Using Shell Scripts for a Better User Experience (source for https://robots.thoughtbot.com/improving-user-experience-with-shell-scripts)

README.md

Server scripts

This is the source for the scripts discussed in https://robots.thoughtbot.com/improving-user-experience-with-shell-scripts

Both scripts are in the bin/ directory of the repo that contains all the markdown documents for blog posts. Users run bin/server and everything is automatically set up for them to view a local preview of the blog. bin/server-setup is a dependency of bin/server and is never run directly by users.

Maitre-d is the name of the "blog engine" discussed in the article.

server

#!/usr/bin/env bash

LOG_FILE=tmp/server.log

green() { printf "\033[32m${1}\033[0m\n"; }
yellow() { printf "\033[33m${1}\033[0m\n"; }
red() { printf "\033[31m${1}\033[0m\n"; }

yellow "Setting up server (may take a while)..."

if ./bin/server-setup "$LOG_FILE"; then
  green "...success! Starting server, visit http://localhost:3002"
  echo
  cd tmp/maitre-d && bundle exec rails server --port 3002
else
  red "Oops, something went wrong :/"
  red "Look in tmp/server.log for the setup log"
  red "If you have any questions, please ask in the #blog channel"
fi

server-setup

#!/usr/bin/env bash

# Exit if any subcommand fails
set -e

# Log everything to $LOG_FILE
LOG_FILE="$1"
exec > "$LOG_FILE" 2>&1

mkdir -p tmp
cd tmp

# Maitre-d is the blog engine
# Get the latest Maitre-d

if test -d maitre-d; then
  cd maitre-d
  git pull
else
  git clone --depth=1 [email protected]:thoughtbot/maitre-d
  cd maitre-d
fi

# Check Ruby installation
rubyversion="$(< .ruby-version)"
if ! ruby --version | grep -Fq "$rubyversion"; then

  if command -v rbenv > /dev/null; then
    # install via rbenv
    printf 'Installing the version of Ruby for this project via rbenv...\n'
    brew update
    brew upgrade ruby-build || true
    rbenv install
  elif command -v rvm > /dev/null; then
    # install via RVM
    printf 'Installing the version of Ruby for this project via RVM...\n'
    source "$HOME/.rvm/scripts/rvm"
    rvm install "$rubyversion" && rvm use "$rubyversion"
  elif command -v ruby-install > /dev/null; then
    # install via ruby-install
    printf 'Installing the version of Ruby for this project via ruby-install...\n'
    ruby-install ruby "$rubyversion" && chruby "$rubyversion"
  else
    printf 'You do not have the correct version of Ruby installed.\n'
  fi
fi

# Set up Ruby dependencies via Bundler
gem install bundler --conservative --no-document
bundle check || bundle install --without test

# Set up Maitre-d ENV variables to point to parent robots folder
cp .sample.env .env
echo "LOCAL_POSTS_PATH=../../source/posts/**/*.md" >> .env

# Maitre-d requires a database
bundle exec rake db:setup

# Add FILEPICKER_API_KEY if not set to the real value
if ! grep -Fq FILEPICKER_API_KEY=A .env; then
  heroku join --app maitre-d-staging
  heroku config:get FILEPICKER_API_KEY --shell --app maitre-d-staging >> .env
fi

@JoelQ

I guess you could also do something like this (triggered whenever you modify a file in the markdown repo on your local machine, e.g. using rb-fsevent):

git status --porcelain | awk '{ print $2 }' | while read name; do cat $name | curl -X POST -d @- 'http://robots.thoughtbot.com/preview'; done;

It loops through changed markdown files and POSTs their contents to the /preview action on your production server. The /preview action creates a draft from the data and allows you to preview your new post right in your production blog.

I probably messed up the command, but I hope you get the idea - which is to avoid setting up anything on local machines.