New post: Poor man's multi-process development

src/posts/2023-10-09-poor-mans-multi-process-development/index.md

@@ -0,0 +1,36 @@
+---
+highlight: true
+path: poor-mans-multi-process-development
+title: Poor man's multi-process development
+categories:
+  - engineering
+author: miguel-palhas
+date: 2023-10-09
+intro: >-
+  Sharing a quick tip that has saved me hours of fiddling with Docker and other complex tools, while retaining an easy-to-use development environment
+---
+
+Local development often involves running many processes simultaneously. e.g.: a web server along with a database or blockchain node.
+
+This is often done with `docker-compose`, or other equivalent orchestration tools. However, these tools completely change the workflow, and hinder other abilities, such as debugging, live-reloading and interoperability with your host system (your mileage may vary depending on your actual tech stack).
+The naive solution, opening multiple terminal splits and manually run each command, works, but makes it less than ideal to onboard new developers, and can be error-prone.
+
+In fact, there's a simple Bash setup that can be used when setting up small projects:
+
+```bash
+#!/bin/bash -ue
+
+trap 'kill %1; kill %2' SIGINT
+
+yarn dev | sed -e 's/^/\x1b[0;31m[web]\x1b[0m /' &
+anvil | sed -e 's/^/\x1b[0;32m[anvil]\x1b[0m /' &
+
+wait
+```
+
+This does most of what I usually need out of Docker in my development environment:
+
+- Spins up multiple processes in the same command
+- Retains the output of each of them
+- Prefixes each processes' output with a custom tag & color
+- Ctrl+C still works, killing all running processes (that's where `trap` comes in)