Max Schwenk maschwenk

Shipping tsrs-lint on npm

The tsrslint PR merged into oxc-project/tsrs today. It adds crates/tsrs_lint with a tsrs-lint binary that speaks tsgolint's headless protocol and covers all 61 tsgolint rules at v7.0.2005. Nothing ships it yet. release.yml only builds tsrs_cli, and tools/tsgolint/README.md says npm packaging is out of scope. The tracking issue asks for tsrs to become the type-aware backend for Oxlint, and the goal is one package rather than a separate fork (@maschwenk/tsrslint). The open question is how to package it.

In a quick check on a large private codebase, upstream tsrs-lint produced the same diagnostics as the fork under oxlint, at about the same speed. Peak memory was noticeably higher than the fork's.

Question 1: where does the binary ship?

option	what it is	upside	downside

typescript-benchmarking runs for microsoft/TypeScript#64474

Raw results for the numbers in microsoft/TypeScript#64474, microsoft/TypeScript#64475 and microsoft/TypeScript#64476.

Setup

Apple M5 Max, 18 cores, 128 GB, macOS 26.6.2
Go 1.27.1, go build ./cmd/tsc in tsc/ (libs embedded)
scenarios from microsoft/typescript-benchmarking at ab43240f, cloned and installed by setup.sh (same as the scenario setup scripts but with

Shit to setup your computer with

Install oh my zsh

sh -c "$(curl -fsSL https://raw.github.com/ohmyzsh/ohmyzsh/master/tools/install.sh)"

Put this into your ~./.zshrc file

Problem

# Factory uses Faker::Name.first_name under the hoode -- evaluates to something with 'frank' as a substring
# Now you have two users named frank without realizing it
let(:user) { create :user } 
let(:another_user) { create :user, name: 'frank' }

it do
 filter_for_frank

Request spec subjects

Who TF cares honestly? 🤷‍♂️ 🤷‍♀️

Great question. Let me preface this by saying I HATE talking about RSpec best practices. It's irrelevant to shipping software and tends to churn up a lot of long, useless discussion. However, I think it's useful to consider best practices when there is a demonstrable functional (not stylistic) difference between two approaches that can affect maintainability, flexibility, and speed.

One style I see quite a lot looks a bit like this:

The `before` hook tower of doom

wget https://downloads.pwnedpasswords.com/passwords/pwned-passwords-1.0.txt.7z 5GB file, expands to ~10GB
brew install unar (optional if you have a cli tool that can unarchive 7z files)
unar pwned-passwords-1.0.txt.7z
fgrep $(echo -n "password youre looking for" | openssl sha1 | tr '[:lower:]' '[:upper:]') pwned-passwords-1.0.txt

$(echo -n "password youre looking for" | openssl sha1 | tr '[:lower:]' '[:upper:]') gives you a SHA1, uppercased

Keybase proof

I hereby claim:

I am maschwenk on github.
I am maschwenk (https://keybase.io/maschwenk) on keybase.
I have a public key ASAyC0V4YqxxtZUPz2MjpoYg2zi7qMJuhvaLhGQ3ZZ2qMAo

To claim this, I am signing this object:

	import { parse } from 'csv-parse@5.3.10/sync';
	import * as fs from 'fs';

	const COST_PER_MINUTE = 0.0028;

	interface ContainerBuild {
	project_name: string;
	build_count: number;
	minutes_saved: number;
	minutes_billed: number;

	#!/usr/bin/env ruby
	# 1. put this in brew path `/usr/local/bin/`
	# 2. make it executable `chmod +x /usr/local/bin/recent-branches`
	# 3. in git config `~/.gitconfig` place an alias to the command inside `[alias]` block
	# `br = recent-branches -n 20` (n being the number of branches you wanna show by default)
	require 'optparse'
	require 'English'

	options = { number: 20 }

	# vim:ft=zsh ts=2 sw=2 sts=2
	#
	# agnoster's Theme - https://gist.github.com/3712874
	# A Powerline-inspired theme for ZSH
	#
	# # README
	#
	# In order for this theme to render correctly, you will need a
	# [Powerline-patched font](https://github.com/Lokaltog/powerline-fonts).
	# Make sure you have a recent version: the code points that Powerline