]> git.bitcoin.ninja Git - rust-lightning/log
rust-lightning
6 months agoMove serialization of `channel_state` fields to `channel_state.rs` 2024-06-channel_state-module-123
Matt Corallo [Tue, 4 Jun 2024 14:46:15 +0000 (14:46 +0000)]
Move serialization of `channel_state` fields to `channel_state.rs`

1f616c0058f77e074c429fa7fb8b2f4594cdb9ad moved a handful of structs
to the new `channel_state.rs` but forgot to move their
serialization logic with them, which is corrected here.

6 months agoMove the public path for `channel_state` types to their new path
Matt Corallo [Tue, 4 Jun 2024 01:25:52 +0000 (01:25 +0000)]
Move the public path for `channel_state` types to their new path

6 months agoReformat `channel_state.rs` to match rustfmt requirements
Matt Corallo [Tue, 4 Jun 2024 01:13:40 +0000 (01:13 +0000)]
Reformat `channel_state.rs` to match rustfmt requirements

6 months agoMove the public channel state API into a new module
Matt Corallo [Tue, 4 Jun 2024 01:06:54 +0000 (01:06 +0000)]
Move the public channel state API into a new module

Our "what is the channel's current state" structs have slowly
grown to be rather nontrivial, and now include eight structs with
many fields. Thus, it makes sense to pull them out of
`ln::channelmanager` and into their own module.

This also makes things easier for the C bindings which don't
support `pub use` from a private module.

6 months agoMerge pull request #3062 from TheBlueMatt/2024-04-123-bindings
Matt Corallo [Tue, 14 May 2024 21:56:44 +0000 (14:56 -0700)]
Merge pull request #3062 from TheBlueMatt/2024-04-123-bindings

[0.0.123-bindings] Bindings changes for 0.0.123

6 months agoMove `Persister` back to lots of generic bounds
Matt Corallo [Sat, 11 May 2024 16:01:32 +0000 (16:01 +0000)]
Move `Persister` back to lots of generic bounds

...as we currently struggle with `AChannelManager` on it

6 months agoDrop `SerialId` type as bindings don't support primitive aliasing
Matt Corallo [Sat, 11 May 2024 16:13:02 +0000 (16:13 +0000)]
Drop `SerialId` type as bindings don't support primitive aliasing

6 months agoHide `Direction::select_node_id` due to lifetimes
Matt Corallo [Sat, 11 May 2024 16:12:39 +0000 (16:12 +0000)]
Hide `Direction::select_node_id` due to lifetimes

6 months agoAvoid slices without inner references
Matt Corallo [Thu, 28 Sep 2023 03:05:09 +0000 (03:05 +0000)]
Avoid slices without inner references

As we cannot express slices without inner references in bindings
wrappers.

6 months agoReplace `EventsProvider` on `OnionMessageHandler` with a single fn
Matt Corallo [Sun, 17 Dec 2023 19:43:28 +0000 (19:43 +0000)]
Replace `EventsProvider` on `OnionMessageHandler` with a single fn

`OnionMessageHandler`s are expected to regularly release a set of
nodes which we need to directly connect to to deliver onion
messages. In order to do so, they currently extend
`EventsProvider`, returning that set as `Event::ConnectionNeeded`s.

While this works fine in Rust, the `EventsProvider` interface
doesn't map well in bindings due to it taking a flexible trait impl
as a method argument.

Instead, here, we convert `OnionMessageHandler` to include a single
function which returns nodes in the form of a `node_id` and
`Vec<SocketAddress>`. This is a bit simpler, if less flexible, API,
and while largely equivalent, is easier to map in bindings.

6 months agoMake the custom message traits cloneable as they're deep in nested structs
Matt Corallo [Fri, 24 Sep 2021 18:44:32 +0000 (18:44 +0000)]
Make the custom message traits cloneable as they're deep in nested structs

6 months agoMake ChannelMonitor always clonable
Matt Corallo [Mon, 1 Feb 2021 01:12:50 +0000 (20:12 -0500)]
Make ChannelMonitor always clonable

Rather than `ChannelMonitor` only being clonable when the signer is
clonable, we require all signers to be clonable and then make all
`ChannelMonitor`s clonable.

6 months ago`crate`-only several BOLT12 methods that require unbounded generics
Matt Corallo [Thu, 28 Sep 2023 03:03:38 +0000 (03:03 +0000)]
`crate`-only several BOLT12 methods that require unbounded generics

These are not expressible in C/most languages, and thus must be
hidden.

