fix(iroh): don't kill noq endpoint on first transport recv error #17800

Workflow file for this run

	name: CI

	on:
	pull_request:
	types: ["labeled", "unlabeled", "opened", "synchronize", "reopened"]
	merge_group:
	push:
	branches:
	- main

	concurrency:
	group: ci-${{ github.workflow }}-${{ github.event.pull_request.number \|\| github.ref }}
	cancel-in-progress: true

	env:
	RUST_BACKTRACE: 1
	RUSTFLAGS: -Dwarnings
	RUSTDOCFLAGS: -Dwarnings
	MSRV: "1.91"
	SCCACHE_CACHE_SIZE: "10G"
	IROH_FORCE_STAGING_RELAYS: "1"

	jobs:
	pick-runner:
	uses: "./.github/workflows/pick-runner.yml"
	with:
	min_free_runners: 3
	secrets:
	RUNNER_READ_TOKEN: ${{ secrets.RUNNER_READ_TOKEN }}

	tests:
	name: CI Test Suite
	if: "github.event_name != 'pull_request' \|\| ! contains(github.event.pull_request.labels.*.name, 'flaky-test')"
	uses: "./.github/workflows/tests.yaml"

	cross_build:
	name: Cross Build Only
	if: "github.event_name != 'pull_request' \|\| ! contains(github.event.pull_request.labels.*.name, 'flaky-test')"
	timeout-minutes: 30
	needs: pick-runner
	runs-on: ${{ fromJSON(needs.pick-runner.outputs.overflow_1) }}
	# runs-on: [self-hosted, linux, X64]
	env:
	RUSTFLAGS: "-Dwarnings --cfg skip_patchbay"
	strategy:
	fail-fast: false
	matrix:
	target:
	# cross tests are currently broken vor armv7 and aarch64
	# see https://github.com/cross-rs/cross/issues/1311
	# - armv7-linux-androideabi
	# - aarch64-linux-android
	# Freebsd execution fails in cross
	# - i686-unknown-freebsd # Linking fails :/
	- x86_64-unknown-freebsd
	# Netbsd execution fails to link in cross
	# - x86_64-unknown-netbsd
	steps:
	- name: Checkout
	uses: actions/checkout@v6
	with:
	submodules: recursive

	- name: Install rust stable
	uses: dtolnay/rust-toolchain@stable

	- name: Cleanup Docker
	continue-on-error: true
	run: \|
	docker kill $(docker ps -q)

	# See https://github.com/cross-rs/cross/issues/1222
	- name: Cache cross
	id: cache-cross
	uses: actions/cache@v5
	with:
	path: ~/.cargo/bin/cross
	key: cross-${{ runner.os }}-${{ runner.arch }}

	- name: Install cross
	if: steps.cache-cross.outputs.cache-hit != 'true'
	uses: taiki-e/install-action@cross

	- name: build
	# cross tests are currently broken vor armv7 and aarch64
	# see https://github.com/cross-rs/cross/issues/1311. So on
	# those platforms we only build but do not run tests.
	run: cross build --all --target ${{ matrix.target }}
	env:
	RUST_LOG: ${{ runner.debug && 'TRACE' \|\| 'DEBUG'}}
	CARGO_TARGET_X86_64_UNKNOWN_LINUX_GNU_LINKER: cc

	android:
	name: Android Build & Test
	if: "github.event_name != 'pull_request' \|\| ! contains(github.event.pull_request.labels.*.name, 'flaky-test')"
	timeout-minutes: 60
	# runs-on: [self-hosted, linux, X64]
	needs: pick-runner
	runs-on: ${{ fromJSON(needs.pick-runner.outputs.overflow_2) }}
	strategy:
	fail-fast: false
	matrix:
	include:
	- target: aarch64-linux-android
	- target: armv7-linux-androideabi
	- target: x86_64-linux-android
	api-level: 35
	steps:
	- name: Checkout
	uses: actions/checkout@v6

	- name: Enable KVM for nested virtualization
	if: matrix.target == 'x86_64-linux-android'
	run: \|
	echo 'KERNEL=="kvm", GROUP="kvm", MODE="0666"' \
	\| sudo tee /etc/udev/rules.d/99-kvm4all.rules
	sudo udevadm control --reload-rules
	sudo udevadm trigger --name-match=kvm

	- name: Set up Rust
	uses: dtolnay/rust-toolchain@stable
	with:
	target: ${{ matrix.target }}

	- name: Setup Java
	uses: actions/setup-java@v5
	with:
	distribution: "temurin"
	java-version: "17"

