Got a great question in Twitter DMs. How do you stay motivated on working with boring/non-hype tech? I’m always hyped about a new stack and can’t make progress building a SAAS. It’s hard out there. If you read twitter, hacker news, etc, there is always something

I’ve been working on Rails applications for over 10 years. In that time I’ve figured out some guidelines on when and when not to use caching. Years ago, I’d cache almost everything. When joining GitHub, I remember being shocked by how little caching was actually used for

Rails 7 has added the load_async method to ActiveRecord. This looks like it could be a huge performance win for many Rails applications. Normally, a Rails app will execute all of its queries serially (one after another). The execution of the request waits while the query is executed. With

I wrote a little CORS tester tool to make testing urls for CORS easier. Mostly I was interested in experimenting with Cloudflare workers more. The entire site is deployed as a Cloudflare worker and served by a single js file. It worked real nicely for a small site. For a

View this post on Dev.to.

Here’s how to deploy to Heroku using GitHub Actions. Without adding any new dependencies. First, you need a Heroku auth token. You can generate one on your machine using the following: heroku authorizations:create Copy the Token value and add it as a secret to your repository (Go to

If you’re using Gems published to GitHub Package Registry. source "https://rubygems.pkg.github.com/mscoutermarsh" do gem "omniauth-auth0", "2.1.0" end You need credentials to install them. An easy trick for doing this is using bundler’s built in environment variable support. Add the following config var

View this post on Dev.to.

How to do it In CloudFlare, setup your DNS entry to look like this: The key part is, leaving the “cloud” set to DNS only (only for setup, you can enable CDN after setup). Once you do that, click “Validate” in Azure and you’ll get that “Domain Ownership” box

Here is how I got NPM modules working with Cloudflare workers + Serverless. 1. Install webpack npm install --save-dev webpack 1. Add a webpack config. webpack.config.js module.exports = { target: 'webworker' }; 1. Add build to your package.json "build": "webpack yourFunctionFile.js", 1. Update your serverless.yml. functions: helloWorld: worker:

Key is: use the hightlightjs shim. Install it using yarn. yarn add highlightjs Inside one of your packs (or application.js), add: import hljs from 'highlightjs' hljs.initHighlightingOnLoad();

Quickly improve the effectiveness of your communication via Slack by making your messages actionable. Bad Hey, can you do me a favor? Better Hey, can you do me a favor? This user account is locked out. Best Hey, can you do me a favor? This user account is locked out

I’ve found myself googling this every couple months. So I’m posting an example here to save you and myself in the future. If you’re creating an API user with full access to only a SINGLE S3 bucket, this example policy will work for you. If you want

Text fields in Postgres are case sensitive. This is a problem when storing emails because coutermarsh.mike@gmail.com and Coutermarsh.Mike@Gmail.com are not equal. If you do any lookups by email address, you have a few options. 1. always use downcase on write 2. use a functional

Quick postgres tip. If you have a JSONB column named “options” and it has a key “favorite_color”. Here is an example on how you select the value from it: SELECT options->'favorite_color' FROM table_name; Filter by a jsonb value: SELECT COUNT(*) FROM table_name WHERE (options @> '{

In the past 3 years, we’ve gone through several iterations of how we host avatars on Product Hunt. There were a lot of mistakes. This is what I recommend doing if you’re adding avatars to your web app. Short version: 1. Upload the original image directly to S3.

Every case is different. But I’d be wary of ever creating a dedicated model for a UserProfile. It sounds like a very general “throw everything here” type of model. Which is something you should avoid. It’s better to be very specific about the type of data being stored.

Here are my top 3 suggestions to improve your Rails codebase. 1. Learn the presenter and service object patterns. This allows you to break down complex logic into small, self contained and easily tested classes. This makes writing/changing/reading code easy. Remember to namespace things and create clear separation

Just setup our Capybara tests to use Chrome driver. Here is how I did it. First, added group :test do gem 'capybara' gem 'selenium-webdriver' gem 'chromedriver-helper' #<- New! end Then in the spec_helper.rb Capybara.register_driver :selenium do |app| Capybara::Selenium::Driver.new(app, browser: :chrome) end Capybara.javascript_driver = :chrome Capybara.configure do |config| config.default_max_wait_time = 10 # seconds config.default_driver = :selenium end Finally, w

The Rails mailer preview URL is impossible to remember. URL: http://localhost:3000/rails/mailers So I’m putting it here. Hopefully this just saved you time while googling. If you need more info about how to use previews. Read this.

Here’s how we purge images from imgix’s cache at Product Hunt with Ruby. If you’re googling around for a solution, hopefully this code helps you out. Imgix docs: https://docs.imgix.com/setup/purging-images # usage: Set env variable IMGIX_API_KEY to your api key # Need to

If you’re getting this error with the Twitter gem’s update_with_media method. And googling around for the solution. I’m here to help you. Twitter::Error::UnacceptableIO The IO object for media must respond to to_io Why this happens When you call open in Ruby with

View this post on Quora.

View this post on Quora.