Satoshi Endpoint Proxy

[arik-so]

The server --proxy option is helpful for testing mainnet inscriptions' behavior locally, but a lot of inscriptions that require such testing rely on a specific dynamically inscribed ordinal, which is usually uninscribed locally, and breaks testability.

I wonder whether allowing the --proxy option to also apply as a fallback for the /r/sat/:sat/at/:index endpoint might be worth pursuing? That should help with resolving dynamic content.

[arik-so]

If the Self::proxy() return type were correct, the following spaghetti code might be an approximation:

  async fn sat_inscription_at_index(
    Extension(index): Extension<Arc<Index>>,
    Extension(server_config): Extension<Arc<ServerConfig>>,
    Path((DeserializeFromStr(sat), inscription_index)): Path<(DeserializeFromStr<Sat>, isize)>,
  ) -> ServerResult<Json<api::SatInscription>> {
    task::block_in_place(|| {
      if !index.has_sat_index() {
        return if let Some(proxy) = server_config.proxy.as_ref() {
          Self::proxy(proxy, &format!("r/sat/{}/at/{}", sat, inscription_index))
        } else {
          Err(ServerError::NotFound(
            "this server has no sat index".to_string(),
          ))
        };
      }

      let id = index.get_inscription_id_by_sat_indexed(sat, inscription_index)?;
      if id.is_none() {
        if let Some(proxy) = server_config.proxy.as_ref() {
          return Self::proxy(proxy, &format!("r/sat/{}/at/{}", sat, inscription_index));
        }
      }

      Ok(Json(api::SatInscription { id }))
    })
  }

[raphjaph]

Yes, this is a good idea. I even have an issue for this #3806

[arik-so]

Oh oops, sorry!