	- name: Setup Android SDK
	uses: android-actions/setup-android@v4

	# Pinned to a release tag (rather than @main) for supply-chain hygiene.
	- name: Setup Android NDK
	uses: nttld/setup-ndk@v1
	id: setup-ndk
	with:
	ndk-version: r25c
	add-to-path: true

	- name: Install cargo-ndk
	run: cargo install --version 4.1.2 cargo-ndk

	# Replace cargo-ndk-runner with a wrapper that forwards RUST_LOG
	# and RUST_BACKTRACE into the emulator shell. The upstream binary
	# does `adb shell <bin>` without inheriting any host env.
	- name: Wrap cargo-ndk-runner to forward logging env
	if: matrix.target == 'x86_64-linux-android'
	run: \|
	cat > "$(which cargo-ndk-runner)" <<'WRAPPER'
	#!/bin/sh
	set -e
	dev=/data/local/tmp/$(basename "$1")
	adb push "$1" "$dev" >/dev/null
	adb shell chmod 755 "$dev"
	shift
	adb shell "RUST_LOG=$RUST_LOG RUST_BACKTRACE=$RUST_BACKTRACE $dev $*"
	WRAPPER
	chmod +x "$(which cargo-ndk-runner)"

	- name: Build
	if: matrix.target != 'x86_64-linux-android'
	env:
	ANDROID_NDK_HOME: ${{ steps.setup-ndk.outputs.ndk-path }}
	run: cargo ndk --target ${{ matrix.target }} build

	# Build all the test binaries up front so the emulator boot can
	# overlap with compilation. cargo-ndk runs `cargo test --no-run`
	# against each crate; the binaries land in
	# target/$CARGO_NDK_TARGET/debug/deps/.
	- name: Build workspace test binaries
	if: matrix.target == 'x86_64-linux-android'
	env:
	ANDROID_NDK_HOME: ${{ steps.setup-ndk.outputs.ndk-path }}
	CARGO_NDK_TARGET: ${{ matrix.target }}
	run: \|
	cargo ndk test --no-run -p iroh-base --all-features
	cargo ndk test --no-run -p iroh-dns --features tls-ring
	cargo ndk test --no-run -p iroh-relay --features tls-ring,metrics
	cargo ndk test --no-run -p iroh --features tls-ring,metrics,portmapper,test-utils

	- name: Run workspace tests on Android emulator
	if: matrix.target == 'x86_64-linux-android'
	uses: reactivecircus/android-emulator-runner@v2
	env:
	ANDROID_NDK_HOME: ${{ steps.setup-ndk.outputs.ndk-path }}
	# Forwarded into the emulator by the cargo-ndk-runner wrapper.
	RUST_LOG: trace
	RUST_BACKTRACE: "1"
	CARGO_NDK_TARGET: ${{ matrix.target }}
	with:
	api-level: ${{ matrix.api-level }}
	arch: x86_64
	target: google_apis
	force-avd-creation: false
	# `-dns-server` + `-netfast` + `-no-metrics` is the combination reported to give
	# working public-internet connectivity on the GH-hosted runner; without them wlan0
	# stays NO-CARRIER and every connect() to a public IP fails with ENETUNREACH.
	# See https://github.com/ReactiveCircus/android-emulator-runner/issues/348#issuecomment-2578082030
	emulator-options: -no-window -no-audio -no-boot-anim -gpu swiftshader_indirect -dns-server 8.8.8.8 -netfast -no-metrics
	disable-animations: true
	script: \|
	adb wait-for-device
	# Wait for full boot and a working outbound TCP. `sys.boot_completed=1` fires
	# before the radio/Wi-Fi stack comes up, and `ip route get 8.8.8.8` reports
	# success while the kernel's per-network fwmark route is still missing. Probe
	# an actual TCP connect to 8.8.8.8:53 instead - it exercises the same path
	# the test sockets will use.
	adb shell 'i=0; while [ -z "$(getprop sys.boot_completed \| tr -d "\r")" ]; do i=$((i+1)); if [ $i -gt 300 ]; then exit 1; fi; sleep 2; done'
	adb shell 'i=0; while ! nc -w 2 8.8.8.8 53 </dev/null >/dev/null 2>&1; do i=$((i+1)); if [ $i -gt 30 ]; then exit 1; fi; sleep 2; done'
	# Disable IPv6: the emulator has site-local + link-local v6 but no global v6
	# route, so connect() to a global v6 address hangs instead of returning ENETUNREACH.
	adb root && adb wait-for-device
	adb shell sysctl -w net.ipv6.conf.all.disable_ipv6=1 \|\| true
	adb shell sysctl -w net.ipv6.conf.default.disable_ipv6=1 \|\| true
	adb shell sysctl -w net.ipv6.conf.lo.disable_ipv6=1 \|\| true
	adb shell ip addr
	cargo ndk test -p iroh-base --all-features
	cargo ndk test -p iroh-dns --features tls-ring
	cargo ndk test -p iroh-relay --features tls-ring,metrics
	cargo ndk test -p iroh --features tls-ring,metrics,portmapper,test-utils

