#ilovefs SharePic generator. https://sharepic.fsfe.org
You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
Renovate Bot 6aae5a09d7 backend: golang: 1.17 -> 1.18 (#23) 2 months ago
.reuse style selection of templates, improve form 4 months ago
LICENSES backend: switch to Go Imagick for ImageMagick 4 months ago
backend backend: golang: 1.17 -> 1.18 (#23) 2 months ago
frontend frontend: add hashtag in text 3 months ago
.dockerignore backend: Dockerize and limit ImageMagick 5 months ago
.drone.yml backend: introduce test run 3 months ago
.gitignore backend: reduce privileges on Linux via seccomp 3 months ago
README.md backend: replace syscall list with syscallset 2 months ago
docker-compose.dev.yml differentiate between local development mode and production mode 4 months ago
docker-compose.yml backend: global configuration via YAML 4 months ago
renovate.json chore(deps): add renovate.json 3 months ago
renovate.json.license make REUSE compliant 3 months ago
shell.nix backend: reduce privileges on Linux via seccomp 3 months ago


SharePic Generator

Build Status REUSE status

A SharePic - a compound of to share and picture - is a themed picture to communicate a quote of a person. This repository contains a web application to generate such SharePics for the FSFE's #ilovefs campaign.

Getting Started

To run this SharePic Generator, Docker and Docker Compose needs to be installed. The software can be started for local development via:

docker-compose -f docker-compose.yml -f docker-compose.dev.yml up -d --build

When it is running, it is accessible at http://localhost:8830/.

Adding a new template to the backend

A template consists of two identically named files - one with the .svg and one with the .yml extension - within the backend/templates directory. Creating those two files is all it takes for adding a new template.

During the backend software's compile time, those files are sourced and included in the binary. When starting up, those files are parsed and prepared within the program's memory.

Customized SVG file

To use a SVG graphic as the base, some manual modifications needs to be performed to the file. It's recommended to use a text editor therefore.

  • Create a pseudo-image tag like the following one where the user's image should appear:
    <image . . . xlink:href="data:image/jpeg;base64,{{.ImageData}}" />
  • Replace the one-line text for the user's name with: {{.AuthorName}}.
  • Replace the one-line text for the user's position or description with: {{.AuthorDesc}}.

YAML Configuration

Note: The file extension must be .yml, not .yaml.

The YAML file describes both the created sharepic as well as the parameters for the multi-lined user message. A described example configuration follows.

# Resulting sharepic's dimensions.
  width: 640
  height: 480

# Size the user submitted picture should be reduced to.
# This value should roughly match the image tag within the SVG.
  width: 80
  height: 87

# Font to be used for the message - must be installed on the system.
# The font sizes will all be tried, the biggest possible one will be used.
  name: Liberation-Sans
    - 22
    - 26
    - 30
    - 34

# Maximum amount of characters for the message, the author, and its description.
  message: 200
  author: 50
  description: 50

# The message will appear within an (invisible) box with its own dimensions,
# which must be smaller than those of the image.
# The margin defines the offset from the top-left point of the image.
  width: 289
  height: 220

  margin_width: 27
  margin_height: 72


The application consists of two components: a frontend and a backend. On the frontend, an HTTP server provides an input mask that can be used to create SharePics. Then the backend receives this data and creates the SharePic, which is delivered via the frontend.

To simplify deployment, Docker Compose is a key component. This completely takes care of creating the necessary containers, isolating them and allowing them to communicate with each other.


The frontend is an Apache web server, which delivers the web page with the input mask as static files. To communicate with the backend, the Apache httpd proxies the corresponding HTTP POST request to the backend.

Since the web page does not require any compilation, it is mounted read-only as a volume into the frontend container in the local development mode (using the -f docker-compose.dev.yml argument). Thus, changes can be made live without restarting. In production mode, it will be copied statically to the container.


The backend consists of an application written for this purpose in the Go programming language, which also starts an HTTP server. On the /sharepic endpoint it expects the POST request of the frontend and creates the SharePic accordingly.

Technically, the SVG image file is used as a template and modified by the sanitized input data. Then ImageMagick is used to convert the customized SVG to a PNG file.

All configuration is achieved by the YAML files within backend/inc. While the backend/inc/backend.yml file defines global settings, a specific .yml file for each template in backend/inc/templates configures template related settings, i.e. the font.

Security Considerations

The most critical component would seem to be ImageMagick, which has a reputation for security issues. Thus, ImageMagick is limited to the necessary modules by a custom Security Policy in the backend/inc/policy.xml file. Further, the backend software also checks that only allowed file types can be uploaded.

The software in the backend container runs with limited user rights and a seccomp-bpf filter applied. Also, the container is not in the default network, so no outgoing connections to the Internet should be possible. From the outside, the backend container should not be accessible, but only the /sharepic HTTP endpoint through the Apache reverse proxy in the frontend.

However, the frontend container is also configured to run the Apache web server with limited user rights.

There is no bidirectional writable volume between container and host. The frontend has read-only rights to the data on the web page. The backend stores temporary files in a volatile manner, which it deletes immediately after each operation.

Privacy Considerations

Despite the fact that the software is based on a typical client-server model, efforts were made to minimize the amount of data generated and not to store it anywhere permanently.

The Apache2 web server in the frontend was configured to not log personal data like IP addresses.

Both the backend software's architecture as well as its deployment is built with the privacy aspect in mind. No personal data should be logged and when creating files are necessary, those are written on a tmpfs and are immediately deleted afterwards.

More details regarding this are present in issue #1.


This program is REUSE compliant. Each file contains or is accompanied by licensing and copyright information. The majority of the code is licensed under AGPL-3.0-or-later.