6 months agoMark several types no-export which should be exported eventually
Matt Corallo [Wed, 19 Jul 2023 20:09:23 +0000 (20:09 +0000)]
Mark several types no-export which should be exported eventually

6 months agoHard-code scorer parameters to `ProbabilisticScoringFeeParameters`
Matt Corallo [Sat, 21 Oct 2023 02:28:53 +0000 (02:28 +0000)]
Hard-code scorer parameters to `ProbabilisticScoringFeeParameters`

The scorer currently relies on an associated type for the fee
parameters. This isn't supportable at all in bindings, and for
lack of a better option we simply hard-code the parameter for all
scorers  to `ProbabilisticScoringFeeParameters`.

6 months agoUse `[u8; 3]` not `CurrencyCode` to skip type aliases to primitives
Matt Corallo [Sat, 11 May 2024 20:05:26 +0000 (20:05 +0000)]
Use `[u8; 3]` not `CurrencyCode` to skip type aliases to primitives

6 months agoMark a few offers fields no-export as we have no mapping for them
Matt Corallo [Sat, 11 May 2024 20:05:15 +0000 (20:05 +0000)]
Mark a few offers fields no-export as we have no mapping for them

6 months agoAvoid enums containing references with lifetimes
Matt Corallo [Sun, 5 Mar 2023 20:38:42 +0000 (20:38 +0000)]
Avoid enums containing references with lifetimes

Having struct fields with references to other structs is tough in
our bindings logic, but even worse if the fields are in an enum.
Its simplest to just take the clone penalty here.

6 months agoExport `outbound_payment` structs in their respective modules
Matt Corallo [Tue, 28 Feb 2023 21:45:14 +0000 (21:45 +0000)]
Export `outbound_payment` structs in their respective modules

Re-exports in Rust make `use` statements a little shorter, but for
otherwise don't materially change a crate's API. Sadly, the C
bindings generator currently can't figure out re-exports, but it
also exports everything into one global namespace, so it doesn't
matter much anyway.

6 months agoUse an explicit `Sign` type on the `ChannelMonitor` read tuple
Matt Corallo [Sat, 24 Dec 2022 04:16:48 +0000 (04:16 +0000)]
Use an explicit `Sign` type on the `ChannelMonitor` read tuple

The bindings currently get confused by the implicit `Sign` type, so
we temporarily remove it on the `impl` here.

6 months agoRestrict ChannelInfo::as_directed_from visibility
Jeffrey Czyz [Tue, 29 Mar 2022 15:20:39 +0000 (10:20 -0500)]
Restrict ChannelInfo::as_directed_from visibility

Bindings can't handle references in return types, so reduce the
visibility to pub(crate).

6 months agoMake `as_directed_to` non-public
Matt Corallo [Tue, 1 Mar 2022 03:46:52 +0000 (03:46 +0000)]
Make `as_directed_to` non-public

...as the bindings generation does not currently have the ability
to map a reference to a `NodeId` inside a tuple.

6 months agoStop relying on a `Clone`able `NetworkGraph` ref in `DefaultRouter`
Matt Corallo [Tue, 23 Jan 2024 19:55:24 +0000 (19:55 +0000)]
Stop relying on a `Clone`able `NetworkGraph` ref in `DefaultRouter`

While there's not really much harm in requiring a `Clone`able
reference (they almost always are), it does make our bindings
struggle a bit as they don't support multi-trait bounds (as it
would require synthesizing a new C trait, which the bindings don't
do automatically). Luckily, there's really no reason for it, and we
can just call the `DefaultMessageRouter` directly when we want to
route a message.

6 months ago[UPSTREAM] Allow any `Deref` to an `EntropySource` in `BlindedPath`
Matt Corallo [Mon, 13 May 2024 18:01:20 +0000 (18:01 +0000)]
[UPSTREAM] Allow any `Deref` to an `EntropySource` in `BlindedPath`

This matches our normal API semantics and allows, for example,
`Arc`s to `EntropySource`s.

6 months ago[UPSTREAM] impl `Clone` on BOLT12 builders in bindings
Matt Corallo [Mon, 13 May 2024 14:39:19 +0000 (14:39 +0000)]
[UPSTREAM] impl `Clone` on BOLT12 builders in bindings

This is required for GC'd languages which need to be able to clone
in order to pass owned objects to Rust.

6 months ago[UPSTREAM] impl `Clone` on unsigned BOLT12 types
Matt Corallo [Mon, 13 May 2024 14:36:01 +0000 (14:36 +0000)]
[UPSTREAM] impl `Clone` on unsigned BOLT12 types

This is required for bindings as passing types from Rust to GC'd
languages can't map the concept of a type that has a lifetime of
the called function but instead needs to clone for safety.

