Repository navigation
chore: update deps async-hwi and bdk_esplora - #369
Conversation
- update async-hwi to 0.0.35: moves async-hwi from reqwest 0.11 to 0.13. That removes h2 0.3 (RUSTSEC-2026-0258), rustls-pemfile (unmaintained warning RUSTSEC-2025-0134), hyper 0.14, and that copies of rustls 0.21 and webpki-roots 0.25. It compiles with no code changes. - bdk_esplora: default-features = false: the default feature, blocking-https, pulls in minreq 2 with rustls 0.21 and webpki-roots 0.25. bdk-cli only uses the async client (async-https, tokio), so this was unused code carrying the old TLS stack. - targetted cargo update -p for quinn-ptoto to 0.11.19, rustls to 0.23.46, rustls-webpki 0.103.16, webpki-roots 1.0.11, zstd-sys 2.1.2 and tokio to 1.54.1
bdk_testenv 0.13.1's `wait_until_electrum_sees_block` returns on the first queued header notification, which can be a stale one left over from an earlier `mine_blocks` call. A following `full_scan` could then run before electrs had indexed the confirmation blocks, intermittently failing tests such as `test_create_tx_multiple_recipients` with a confirmed balance of 0. Add a `wait_for_electrum_tip` helper to `tests/common` that reads bitcoind's current block height and drains header notifications until electrs reports that height, and use it in place of all calls to `wait_until_electrum_sees_block` in the electrum and esplora online tests. The same bug is fixed upstream in bdk (a98cf7c4) but not yet released; the helper can be dropped once a new bdk_testenv is published.
Codecov Report✅ All modified and coverable lines are covered by tests. Additional details and impacted files@@ Coverage Diff @@
## master #369 +/- ##
=======================================
Coverage 59.08% 59.08%
=======================================
Files 25 25
Lines 4646 4646
=======================================
Hits 2745 2745
Misses 1901 1901
Flags with carried forward coverage won't be shown. Click here to find out more. ☔ View full report in Codecov by Harness. 🚀 New features to boost your workflow:
|
|
tACK d2317d2 Confirmed the 16 RUSTSEC advisories were addressed between branch point 66d5618 and d2317d2. I generated an audit diff to confirm, details below. Cargo audit diffdiff --git a/tmp/before.txt b/tmp/after.txt
index 332860e..2819cd3 100644
--- a/tmp/before.txt
+++ b/tmp/after.txt
@@ -1,26 +1,11 @@
crypto-bigint vuln RUSTSEC-2026-0387
-h2 vuln RUSTSEC-2026-0258
-quinn-proto vuln RUSTSEC-2026-0338
-quinn-proto vuln RUSTSEC-2026-0339
-quinn-proto vuln RUSTSEC-2026-0340
-quinn-proto vuln RUSTSEC-2026-0341
-quinn-proto vuln RUSTSEC-2026-0342
rustls vuln RUSTSEC-2026-0346
-rustls vuln RUSTSEC-2026-0346
-rustls vuln RUSTSEC-2026-0347
-rustls vuln RUSTSEC-2026-0348
-rustls vuln RUSTSEC-2026-0349
rustls vuln RUSTSEC-2026-0349
-rustls-pemfile warn RUSTSEC-2025-0134
rustls-webpki vuln RUSTSEC-2026-0098
rustls-webpki vuln RUSTSEC-2026-0099
rustls-webpki vuln RUSTSEC-2026-0104
rustls-webpki vuln RUSTSEC-2026-0382
-rustls-webpki vuln RUSTSEC-2026-0382
serde_cbor warn RUSTSEC-2021-0127
-tokio warn RUSTSEC-2026-0386
-webpki-root-certs vuln RUSTSEC-2026-0383
-webpki-roots vuln RUSTSEC-2026-0381
webpki-roots vuln RUSTSEC-2026-0381
zip vuln RUSTSEC-2026-0367
zip vuln RUSTSEC-2026-0368
@@ -34,4 +19,3 @@ zstd-safe warn RUSTSEC-2026-0401
zstd-safe warn RUSTSEC-2026-0402
zstd-safe warn RUSTSEC-2026-0404
zstd-safe warn RUSTSEC-2026-0406
-zstd-sys vuln RUSTSEC-2026-0410
cargo audit -q | awk '
/^Crate:/ { crate = $2; type = "vuln" }
/^Warning:/ { type = "warn" }
/^ID:/ { print crate, type, $2 }
' | sort # > /tmp/{before|after}.txtAll tests passed for each of the two commits, also when cherry-picking the test fix in isolation. With regards to:
I was unable to reproduce the intermittently failing tests in If anyone has advice on how to coax some of them into failing, I would appreciate pointers. |
vadim-anfv
left a comment
vadim-anfv
left a comment
There was a problem hiding this comment.
tACK d2317d2
Ran cargo audit before and after. It does not go green and cannot here: rustls, rustls-webpki and webpki-roots are each in the graph twice, this bumps the modern copy, and the old one still comes through minreq via bitcoincore-rpc and the bitcoind/electrsd harness. Tracked in #308.
I could not reproduce the flake either. At c5d3230, with the two waits in fund_and_sync_wallet printing the notification queue against bitcoind's tip, 15 runs of test_create_tx_multiple_recipients gave one notification each time, always at the tip. Not an objection to the fix. If you have a setup where it does fail, that would be worth having in the thread.
|
Thank you @bvotteler and @vadim-anfv for the reviews. I will setup an envt and see if I can reproduce the failure and share. Since it is non-blocking, I will merge the PR. Thank you. |
Description
Dependency updates (
chore: update deps async-hwi and bdk_esplora)async-hwi0.0.32 to 0.0.35. It moves from reqwest 0.11 to 0.13, which removes h2 0.3 (RUSTSEC-2026-0258),rustls-pemfile(unmaintained, RUSTSEC-2025-0134), hyper 0.14, and copies of rustls 0.21 / webpki-roots 0.25.default-features = falseonbdk_esplora. The defaultblocking-httpsfeature pulled in minreq 2 with rustls 0.21 / webpki-roots 0.25, but bdk-cli only uses the async client.cargo update -pfor patch releases:**fix test: wait for electrs to reach bitcoind's tip in online tests
bdk_testenv0.13.1'swait_until_electrum_sees_blockreturns on the first queued header notification, which can be stale from an earliermine_blockscall.full_scancould then run before the confirmation blocks were indexed, intermittently failing e.g.test_create_tx_multiple_recipientswith a confirmed balance of 0. A localwait_for_electrum_tiphelper waits until electrs reports bitcoind's current height. This is fixed upstream in bdk (a98cf7c4) but not released yet; the helper can be removed once it is.Notes to the reviewers
cargo update: a full update also bumpsbdk_wallet3.1 to 3.2, which introduces deprecation warnings.async-hwi0.0.35 compiles with no code changes.Checklists
All Submissions:
cargo fmtandcargo clippybefore committing