According to BCMR spec, if bare domains are un op_return, we should look up "well known" path for the domain. Fallback to previous behaviour if well known is not found.
372 lines
11 KiB
Rust
372 lines
11 KiB
Rust
// Copyright (C) 2024-2026 Whiterun LLC
|
|
//
|
|
// This software is licensed under the GNU Affero General Public License (AGPL), version 3.0 or later.
|
|
// A copy of the license can be found in the LICENSE file or at https://www.gnu.org/licenses/agpl-3.0.html
|
|
|
|
use std::{
|
|
sync::{atomic::AtomicBool, atomic::Ordering, Arc},
|
|
time::Duration,
|
|
};
|
|
|
|
use crate::bcmr::parsedbcmr::{parse_bcmr_json, SOURCE_ON_CHAIN};
|
|
use crate::db::bcmr::insert_bcmr_data;
|
|
use crate::db::bcmr::update_bcmr_failure;
|
|
use crate::{
|
|
bcmr::parse_bcmr_from_opreturn,
|
|
db::bcmr::{get_entries_missing_bcmr_download, AuthChainEntry},
|
|
};
|
|
use anyhow::*;
|
|
use bitcoin_hashes::hex::ToHex;
|
|
use bitcoincash::Script;
|
|
use log::{info, warn};
|
|
use rand::thread_rng;
|
|
use serde_json::Value;
|
|
use sqlx::SqlitePool;
|
|
use tokio::task::JoinHandle;
|
|
|
|
use rand::seq::SliceRandom;
|
|
use std::result::Result::Ok;
|
|
|
|
use super::utilurl::get_url;
|
|
|
|
// We are generous on timeout to allow for slow ipfs gateway
|
|
const DOWNLOAD_TIMEOUT: Duration = Duration::from_secs(60);
|
|
const IPFS_GATEWAYS: [&str; 2] = ["https://ipfs.io/ipfs/", "https://w3s.link/ipfs/"];
|
|
const MAX_BCMR_SIZE: usize = 1024 * 1024 * 100; // 100 MB
|
|
|
|
const MAX_PARALLEL_DOWNLOADS: usize = 5;
|
|
|
|
const WELL_KNOWN_PATH: &str = "/.well-known/bitcoin-cash-metadata-registry.json";
|
|
|
|
/// Resolve an on-chain BCMR URI into one or more candidate URLs to try in order.
|
|
///
|
|
/// Handles three cases:
|
|
/// - `ipfs://<cid>` → a single random IPFS gateway URL.
|
|
/// - Bare domain (authority only, no path) → BCMR v2 well-known URI first, then
|
|
/// the bare root URL as a fallback for bug-compatibility with tokens that
|
|
/// served the registry directly at `/`.
|
|
/// - Anything with a path → used as-is (with `https://` prepended if missing).
|
|
fn resolve_uri_candidates(uri: &str) -> Vec<String> {
|
|
if let Some(stripped) = uri.strip_prefix("ipfs://") {
|
|
return vec![format!(
|
|
"{}{}",
|
|
IPFS_GATEWAYS.choose(&mut thread_rng()).unwrap(),
|
|
stripped
|
|
)];
|
|
}
|
|
|
|
let (scheme, rest) = match uri.split_once("://") {
|
|
Some((s, r)) => (s, r),
|
|
None => ("https", uri),
|
|
};
|
|
|
|
let (authority, path) = match rest.split_once('/') {
|
|
Some((a, p)) => (a, format!("/{p}")),
|
|
None => (rest, String::new()),
|
|
};
|
|
|
|
if path.is_empty() || path == "/" {
|
|
vec![
|
|
format!("{scheme}://{authority}{WELL_KNOWN_PATH}"),
|
|
format!("{scheme}://{authority}/"),
|
|
]
|
|
} else {
|
|
vec![format!("{scheme}://{authority}{path}")]
|
|
}
|
|
}
|
|
|
|
pub struct BCMRDownloader {
|
|
db: SqlitePool,
|
|
keep_running: Arc<AtomicBool>,
|
|
download_task: Option<JoinHandle<()>>,
|
|
}
|
|
|
|
const LOOP_SLEEP_TIME: Duration = Duration::from_secs(10);
|
|
|
|
async fn get_download_candidates(db: &SqlitePool) -> Vec<AuthChainEntry> {
|
|
match get_entries_missing_bcmr_download(db).await {
|
|
Ok(c) => c,
|
|
Err(e) => {
|
|
warn!("bcmr: Failed to fetch bcmr download candidates: {e}");
|
|
Vec::default()
|
|
}
|
|
}
|
|
}
|
|
|
|
async fn fetch_bcmr(
