minimalist-fulcro-template-fly

This is a clone on minimalist-fulcro-template extended with PostgreSQL and deploy to Fly.io.

Key differences:

You need to run docker compose up in this folder to start PostgreSQL
Evaluating com.example.server.main will not only start the server but also apply DB migrations from resources/db-init.sql
Pathom resolvers can use (:conn env) with next.jdbc/execute! to run queries
You must manually call (start) in the main ns to start the server

Operation

Local

Run via docker

docker compose up        # start Postgres
docker build -t ex-fly . # build the image
# Run the image, pointing to Postgres running on the host:
docker run --rm -it -p 9999:8008 \
  -e DATABASE_URL="jdbc:postgresql://postgres:FulcroRulez@host.docker.internal:5432/postgres"  \
  ex-fly:latest

Fly

Fly setup

fly launch --name mft-fly --no-deploy - here we choose the app name mft-fly
fly postgres create and choose app name, e.g. mft-fly-db This will print a connect string such as postgres://postgres:sxRhUb2HlJUjSuc@mft-fly-db.internal:5432
Set the DB url env var (see the config ns) for the app using the value: fly secrets set -a mft-fly DATABASE_URL="postgres://postgres:sxRhUb2HlJUjSuc@mft-fly-db.internal:5432"
Deploy: fly deploy
Access your deployed app under its name, here: https://mft-fly.fly.dev/
For Calva, the server is started automatically during jack-in
Factor out resolvers ns & make it so that you only need to load pathom ns to get resolver changes in

Original minimalist-fulcro-template instructions

A template for starting a new, minimalistic, full-stack Fulcro application. Intended for playing with and learning Fulcro, not for production apps, and therefore simpler than the official fulcro-template. It is a good starting point for your learning projects that is hopefully simple enough for you to understand.

TIP: For an even simpler template with an in-browser backend, see minimalist-fulcro-template-backendless.

Creating a new application from the template

Download or clone this repository to your computer and start hacking away.

Explanation

You will run shadow-cljs, which will watch, compile, and update the sources and separately you will start a HTTP server from the REPL to serve the application and process Pathom requests.

Usage

Prerequisites: same as shadow-cljs'.

First, install frontend dependencies via npm, yarn, or similar:

npm install # or yarn install # reportedly yarn < v3

then start the application either via

npx shadow-cljs watch main

or, if you have Babashka installed, via

bb run

NOTE: For Calva, it is best to start a client build and REPL from the editor - run Jack-in, selecting shadow-cljs then the :main build.

Now start the server: load com.example.server.main into the Shadow REPL - this will also evaluate the (defonce stop-fn (atom start)) line, starting the server.

NOTE: Now if you ever change Pathom resolvers or something, run the restart code in the (comment ...) in the main ns.

Finally, navigate to http://localhost:8008. Note: You can switch to the browser REPL by evaluating (shadow/repl :main) in the REPL. (Calva does the latter for you).

Create a standalone build

You can also compile the sources into a directory via

npx shadow-cljs release main
# or: bb build

Why is this not suitable for production?

No thought was given to security, performance, monitoring, error tracking and other important production concerns. It also bakes in fulcro-troubleshooting, which you do not want unnecessarily increasing your bundle size in production settings. So if you want to use the template as a starting point for a production application, you will need to add those yourself.

License

Distributed under the Unlicense.

brancusi / fulcro-recipe-builder-demo