Bitcoin Core

test: add coverage for importdescriptors while wallet is rescanning

Co-authored-by: w0xlt <woltx@protonmail.com>

test: updated different_key to be different_field and also used a single assert_equal with 3 args instead of multiple assert_equals

iwyu: Fix warnings in `src/bench` and treat them as error

refactor: Drop unused includes after iwyu CI bump

Now that the C++ std-lib was bumped in the CI task in the prior commit,
some unused includes can (and must) be dropped.

ci: Bump APT_LLVM_V-based task configs to Ubuntu 26.04

Their clang version is pinned, so the only relevant change should be a
more recent cmake, more recent C++ stdlib, and more recent Python
version.

ci: Debian Trixie -> Ubuntu 26.04

net: move cs_main up in FetchBlock to fix rpc assert crash

1. FetchBlock runs in a http worker thread. It acquires a PeerRef,
locks cs_main, then may later call BlockRequested which asserts
that CNodeState exists for the peer.

2. FinalizeNode may run in either the bitcoind or b-net threads. It
locks cs_main, fetches a PeerRef from RemovePeer, fetches a CNodeState,
and later removes it from m_node_states.

Because of the lock placement in FetchBlock, the http worker thread in 1)
can acquire a valid PeerRef and block while the b-net thread in 2) is
cleaning up the peer in FinalizeNode. When the worker thread later acquires
cs_main, it may crash in BlockRequested since no CNodeState exists. Fix
this by acquiring the lock earlier in FetchBlock.

The lock can be replaced with a net-specific lock when the remaining
CNodeState fields are moved to Peer.

test: Use SteadyClockContext in pcp_tests

This is easier to reason about, because it will automatically take care
of properly setting INITIAL_MOCK_TIME in the ctor. Also, it allows to
drop the ElapseTime and replace it with a call to operator+=()

test: Use FakeNodeClock in more places

The context is easier to reason about: E.g.,

* in TestBasicMining it allows to drop manual SetMockTime(0) calls,
* in connections_desirable_service_flags it allows to drop manual calls
to SetMockTime(GetTime<std::chrono::seconds>() + _n_) and replace them
by operator+=(_n_)
* in wallet_tests it clarifies that the mocktime does not persist
outside the AddTx function

test: Add FakeNodeClock m_clock to TestChain100Setup

Currently, all test cases using TestChain100Setup or a derived class
like BuildChainTestingSetup are using mocktime by default due to the
SetMockTime call in the TestChain100Setup ctor.

This is confusing, because test cases using mocktime explicitly seem to
imply that before they set the mocktime, real time was used.

E.g. index_reorg_crash claimed in a comment to "Enable mock time".

Fix this issue by adding a FakeNodeClock m_clock field to
TestChain100Setup. Then, use the m_clock instead of explicit calls to
SetMockTime or to a (now) shadowing local FakeNodeClock variable.

bench, refactor: Use `std::string_view` for `BenchRunner` ctor parameter

This avoids implicit conversions from string literals to `std::string`,
eliminating the need to include `<string>` everywhere the `BENCHMARK()`
macro is used.

ci: use llvm 22.1.7

ci: skip libunwind runtime in LLVM build

Document why we use LIBCXXABI_USE_LLVM_UNWINDER=OFF.

scripted-diff: Rename UNIQUE_NAME to BITCOIN_UNIQUE_NAME

The `nb30.h` Windows header defines `UNIQUE_NAME` as a macro.

This introduces a fragile dependency on header inclusion order: if
Windows headers happen to be included before `UNIQUE_NAME` is used, the
preprocessor expands it into a numeric literal, causing syntax errors.

Rename the macro to `BITCOIN_UNIQUE_NAME` to remove this fragility and
avoid the collision entirely.

-BEGIN VERIFY SCRIPT-
sed -i 's/\<UNIQUE_NAME\>/BITCOIN_UNIQUE_NAME/g' $(git grep -l 'UNIQUE_NAME' ./src/)
-END VERIFY SCRIPT-

test: add coverage for importdescriptor with an encrypted wallet

blockstorage: Remove cs_LastBlockFile recursive mutex

The cs_LastBlockFile mutex is redundant: all critical sections are
already covered by cs_main. This is demonstrated in this patch by
replacing all instances of locking cs_LastBlockFile with pairs of
`AssertLockHeld(::cs_main)` and `EXCLUSIVE_LOCKS_REQUIRED(::cs_main)`
annotations. No additional `::cs_main` LOCK(...)s are introduced.

It is also not clear for which sections `cs_LastBlockFile` is
responsible for. It is annotated for `m_blockfile_cursors`, but
sporadically and inconsistently also covers `m_blockfile_info`.

Since it has no semantic meaning, and seems confusing to developers,
remove it.

test: Limit clocks to one active instance

SteadyClockContext and FakeNodeClock assume they are the only active
instance. Overlapping them in the same scope would silently clobber
each other.

Add a CRTP base class, LimitOne, that asserts at construction if
another instance already exists.

scripted-diff: Rename NodeClockContext to FakeNodeClock

The previous name did not indicate the type was intended for
testing. Renaming to FakeNodeClock makes this explicit and
allows call sites to drop the ctx suffix on the variable name.

Suggested in #34858 review feedback.

-BEGIN VERIFY SCRIPT-
s() { git grep -l "$1" -- src | xargs sed -i "s/$1/$2/g"; }

s '\<NodeClockContext\>' 'FakeNodeClock'
s '\<clock_ctx\>' 'clock'
-END VERIFY SCRIPT-

test: Use NodeClockContext in more call sites

This refactor is a follow-up to commit
faad08e59c4419e09eb75054bf468ca98a837ca8 and does not
change any behavior.

These call sites are clean mechanical swaps. The remaining ones
require non-trivial test refactoring and are left for future
follow-ups.

test: Drop ++ from NodeClockContext default constructor

The increment was originally added so that mocked time would not appear
to go backward relative to real-clock timestamps captured before
construction, since Now<NodeSeconds>() rounds the current time down to
a whole second. In practice the tests do not mix real and mocked
timestamps in a way that exposes this, so the increment is unnecessary.

test: Enter mocktime before peer creation in block_relay_only_eviction

This is a follow-up to commit faad08e59c4419e09eb75054bf468ca98a837ca8.
Hoisting the NodeClockContext above peer creation ensures m_connected is
captured under mocktime, making the MINIMUM_CONNECT_TIME check
deterministic regardless of which peer is selected for eviction.

This is a prerequisite for the next commit, which removes the
one-second advance from the NodeClockContext default constructor.

add release notes

test: check getprivatebroadcast and abortprivatebroadcast throw if the node is running without -privatebroadcast set

rpc: getprivatebroadcastinfo and abortprivatebroadcast throw if -privatebroadcast is disabled

fuzz: fix dead HD keypaths (de)serialization round-trip

`DeserializeHDKeypaths()` was writing into the original `hd_keypaths`
map instead of `deserialized_hd_keypaths`. As a result the latter was
always empty and the round-trip assertion following was trivially true,
so the serialize/deserialize round-trip wasn't actually being exercised.

That bug was introduced with the commit introducing the fuzz target
(commit f898ef65c947776750e49d050633f830546bbdc6, #18994).