|
|
client: &reqwest::Client,
|
|
urls: &[String],
|
|
expected_hash: &str,
|
|
) -> (
|
|
Option<(String, String)>,
|
|
String, /* error */
|
|
bool, /* fatal error (don't try again) */
|
|
) {
|
|
let mut candidate: Option<(String, String)> = None;
|
|
let mut errors: Vec<String> = Vec::new();
|
|
|
|
for url in urls {
|
|
for resolved in resolve_uri_candidates(url) {
|
|
let (contents, actual_hash) =
|
|
match get_url(client, &resolved, DOWNLOAD_TIMEOUT, MAX_BCMR_SIZE).await {
|
|
Ok(c) => c,
|
|
Err(e) => {
|
|
errors.push(e.to_string());
|
|
continue;
|
|
}
|
|
};
|
|
|
|
if actual_hash == expected_hash {
|
|
return (Some((contents, actual_hash)), String::default(), false);
|
|
} else {
|
|
candidate = Some((contents, actual_hash))
|
|
}
|
|
}
|
|
}
|
|
|
|
// No URL's with matching expected hash
|
|
if candidate.is_some() {
|
|
(candidate, String::default(), false)
|
|
} else if errors.is_empty() {
|
|
(None, "No URLs to fetch BCMR from".to_string(), true)
|
|
} else {
|
|
(
|
|
None,
|
|
format!("Errors fetching BCMR: {}", errors.join("; ")),
|
|
false,
|
|
)
|
|
}
|
|
}
|
|
|
|
async fn process_entry(client: &reqwest::Client, db: &SqlitePool, entry: &AuthChainEntry) {
|
|
info!(
|
|
"bcmr: Downloading BCMR for token {}, txid {}",
|
|
entry.token_id.to_hex(),
|
|
entry.txid.to_hex()
|
|
);
|
|
|
|
let Some(raw) = entry.bcmr_data.as_ref() else {
|
|
warn!(
|
|
"bcmr: Missing OP_RETURN BCMR for utxo {} (token {}, txid {}); skipping.",
|
|
entry.utxo.to_hex(),
|
|
entry.token_id.to_hex(),
|
|
entry.txid.to_hex()
|
|
);
|
|
return;
|
|
};
|
|
|
|
let bcmr = match parse_bcmr_from_opreturn(&Script::from(raw.clone())) {
|
|
Some(v) => v,
|
|
None => {
|
|
warn!(
|
|
"bcmr: Invalid BCMR OP_RETURN for token {}, txid {}",
|
|
entry.token_id.to_hex(),
|
|
entry.txid.to_hex()
|
|
);
|
|
return;
|
|
}
|
|
};
|
|
|
|
let (json_str, error, is_fatal) = fetch_bcmr(client, &bcmr.uris, &bcmr.hash.to_hex()).await;
|
|
|
|
let (json_str, actual_hash) = match json_str {
|
|
Some(j) => j,
|
|
None => {
|
|
if let Err(e) = update_bcmr_failure(
|
|
db,
|
|
&entry.utxo,
|
|
&entry.txid,
|
|
&entry.token_id,
|
|
&format!("Failed to fetch BCMR: {error}"),
|
|
is_fatal,
|
|
)
|
|
.await
|
|
{
|
|
warn!("bcmr: Failed to set BCMR error: {e}");
|
|
}
|
|
return;
|
|
}
|
|
};
|
|
|
|
let json: Value = match serde_json::from_str(&json_str) {
|
|
Ok(b) => b,
|
|
Err(e) => {
|
|
if let Err(e2) = update_bcmr_failure(
|
|
db,
|
|
&entry.utxo,
|
|
&entry.txid,
|
|
&entry.token_id,
|
|
&format!("BCMR invalid JSON error: {e}"),
|
|
true,
|
|
)
|
|
.await
|
|
{
|
|
warn!("bcmr: Failed to set BCMR error: {e2}");
|
|
}
|
|
return;
|
|
}
|
|
};
|
|
|
|
let bcmr_parsed = match parse_bcmr_json(
|
|
&json,
|
|
&entry.token_id,
|
|
Some(actual_hash),
|
|
Some(bcmr.hash.to_hex()),
|
|
SOURCE_ON_CHAIN,
|
|
) {
|
|
Ok(b) => b,
|
|
Err(err) => {
|
|
if let Err(e2) = update_bcmr_failure(
|
|
db,
|
|
&entry.utxo,
|
|
&entry.txid,
|
|
&entry.token_id,
|
|
&format!("BCMR contents error: {err}"),
|
|
true,
|
|
)
|
|
.await
|
|
{
|
|
warn!("bcmr: Failed to set BCMR error: {e2}");
|
|
}
|
|
return;
|
|
}
|
|
};
|
|
|
|
if let Err(err) = insert_bcmr_data(db, &entry.token_id, &entry.utxo, &bcmr_parsed).await {
|
|
warn!("bcmr: Failed to insert BCMR data {err}")
|
|
}
|
|
}
|
|
|
|
impl BCMRDownloader {
|
|
pub fn new(db: SqlitePool) -> Self {
