rust-lightning

Restrict CI build matrix to Linux+MSRV for PRs

Only run the full build matrix (Linux/Windows/macOS × stable/beta/MSRV)
on pushes to main. PR and non-main push builds now only run Linux with
the MSRV toolchain (1.75.0), which is the most important gate for
catching issues.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

Split method to reconstruct pending HTLCs into two

In the next commit, we want to dedup fields between the
InboundUpdateAdd::Forwarded's HTLCPreviousHopData and the outer
InboundHTLCOutput/Channel structs, since many fields are duplicated in both
places at the moment. As part of doing this cleanly, we first refactor the
method that retrieves these InboundUpdateAdds for reconstructing the set of
pending HTLCs during ChannelManager deconstruction.

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>

Dedup data in InboundUpdateAdd::Forwarded::hop_data

Previously, the InboundUpdateAdd::Forwarded enum variant contained an
HTLCPreviousHopData, which had a lot of fields that were redundant with the
outer InboundHTLCOutput/Channel structs. Here we dedup those fields, which is
important because the pending InboundUpdateAdds are persisted whenever the
ChannelManager is persisted.

Trivial: use full path in test macros

Useful when using these macros in lightning-tests/upgrade_downgrade_tests

Remove explicit usage of test_default_channel_config

test_default_channel_config is the default
now so it does not need to be set explicitly
in some of the tests.

Removes unnecessary extra None config.

Use handle_and_accept_open_channel in async_signer test

Update changelog and remove manually_accept references

Drop `proptest`s in `lightning-liquidity`

`proptest`'s transitive dependency tree has always been somewhat
large, but one of them (`rusty-fork`'s `tempfile` dependency) just
went ahead with a bump of their `rand` dependency, breaking our
MSRV yet again.

Because we don't actually use `proptest` for anything interesting,
the simplest solution is to simply drop it, which we do here. Note
that we'll likely transition the LSPS5 URL type to simply use the
`bitreq` URL type over the next few days anyway, so there's not
much reason to care about its continued test coverage.

Further, in writing this commit it was discovered that our tests in
`lsps2/utils.rs` were actually broken on the vast majority of
inputs, but proptest wasn't testing with any interesting test cases
at all, causing it to be missed entirely!

Bump `lightning-types` crate version to fix semver tests

5427b0de7e93ce4ccf63c79de756e0da49e33d0b changed the
`lightning-types` API but we forgot to bump the crate version to
make semver tests pass.

Pass the `addr` field of `tor_connect_outbound` to connection setup

When `setup_outbound` was used to setup a connection proxied over Tor,
it previously set the remote address of the peer to the address of the
Tor proxy.

This address of the Tor proxy was assigned to the
`PeerDetails::socket_address` for that peer in
`PeerManager::list_peers`, and if it was not a private IPv4 or IPv6
address, it was also reported to the peer in our init message.

This commit refactors `tor_connect_outbound` to pass its own peer
address parameter directly to the connection setup code.

This peer address will now appear in `PeerManager::list_peers` for
outbound Tor connections made using `tor_connect_outbound`, and will be
reported to the peer in our init message if it is not a private IPv4 or
IPv6 address.

Use `DirEntry::file_type` rather than `metadata...` in `list`

In the discussions at #3799 it was noted that `DirEntry::file_type`
will often use cached information rather than making a fresh
syscall, fixing the `list` race condition where we lose files while
iterating the directory for some filesystems on some Unix
platforms.

For some reason, that fix didn't make it into the merged PR, and we
rather stuck with `DirEntry::metadata()` which *always* does a
fresh syscall and always exhibits the problematic behavior. Here we
simply swap for `DirEntry::file_type` which at least fixes the
issue for "some filesome filesystems (among them: Btrfs, ext2,
ext3, and ext4)" (per `readdir(3)`).

Add a read closure to the `legacy` TLV variant

Update the `legacy` TLV read/write variant signature from
`(legacy, $fieldty, $write)` to `(legacy, $fieldty, $read, $write)`,
adding a read closure parameter matching the `custom` variant's
signature.

The read closure is applied in `_check_missing_tlv!` after all TLV
fields are read but before `static_value` fields consume legacy
values. This preserves backwards compatibility with `static_value` and
`default_value` expressions that reference legacy field variables as
`Option<$fieldty>` during TLV reading.

The read closure signature differs from `custom`:
`FnOnce(Option<&$fieldty>) -> Result<(), DecodeError>`. All existing
usage sites return `Ok(())` as their read closure (no-op).