	cross_test:
	name: Cross Test
	if: "github.event_name != 'pull_request' \|\| ! contains(github.event.pull_request.labels.*.name, 'flaky-test')"
	timeout-minutes: 30
	needs: pick-runner
	runs-on: ${{ fromJSON(needs.pick-runner.outputs.overflow_3) }}
	# runs-on: [self-hosted, linux, X64]
	strategy:
	fail-fast: false
	matrix:
	target:
	- i686-unknown-linux-gnu
	env:
	RUSTFLAGS: "-Dwarnings --cfg skip_patchbay"
	steps:
	- name: Checkout
	uses: actions/checkout@v6
	with:
	submodules: recursive

	- name: Install rust stable
	uses: dtolnay/rust-toolchain@stable

	- name: Cleanup Docker
	continue-on-error: true
	run: \|
	docker kill $(docker ps -q)

	# See https://github.com/cross-rs/cross/issues/1222
	- name: Cache cross
	id: cache-cross
	uses: actions/cache@v5
	with:
	path: ~/.cargo/bin/cross
	key: cross-${{ runner.os }}-${{ runner.arch }}

	- name: Install cross
	if: steps.cache-cross.outputs.cache-hit != 'true'
	uses: taiki-e/install-action@cross

	- name: test
	run: cross test --all --target ${{ matrix.target }} -- --test-threads=12
	env:
	RUST_LOG: ${{ runner.debug && 'TRACE' \|\| 'DEBUG' }}
	CARGO_TARGET_X86_64_UNKNOWN_LINUX_GNU_LINKER: cc

	wasm_test:
	name: Build & test wasm32 for browsers
	needs: pick-runner
	runs-on: ${{ fromJSON(needs.pick-runner.outputs.overflow_4) }}
	# runs-on: [self-hosted, linux, X64]
	env:
	RUSTFLAGS: '--cfg getrandom_backend="wasm_js"'
	steps:
	- name: Checkout sources
	uses: actions/checkout@v6

	- name: Setup node.js version 22.5 # needed for browser-like websocket API support in node.js
	uses: actions/setup-node@v6
	with:
	node-version: 22.5

	- name: Install stable toolchain
	uses: dtolnay/rust-toolchain@stable

	- name: Add wasm target
	run: rustup target add wasm32-unknown-unknown

	- name: Install cargo-binstall
	uses: cargo-bins/cargo-binstall@v1.18.1

	- name: Install wasm-bindgen-test-runner
	# Match the version of wasm-bindgen used in Cargo.lock
	run: cargo binstall wasm-bindgen-cli@0.2.122 --locked --no-confirm

	- name: Install wasm-tools
	uses: bytecodealliance/actions/wasm-tools/setup@v1

	- name: wasm32 build (iroh-base)
	run: cargo build --target wasm32-unknown-unknown -p iroh-base --all-features

	- name: wasm32 build (iroh-relay)
	run: cargo build --target wasm32-unknown-unknown -p iroh-relay

	- name: wasm32 build (iroh)
	run: cargo build --target wasm32-unknown-unknown -p iroh

	# If the Wasm file contains any 'import "env"' declarations, then
	# some non-Wasm-compatible code made it into the final code.
	- name: Ensure no 'import "env"' in iroh-relay Wasm
	run: \|
	! wasm-tools print --skeleton target/wasm32-unknown-unknown/debug/iroh_relay.wasm \| grep 'import "env"'

	- name: Ensure no 'import "env"' in iroh Wasm
	run: \|
	! wasm-tools print --skeleton target/wasm32-unknown-unknown/debug/iroh.wasm \| grep 'import "env"'

	- name: Run integration test in wasm
	run: cargo test -p iroh --test integration --target=wasm32-unknown-unknown

	check_semver:
	needs: pick-runner
	runs-on: ${{ fromJSON(needs.pick-runner.outputs.overflow_5) }}
	# runs-on: [self-hosted, linux, X64]
	env:
	RUSTC_WRAPPER: "sccache"
	SCCACHE_GHA_ENABLED: "on"
	steps:
	- uses: actions/checkout@v6
	with:
	fetch-depth: 0
	- name: Install sccache
	uses: mozilla-actions/sccache-action@v0.0.10
	continue-on-error: true
	- uses: ./.github/workflows/sccache-probe