|
|
Self {
|
|
db,
|
|
keep_running: Arc::new(AtomicBool::new(true)),
|
|
download_task: None,
|
|
}
|
|
}
|
|
|
|
pub fn start(&mut self) -> Result<()> {
|
|
let db = self.db.clone();
|
|
let keep_running = self.keep_running.clone();
|
|
|
|
self.download_task = Some(tokio::spawn(async move {
|
|
let client = reqwest::Client::new();
|
|
|
|
loop {
|
|
if !keep_running.load(Ordering::Relaxed) {
|
|
info!("Exiting bcmr download task");
|
|
return;
|
|
}
|
|
|
|
let mut queue = get_download_candidates(&db).await;
|
|
|
|
// Shuffle so we don't starve any single token if the list is large
|
|
{
|
|
let mut rng = thread_rng();
|
|
queue.shuffle(&mut rng);
|
|
}
|
|
|
|
if queue.is_empty() {
|
|
tokio::time::sleep(LOOP_SLEEP_TIME).await;
|
|
continue;
|
|
}
|
|
|
|
info!("bcmr: {} tokens need BCMR download", queue.len());
|
|
|
|
// Process up to MAX_PARALLEL_DOWNLOADS concurrently
|
|
for chunk in queue.chunks(MAX_PARALLEL_DOWNLOADS) {
|
|
let futures: Vec<_> = chunk
|
|
.iter()
|
|
.map(|entry| process_entry(&client, &db, entry))
|
|
.collect();
|
|
futures::future::join_all(futures).await;
|
|
}
|
|
|
|
tokio::time::sleep(LOOP_SLEEP_TIME).await;
|
|
}
|
|
}));
|
|
|
|
Ok(())
|
|
}
|
|
}
|
|
|
|
impl Drop for BCMRDownloader {
|
|
fn drop(&mut self) {
|
|
self.keep_running
|
|
.store(false, std::sync::atomic::Ordering::SeqCst);
|
|
if let Some(task) = self.download_task.take() {
|
|
task.abort();
|
|
info!("bcmr download task aborted");
|
|
}
|
|
}
|
|
}
|
|
|
|
#[cfg(test)]
|
|
mod tests {
|
|
use super::*;
|
|
|
|
#[test]
|
|
fn bare_domain_resolves_to_well_known_then_root() {
|
|
let candidates = resolve_uri_candidates("chipnet-16.paryonusd.com");
|
|
assert_eq!(
|
|
candidates,
|
|
vec![
|
|
"https://chipnet-16.paryonusd.com/.well-known/bitcoin-cash-metadata-registry.json"
|
|
.to_string(),
|
|
"https://chipnet-16.paryonusd.com/".to_string(),
|
|
]
|
|
);
|
|
}
|
|
|
|
#[test]
|
|
fn bare_domain_with_scheme_and_trailing_slash() {
|
|
let candidates = resolve_uri_candidates("https://example.com/");
|
|
assert_eq!(
|
|
candidates,
|
|
vec![
|
|
"https://example.com/.well-known/bitcoin-cash-metadata-registry.json".to_string(),
|
|
"https://example.com/".to_string(),
|
|
]
|
|
);
|
|
}
|
|
|
|
#[test]
|
|
fn url_with_path_is_used_as_is() {
|
|
let candidates = resolve_uri_candidates("https://example.com/path/registry.json");
|
|
assert_eq!(
|
|
candidates,
|
|
vec!["https://example.com/path/registry.json".to_string()]
|
|
);
|
|
}
|
|
|
|
#[test]
|
|
fn schemeless_url_with_path_gets_https() {
|
|
let candidates = resolve_uri_candidates("example.com/registry.json");
|
|
assert_eq!(
|
|
candidates,
|
|
vec!["https://example.com/registry.json".to_string()]
|
|
);
|
|
}
|
|
|
|
#[test]
|
|
fn ipfs_uri_uses_a_gateway() {
|
|
let candidates = resolve_uri_candidates("ipfs://QmAbc123");
|
|
assert_eq!(candidates.len(), 1);
|
|
assert!(candidates[0].ends_with("/QmAbc123"));
|
|
assert!(IPFS_GATEWAYS.iter().any(|gw| candidates[0].starts_with(gw)));
|
|
}
|
|
|
|
#[test]
|
|
fn domain_with_port_is_treated_as_bare() {
|
|
let candidates = resolve_uri_candidates("example.com:8443");
|
|
assert_eq!(
|
|
candidates,
|
|
vec![
|
|
"https://example.com:8443/.well-known/bitcoin-cash-metadata-registry.json"
|
|
.to_string(),
|
|
"https://example.com:8443/".to_string(),
|
|
]
|
|
);
|
|
}
|
|
}
|