},"value":2} | |
| alice simoleans after: balance {"brand":{},"value":1} |
| // | |
| // Large data for marshal benchmarking | |
| // | |
| // 'serialized.js' contains JSON-encoded capdata ({ body, slots }) | |
| // with a 621kB body and 643 slots. It represents the state of a | |
| // Wallet after many trade operations had been executed (and not all | |
| // of them were cleaned up: think of it as a worst-case serialization | |
| // task). | |
| import '@endo/init'; |
--- a/golang/cosmos/CHANGELOG.md
+++ b/golang/cosmos/CHANGELOG.md
@@ -3,6 +3,74 @@
All notable changes to this project will be documented in this file.
See [Conventional Commits](https://conventionalcommits.org) for commit guidelines.
+## [0.34.0](https://github.com/Agoric/agoric-sdk/compare/@agoric/cosmos@0.32.2...@agoric/cosmos@0.34.0) (2023-05-19)| #!/usr/bin/env node | |
| // @ts-nocheck | |
| import '@endo/init'; | |
| import fs from 'fs'; | |
| import sqlite3 from 'better-sqlite3'; | |
| import { readZip } from '@endo/zip'; | |
| let [dbfn] = process.argv.slice(2); | |
| const db = sqlite3(dbfn); |
Getting Claude-on-web to use a single Github repo, circa 09-may-2026
(note to self, or others, in case it helps)
I keep my service vendors separated in distinct browser profiles: I do all my Claude work in one, and my GitHub work in another. I use separate email (sub-)addresses for each. The idea was to reduce confused-deputy attacks and cross-vendor tracking. I can't say I recommend doing things this way, but it's what I do.
I was working with Claude Code on my laptop (the CLI/TUI thing you install from Homebrew) to build a web service, and it looked like it was going to take a while, so I wanted to move the work to some server so I could close my laptop. Also, I wanted to be able to visit the new site from my phone. I figured the claude-on-web tool (https://claude.ai/code , the one that runs on an Anthropic VM) would be a good approach (turns out maybe not). So it was time to learn how to set that up.