ronmi/prpr repository overview

Simple application fetching static HTML via Firefox Marionette.

It is designed most for internal using. YOU HAVE TO PREVENT SECURITY ISSUE YOURSELF.

⁠Configuration

Prpr accepts following environmental variables:

• SECRET: shared secret. Default to "" which disables authenticating.
• FIREFOX: path to Firefox binary. Default to "firefox".
• DEBUG_FIREFOX: disable --headless mode.
• FIREFOX_OPTS: extra command line arguments passed to Firefox.
• BIND: binding address passed to ListenAndServe. Default to ":9801".
• QUEUE_SIZE: max concurrent loading pages. Default to 1.

It always passes --marionette and --safe-mode to Firefox.

⁠Extra configuration for Docker

The docker image ronmi/prpr will create new profile directory in /tmp each run, which might waste disk space if you start/stop it frequently.

You can mount a volume somewhere and pass it in envvar FIREFOX_PROFILE

docker run -d --name prpr \
  -v "$(pwd)/data:/fxprofile" \
  -e "FIREFOX_PROFILE=/fxprofile" \
  -p "127.0.0.1:9801:9801" \
  ronmi/prpr

⁠How it fetches HTML

1. Allocate a free browser window (blocked if none free)
2. Switch to allocated window
3. Navigate to the url
4. Wait a second
5. Every second, check if the HTML element presents
6. If it does exists, run a piece of JS to fetch HTML and return
7. Returns error after 10 tries

It is much more complex in real code as prpr now supports concurrent loading.

⁠Usage

wget \
  --post-data='uri=http://google.com&wait=queryString&secret=mysecret' \
  http://127.0.0.1:9801/grab

• uri: URL you want to fetch from.
• wait: CSS-selector to wait for.
• secret: your shared secret.

It returns HTML code if success, 400 if shared secret mismatch and 500 when failed.

⁠WARNING

• To keep code simple, prpr just waits 10 seconds before actually connect to firefox. IT DOES NOT SUPPORT RECONNECTING.
• For best security, DO NOT EXPOSE IT TO PUBLIC NETWORK.

MAKE SURE YOU KNOW WHAT YOU ARE FETCHING!

⁠License

WTFPL