add README

.github/workflows/build-and-publish.yml

@@ -23,8 +23,6 @@ jobs:
 
       - name: Set up docker buildx
         uses: docker/setup-buildx-action@v3
-        with:
-          platforms: linux/amd64,linux/arm/v7,linux/arm64
 
       - name: Login to GitHub Packages
         uses: docker/login-action@v2
@@ -37,6 +35,7 @@ jobs:
         uses: docker/build-push-action@v5
         if: ${{ github.ref_type == 'tag' }}
         with:
+          platforms: linux/amd64,linux/arm/v7,linux/arm64
           context: ./
           tags: "ghcr.io/${{ github.repository }}:${{ github.ref_name }}"
           push: true
@@ -45,6 +44,7 @@ jobs:
         uses: docker/build-push-action@v4
         if: ${{ github.ref_type != 'tag' }}
         with:
+          platforms: linux/amd64,linux/arm/v7,linux/arm64
           context: ./
           tags: "ghcr.io/${{ github.repository }}:latest"
           push: ${{ github.ref == 'refs/heads/main' }}

README.md

@@ -1,3 +1,28 @@
 # omnilock
 
-A zero dependency lock manager
+A zero dependency locker.
+
+## How does it work?
+
+This locker uses TCP connections to keep track of locks.
+This way, a lock is freed as soon as the locking process ends (some way or another) or it's unlocked explicitly.
+
+## Protocol
+
+1. Open a TCP connection to the omnilock server
+2. Send a lock command through the established connection (e.g. `lock my_lock`)
+3. Read the response, it may be either `success`, `failed` or `error: something went wrong`
+4. Do your things on the client side
+5. Either unlock the lock explicitly: `unlock my_lock` or just let the process exit so the OS closes the connection
+
+Note: All commands and responses end with `\n` (newline).
+
+## Usage as docker image
+
+Start the server with:
+
+```shell
+docker run --rm -it -p 7194:7194 ghcr.io/nitwhiz/omnilock:lastest
+```
+
+And connect to it via port 7194.