	- name: Setup Environment (PR)
	if: ${{ github.event_name == 'pull_request' }}
	shell: bash
	run: \|
	echo "HEAD_COMMIT_SHA=$(git rev-parse origin/${{ github.base_ref }})" >> ${GITHUB_ENV}
	- name: Setup Environment (Push)
	if: ${{ github.event_name == 'push' \|\| github.event_name == 'merge_group' }}
	shell: bash
	run: \|
	echo "HEAD_COMMIT_SHA=$(git rev-parse origin/main)" >> ${GITHUB_ENV}
	- name: Check semver
	# uses: obi1kenobi/cargo-semver-checks-action@v2
	uses: n0-computer/cargo-semver-checks-action@feat-baseline
	with:
	package: iroh, iroh-base, iroh-dns-server, iroh-relay, iroh-net-report
	baseline-rev: ${{ env.HEAD_COMMIT_SHA }}
	use-cache: false

	check_external_types:
	timeout-minutes: 30
	name: Check external types in public API
	if: "github.event_name != 'pull_request' \|\| ! contains(github.event.pull_request.labels.*.name, 'flaky-test')"
	needs: pick-runner
	runs-on: ${{ fromJSON(needs.pick-runner.outputs.overflow_9) }}
	# runs-on: [self-hosted, linux, X64]
	env:
	RUSTC_WRAPPER: "sccache"
	SCCACHE_GHA_ENABLED: "on"
	# The tool drives an internal rustdoc invocation that only emits JSON
	# for the analyzer, so the workflow-wide `-Dwarnings` would just turn
	# unrelated rustdoc lints (e.g. `private-intra-doc-links`) into hard
	# failures here. Clear it for this job.
	RUSTDOCFLAGS: ""
	# Pin to the nightly that the pinned `cargo-check-external-types`
	# release was last tested against. Update both together.
	CARGO_CHECK_EXTERNAL_TYPES_VERSION: "0.4.0"
	TOOLCHAIN: "nightly-2025-10-18"
	steps:
	- uses: actions/checkout@v6
	- uses: dtolnay/rust-toolchain@master
	with:
	toolchain: ${{ env.TOOLCHAIN }}
	- name: Install sccache
	uses: mozilla-actions/sccache-action@v0.0.9

	- name: Install cargo-binstall
	uses: cargo-bins/cargo-binstall@v1.18.1

	- uses: taiki-e/install-action@cargo-make

	- name: Install cargo-check-external-types
	run: cargo binstall cargo-check-external-types@${{ env.CARGO_CHECK_EXTERNAL_TYPES_VERSION }} --locked --no-confirm

	- name: Check external types
	run: cargo make check-external-types

	check_fmt:
	timeout-minutes: 30
	name: Checking fmt
	runs-on: ubuntu-latest
	env:
	RUSTC_WRAPPER: "sccache"
	SCCACHE_GHA_ENABLED: "on"
	steps:
	- uses: actions/checkout@v6
	- uses: dtolnay/rust-toolchain@stable
	with:
	components: rustfmt
	- uses: mozilla-actions/sccache-action@v0.0.10
	continue-on-error: true
	- uses: ./.github/workflows/sccache-probe
	- uses: taiki-e/install-action@cargo-make
	- run: cargo make format-check

	check_docs:
	timeout-minutes: 30
	name: Checking docs
	needs: pick-runner
	runs-on: ${{ fromJSON(needs.pick-runner.outputs.overflow_6) }}
	# runs-on: [self-hosted, linux, X64]
	env:
	RUSTC_WRAPPER: "sccache"
	SCCACHE_GHA_ENABLED: "on"
	RUSTDOCFLAGS: --cfg docsrs
	steps:
	- uses: actions/checkout@v6
	- uses: dtolnay/rust-toolchain@master
	with:
	toolchain: nightly-2025-10-09
	- name: Install sccache
	uses: mozilla-actions/sccache-action@v0.0.10
	continue-on-error: true
	- uses: ./.github/workflows/sccache-probe

