const ENVIRONMENT_VARIABLES_AND_EXAMPLES: &str = r#"
Environment variables:
    RERUN_CHUNK_MAX_BYTES     Maximum chunk size threshold for the compactor.
    RERUN_CHUNK_MAX_ROWS      Maximum chunk row count threshold for the compactor (sorted chunks).
    RERUN_CHUNK_MAX_ROWS_IF_UNSORTED
                              Maximum chunk row count threshold for the compactor (unsorted chunks).
    RERUN_SHADER_PATH         The search path for shader/shader-imports. Only available in developer builds.
    RERUN_TRACK_ALLOCATIONS   Track memory allocations to diagnose memory leaks in the viewer.
                              WARNING: slows down the viewer by a lot!
    RERUN_MAPBOX_ACCESS_TOKEN The Mapbox access token to use the Mapbox-provided backgrounds in the map view.
    RUST_LOG                  Change the log level of the viewer, e.g. `RUST_LOG=debug`.
    WGPU_BACKEND              Overwrites the graphics backend used, must be one of `vulkan`, `metal` or `gl`.
                              Default is `vulkan` everywhere except on Mac where we use `metal`. What is
                              supported depends on your OS.
    WGPU_POWER_PREF           Overwrites the power setting used for choosing a graphics adapter, must be `high`
                              or `low`. (Default is `high`)


Examples:
    Open a Rerun Viewer that listens for incoming SDK connections:
        rerun

    Load some files and show them in the Rerun Viewer:
        rerun recording.rrd mesh.obj image.png https://example.com/recording.rrd

    Open an .rrd file and stream it to a Web Viewer:
        rerun recording.rrd --web-viewer

    Host a Rerun TCP server which listens for incoming TCP connections from the logging SDK, buffer the log messages, and serves the results over WebSockets:
        rerun --serve-web

    Host a Rerun Server which serves a recording over WebSocket to any connecting Rerun Viewers:
        rerun --serve-web recording.rrd

    Connect to a Rerun Server:
        rerun ws://localhost:9877

    Listen for incoming TCP connections from the logging SDK and stream the results to disk:
        rerun --save new_recording.rrd
"#;