Is there an existing issue for this?
I searched for existing issues and found #32276 / #32325, which looks related because it also involves cluster_linearize, MultiIntBitSet, GCC and array-bounds diagnostics.
However, this report is about cluster_linearize_tests failing in a normal Bitcoin Core v31.0 test build with GCC 16 on Arch Linux, not the Valgrind fuzz CI warning-as-error case addressed
there.
Current behaviour
When building Bitcoin Core v31.0 on Arch Linux with the current GCC toolchain, the unit test cluster_linearize_tests fails.
This was found while preparing an Arch Linux packaging merge request for unrelated packaging fixes:
https://gitlab.archlinux.org/archlinux/packaging/packages/bitcoin/-/merge_requests/4
- backporting the upstream Boost >= 1.91 compatibility patch
- adding missing
pythontocheckdepends
The Arch package maintainer suggested that an upstream commit would be preferred if this test issue needs to be addressed in the package.
Expected behaviour
cluster_linearize_tests should pass with the current GCC toolchain, as it does with GCC 15.2.1 in the same clean Arch Linux chroot setup.
Steps to reproduce
On Arch Linux, build Bitcoin Core v31.0 in a clean chroot with the current GCC toolchain and run the test suite.
In the Arch packaging setup this is done through pkgctl / clean chroot build, with tests enabled through the package check() function:
ctest --test-dir build
The relevant test is:
cluster_linearize_tests
Relevant log output
With the older GCC 15.2.1 toolchain, the same package build was tested without the local test patch. In that clean chroot run, cluster_linearize_tests passed and the full test suite completed successfully
With the current GCC 16 toolchain, cluster_linearize_tests fails. I locally tested the patch below, and with that patch the failing test passed again with the current toolchain.
I do not know whether this is the correct upstream fix; I am including it only as a data point and possible direction.
--- a/src/test/cluster_linearize_tests.cpp
+++ b/src/test/cluster_linearize_tests.cpp
@@ -56,7 +56,7 @@
void TestOptimalLinearization(std::span<const uint8_t> enc, std::initializer_list<DepGraphIndex> optimal_linearization)
{
- DepGraphIndex tx_count = 0;
+ DepGraphIndex position_range = 0;
FastRandomContext rng;
auto test_fn = [&]<typename SetType>() {
@@ -100,19 +100,22 @@
SanityCheck(depgraph, lin);
BOOST_CHECK(std::ranges::equal(lin, optimal_linearization));
}
- tx_count = depgraph.PositionRange();
+ position_range = depgraph.PositionRange();
};
- // Always run with 64-bit set types
- // - The native one that will be used on this platform.
- test_fn.template operator()<BitSet<64>>();
- // - The one used on 32-bit platforms.
- test_fn.template operator()<bitset_detail::MultiIntBitSet<uint32_t, 2>>();
- // - An 8-bit one, which is maximally different in terms of bitset behavior.
- test_fn.template operator()<bitset_detail::MultiIntBitSet<uint8_t, 8>>();
+ // Always run with a set type that can hold all encoded test clusters.
+ test_fn.template operator()<BitSet<256>>();
+
+ // Also run with 64-bit set types if the cluster doesn't use indexes above 63.
+ if (position_range <= 64) {
+ // - The native one that will be used on this platform.
+ test_fn.template operator()<BitSet<64>>();
+ // - The one used on 32-bit platforms.
+ test_fn.template operator()<bitset_detail::MultiIntBitSet<uint32_t, 2>>();
+ }
// Also run with 32-bit set types if the cluster doesn't use indexes above 31.
- if (tx_count <= 32) {
+ if (position_range <= 32) {
// - The native one that will be used on this platform.
test_fn.template operator()<BitSet<32>>();
// - An 8-bit one, which is maximally different in terms of bitset behavior.
How did you obtain Bitcoin Core
Arch Linux package build: https://gitlab.archlinux.org/archlinux/packaging/packages/bitcoin/-/merge_requests/4
What version of Bitcoin Core are you using?
Bitcoin Core v31.0.
Operating system and version
Arch Linux.
The failing setup uses the current Arch GCC 16 toolchain.
The counter-test used GCC 15.2.1 in the same clean chroot setup, where the unpatched test suite passed:
gcc 15.2.1+r604+g0b99615a8aef-1 gcc-libs 15.2.1+r604+g0b99615a8aef-1
Additional context
I am not a C++ developer and cannot fully judge whether the local patch above is semantically correct. It was generated with help from an LLM and then tested locally.
The important observations are:
- cluster_linearize_tests fails with the current GCC 16 toolchain on Arch Linux.
- The same test passes without the local patch when building with GCC 15.2.1.
- The local patch above makes the failing test pass with GCC 16.
- #32276 / #32325 look related, but appear to address a Valgrind fuzz CI array-bounds warning-as-error case rather than this normal unit-test failure.
Assisted-by: OpenAI Codex (GPT-5.5, reasoning: high)
