Skip to content

dev.md

Latest commit

 

History

History
92 lines (66 loc) · 2.01 KB

dev.md

File metadata and controls

92 lines (66 loc) · 2.01 KB

development

  • golang >= 1.24
  • direnv to load environment vars
cp ./.env.example .env

If you are running apps outside of docker, remember to change the postgres, and imgproxy hostnames to "localhost" in .env.

Initialize local env variables using direnv

echo dotenv > .envrc && direnv allow

Boot up database (or bring your own)

docker compose -f docker-compose.yml -f docker-compose.override.yml --profile db up -d

Create db and migrate

make create
make migrate
go run ./cmd/pgs/ssh
# in a separate terminal
go run ./cmd/pgs/web

sign up and upload files

The initial database has no users, you need to sign up via pico/ssh:

go run ./cmd/pico/ssh
# in a separate terminal, complete the signup flow, set your username to "picouser"
ssh localhost -p 2222

Stop the pico SSH server, then you can upload files:

go run ./cmd/pgs/ssh
# in a separate terminal
go run ./cmd/pgs/web
# in a third terminal
echo 'Hello, World!' > file.txt
scp -P 2222 file.txt localhost:/test/file.txt
curl -iH "Host: picouser-test.pgs.dev.pico.sh" localhost:3000/file.txt

deployment

We use an image based deployment, so all of our images are uploaded to ghcr.io/picosh/pico

DOCKER_TAG=latest make bp-all

Once images are built, docker compose is used to stand up the services:

docker compose up -d

This makes use of a production .env.prod environment file which defines the various listening addresses and services that will be started. For production, we add a .envrc containing the following:

export COMPOSE_FILE=docker-compose.yml:docker-compose.prod.yml
export COMPOSE_PROFILES=services,caddy

And symlink .env to .env.prod:

ln -s .env.prod .env

This allows us to use docker-compose normally as we would in development.

For any migrations, logging into the our database server, pulling the changes to migrations and running make latest is all that is needed.