Add unit tests for Tracker

[josecelano]

I'm double-checking the new Axum HTTP tracker and adding more tests.

[josecelano]

Hi @da2ce7 @WarmBeer, I've just realized we have a duplicate struct:

/// Swarm statistics for one torrent.
/// Swarm metadata dictionary in the scrape response.
/// BEP 48: <https://www.bittorrent.org/beps/bep_0048.html>
#[derive(Debug, PartialEq, Default)]
pub struct SwarmMetadata {
    pub complete: u32,   // The number of active peers that have completed downloading (seeders)
    pub downloaded: u32, // The number of peers that have ever completed downloading
    pub incomplete: u32, // The number of active peers that have not completed downloading (leechers)
}

/// Swarm statistics for one torrent.
/// Alternative struct for swarm metadata in scrape response.
#[derive(Debug, PartialEq, Default)]
pub struct SwarmStats {
    pub completed: u32, // The number of peers that have ever completed downloading
    pub seeders: u32,   // The number of active peers that have completed downloading (seeders)
    pub leechers: u32,  // The number of active peers that have not completed downloading (leechers)
}

I introduced the SwarmMetadata to use the same names as in the BEP 48. I think SwamStats is a better name and the names for attributes as well.

I would keep both and use SwarmMetadata only on the scrape context and SwamStats for the rest of the application. Maybe we could use SwarmMetadata when we build the final response.

On the other hand, It could be confusing to use two different names, although the BitTorrent specification (BEPs) do that all the time. The SwarmMetadata.complete != SwamStats.completed is especially misleading.

[josecelano]

Hi @da2ce7 @WarmBeer, this is ready to review. Finally, I've added only unit tests for the core Tracker. I will open new PRs to add more unit tests for the new HTTP Axum implementation. In general, I think we should increase the unit tests and decrease the integration tests.