	- name: Docs
	run: cargo doc --workspace --all-features --no-deps --document-private-items

	clippy_check:
	timeout-minutes: 30
	needs: pick-runner
	runs-on: ${{ fromJSON(needs.pick-runner.outputs.overflow_7) }}
	# runs-on: [self-hosted, linux, X64]
	env:
	RUSTC_WRAPPER: "sccache"
	SCCACHE_GHA_ENABLED: "on"
	steps:
	- uses: actions/checkout@v6
	- uses: dtolnay/rust-toolchain@stable
	with:
	components: clippy
	- name: Install sccache
	uses: mozilla-actions/sccache-action@v0.0.10
	continue-on-error: true
	- uses: ./.github/workflows/sccache-probe

	# TODO: We have a bunch of platform-dependent code so should
	# probably run this job on the full platform matrix
	- name: clippy check (all features)
	run: cargo clippy --workspace --all-features --all-targets --lib --bins --tests --benches --examples

	- name: clippy check (no features)
	run: cargo clippy --workspace --no-default-features --all-targets --lib --bins --tests --benches --examples

	- name: clippy check (default features)
	run: cargo clippy --workspace --all-targets --lib --bins --tests --benches --examples

	check_optional_deps:
	timeout-minutes: 10
	name: Check optional deps are absent without features
	runs-on: ubuntu-latest
	steps:
	- uses: actions/checkout@v6
	- uses: dtolnay/rust-toolchain@stable
	- name: Verify optional deps are absent without features
	run: \|
	fail=0
	for dep in portmapper igd-next swarm-discovery ring aws-lc-rs; do
	echo "cargo tree -p iroh --no-default-features --edges normal --target all -i $dep";
	output=$(cargo tree -p iroh --no-default-features --edges normal --target all -i "$dep" 2>&1) \|\| true
	if ! echo "$output" \| grep -qE "(nothing to print\|did not match any packages)"; then
	echo "ERROR: $dep should not be in the dependency tree without its feature"
	echo "$output"
	fail=1
	fi
	done
	exit $fail

	msrv:
	if: "github.event_name != 'pull_request' \|\| ! contains(github.event.pull_request.labels.*.name, 'flaky-test')"
	timeout-minutes: 30
	name: Minimal Supported Rust Version
	needs: pick-runner
	runs-on: ${{ fromJSON(needs.pick-runner.outputs.overflow_8) }}
	# runs-on: [self-hosted, linux, X64]
	env:
	RUSTC_WRAPPER: "sccache"
	SCCACHE_GHA_ENABLED: "on"
	steps:
	- uses: actions/checkout@v6
	- uses: dtolnay/rust-toolchain@master
	with:
	toolchain: ${{ env.MSRV }}
	- name: Install sccache
	uses: mozilla-actions/sccache-action@v0.0.10
	continue-on-error: true
	- uses: ./.github/workflows/sccache-probe

	- name: Check MSRV all features
	run: \|
	cargo +$MSRV check --workspace --all-targets

	cargo_deny:
	timeout-minutes: 30
	name: cargo deny
	runs-on: ubuntu-latest
	steps:
	- uses: actions/checkout@v6
	- uses: EmbarkStudios/cargo-deny-action@v2
	with:
	arguments: --workspace --all-features
	command: check
	command-arguments: "-Dwarnings"

	netsim-integration-tests:
	permissions: write-all
	if: "github.event_name != 'pull_request' \|\| ! contains(github.event.pull_request.labels.*.name, 'flaky-test')"
	uses: "./.github/workflows/netsim_runner.yaml"
	secrets: inherit
	with:
	branch: ${{ github.ref }}
	max_workers: 6
	netsim_branch: "main"
	sim_paths: ${{ github.event_name == 'pull_request' && 'sims/integration' \|\| 'sims/iroh/iroh.json,sims/integration' }}
	pr_number: ${{ github.event.pull_request.number \|\| '' }}
	filter: ${{ github.event_name == 'pull_request' && 'skip:intg_iroh_full__1_to_3,intg_iroh_full__1_to_1x3,intg_iroh_relay_only__1_to_3' \|\| '' }}

	codespell:
	timeout-minutes: 30
	runs-on: ubuntu-latest
	steps:
	- uses: actions/checkout@v6
	- name: Cache pip
	uses: actions/cache@v5
	with:
	path: ~/.cache/pip
	key: pip-codespell-2.4.1
	- run: pip install --user codespell[toml]==2.4.1
	- run: codespell --ignore-words-list=ans,atmost,crate,inout,ratatui,ser,stayin,swarmin,worl,keep-alives --skip=CHANGELOG.md

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

fix(iroh): don't kill noq endpoint on first transport recv error #17800

Workflow file

fix(iroh): don't kill noq endpoint on first transport recv error #17800

Uh oh!

Workflow file for this run