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

Fix thread starvation in test_single_channel_multiple_mpp

The busy-wait loop polling for PaymentClaimed events had no yield,
causing it to continuously acquire ChannelManager locks via
get_and_clear_pending_events(). This could starve the claim_funds
thread of lock access, preventing it from ever queuing the event.
Add a yield_now() call matching the pattern used by the other two
spin loops in this test.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

Fix docs on ChannelMonitor::payment_preimages

Support deleting legacy forward map persistence in 0.5

In 0.3+, we are taking steps to remove the requirement of regularly persisting
the ChannelManager and instead rebuild the set of HTLC forwards (and the
manager generally) from Channel{Monitor} data.

We previously merged support for reconstructing the
ChannelManager::decode_update_add_htlcs map from channel data, using a new
HTLC onion field that will be present for inbound HTLCs received on 0.3+ only.
The plan is that in upcoming LDK versions, the manager will reconstruct this
map and the other forward/claimable/pending HTLC maps will automatically
repopulate themselves on the next call to process_pending_htlc_forwards.

As such, once we're in a future version that reconstructs the pending HTLC set,
we can stop persisting the legacy ChannelManager maps such as forward_htlcs,
pending_intercepted_htlcs since they will never be used.

For 0.3 to be compatible with this future version, in this commit we detect
that the manager was last written on a version of LDK that doesn't persist the
legacy maps. In that case, we don't try to read the old forwards map and run
the new reconstruction logic only.

Check pruned HTLCs were resolved on startup

In a recent commit, we added support for pruning an inbound HTLC's persisted
onion once the HTLC has been irrevocably forwarded to the outbound edge.

Here, we add a check on startup that those inbound HTLCs were actually handled.
Specifically, we check that the inbound HTLC is either (a) currently present in
the outbound edge or (b) was removed via claim. If neither of those are true,
we infer that the HTLC was removed from the outbound edge via fail and fail the
inbound HTLC backwards.

Trivially refactor reload_node macro

Cleans it up a bit in preparation for adding a new variant in the next commit.

Deterministic reconstruct_manager option in tests

We recently merged (test-only, for now) support for the ChannelManager
reconstructing its set of pending HTLCs from Channel{Monitor} data, rather than
using its own persisted maps. But because we want test coverage of both the new
reconstruction codepaths as well as the old persisted map codepaths,
in tests we would decide between those two sets of codepaths randomly.

We now want to add some tests that require using the new codepaths, so here we
add an option to explicitly set whether to reconstruct or not rather than
choosing randomly.

Prune inbound HTLC onions once forwarded

We store inbound committed HTLCs' onions in Channels for use in reconstructing
the pending HTLC set on ChannelManager read. If an HTLC has been forwarded to
the outbound edge, we no longer need to persist the inbound edge's onion and
can prune it here.

Mark legacy pre-0.3 inbound htlcs on persist

In 0.3+, we are taking steps to remove the requirement of regularly persisting
the ChannelManager and instead rebuild the set of HTLC forwards (and the
manager generally) from Channel{Monitor} data.

We previously merged support for reconstructing the
ChannelManager::decode_update_add_htlcs map from channel data, using a new
HTLC onion field that will be present for inbound HTLCs received on 0.3+ only.

However, we now want to add support for pruning this field once it's no longer
needed so it doesn't get persisted every time the manager gets persisted. At
the same time, in a future LDK version we need to detect whether the field was
ever present to begin with to prevent upgrading with legacy HTLCs present.

We accomplish both by converting the plain update_add option that was
previously serialized to an enum that can indicate whether the HTLC is from
0.2- versus 0.3+-with-onion-pruned (a variant for the latter is added in the
next commit).

Actual pruning of the new update_add field is added in the next commit.

Trivial: document some fields on MonitorRestoreUpdates

Don't double-forward inbounds resolved in holding cell

We recently added support for reconstructing
ChannelManager::decode_update_add_htlcs on startup, using data present in the
Channels. However, we failed to prune HTLCs from this rebuilt map if a given
HTLC was already forwarded+removed from the outbound edge and resolved in the
inbound edge's holding cell.

Here we fix this bug that would have caused us to
double-forward inbound HTLC forwards, which fortunately was not shipped.

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>

Simplify channel_closed check on manager read

This cleanup falls out of the changes made in the previous commit. Separated
out here for reviewability.

De-dup decode_htlcs from monitor only if channel is closed

We recently added support for reconstructing
ChannelManager::decode_update_add_htlcs on startup, using data present in the
Channels. However, we failed to prune HTLCs from this rebuilt map if a given
inbound HTLC was already forwarded to the outbound edge and in the outbound
holding cell (this bug could've caused us to double-forward HTLCs, fortunately
it never shipped).

As part of fixing this bug, we clean up the overall pruning approach by:
1. If the Channel is open, then it is the source of truth for what HTLCs are
outbound+pending (including pending in the holding cell)
2. If the Channel is closed, then the corresponding ChannelMonitor is the
source of truth for what HTLCs are outbound+pending

Previously, we would only consider the monitor's pending HTLCs, which ignored
holding cell HTLCs.

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>

fix: test_payment_path_scoring failing