6 months ago[UPSTREAM] Expose `ChannelManager` offer constructors in bindings
Matt Corallo [Mon, 13 May 2024 14:35:34 +0000 (14:35 +0000)]
[UPSTREAM] Expose `ChannelManager` offer constructors in bindings

6 months ago[UPSTREAM] Mark `io_extras` as no-bindings-export as its not pub
Matt Corallo [Sun, 12 May 2024 00:26:19 +0000 (00:26 +0000)]
[UPSTREAM] Mark `io_extras` as no-bindings-export as its not pub

6 months ago[UPSTREAM] Make `offers::Amount` `Copy` and export it in bindings
Matt Corallo [Sat, 11 May 2024 18:51:48 +0000 (18:51 +0000)]
[UPSTREAM] Make `offers::Amount` `Copy` and export it in bindings

`Amount` is less than two pointers long, so there's no reason it
shouldn't be `Copy`. Further, because its an enum, bindings can't
map a reference to it in an `Option`. Thus, here, we simply make it
`Copy` and return it in `Option`s rather than a reference to it.

6 months ago[UPSTREAM] Use prelude by universal import, rather than single import
Matt Corallo [Sat, 11 May 2024 15:55:36 +0000 (15:55 +0000)]
[UPSTREAM] Use prelude by universal import, rather than single import

This avoids the bindings trying to figure out what a
`lightning::prelude::Vec` is rather than matching it as a `Vec`.

6 months ago[UPSTREAM] Export `ChannelId` in bindings rather than `[u8; 32]`
Matt Corallo [Sat, 11 May 2024 16:16:12 +0000 (16:16 +0000)]
[UPSTREAM] Export `ChannelId` in bindings rather than `[u8; 32]`

Now that `ChannelId` has useful constructors and methods we need
to start exposing it to bindings users rather than relying on using
`[u8; 32]`.

6 months ago[UPSTREAM] Skip `BestBlock::new` in bindings as we already have one
Matt Corallo [Sat, 11 May 2024 16:10:56 +0000 (16:10 +0000)]
[UPSTREAM] Skip `BestBlock::new` in bindings as we already have one

6 months agoMerge pull request #3055 from TheBlueMatt/2024-05-123-backports-2 v0.0.123
Matt Corallo [Thu, 9 May 2024 00:12:17 +0000 (17:12 -0700)]
Merge pull request #3055 from TheBlueMatt/2024-05-123-backports-2

Final backports for 0.0.123

6 months agoBump to release versions, dropping -rc1 2024-05-123-backports-2
Matt Corallo [Wed, 8 May 2024 20:56:59 +0000 (20:56 +0000)]
Bump to release versions, dropping -rc1

6 months agoBump `possiblyrandom` to 0.2
Matt Corallo [Fri, 19 Apr 2024 15:33:58 +0000 (15:33 +0000)]
Bump `possiblyrandom` to 0.2

I'd uploaded `possiblyrandom` 0.1 to reserve the crate name but
with a dummy library so now we need ot use 0.2 for the release.

6 months agoAdd changelog entries for 0.0.123
Matt Corallo [Fri, 19 Apr 2024 15:24:32 +0000 (15:24 +0000)]
Add changelog entries for 0.0.123

6 months agoKeep newlines between releases in CHANGELOG consistent
Matt Corallo [Tue, 9 Apr 2024 14:39:33 +0000 (14:39 +0000)]
Keep newlines between releases in CHANGELOG consistent

6 months agoAdd release note for 0.0.122 with just three PRs.
Matt Corallo [Mon, 8 Apr 2024 16:16:03 +0000 (16:16 +0000)]
Add release note for 0.0.122 with just three PRs.

6 months agoUse consistent byte/char offsets when parsing invoice HRPs
Matt Corallo [Tue, 30 Apr 2024 17:11:54 +0000 (17:11 +0000)]
Use consistent byte/char offsets when parsing invoice HRPs

When parsing lightning-invoice HRPs we want to read them
char-by-char, tracking at which offset different fields were. Prior
to this commit this was done first by reading char-by-char and then
by indexing using the byte offset which works for ASCII strings but
fails on multi-byte characters.

This commit fixes this issue by simply always walking byte-by-byte
and rejecting multi-byte characters which don't belong in HRPs.

6 months agoFix the `full_stack_target` breakage test and doc feerate requests
Matt Corallo [Tue, 7 May 2024 16:34:36 +0000 (16:34 +0000)]
Fix the `full_stack_target` breakage test and doc feerate requests

When we added the additional deust exposure checks in
702196819e6445048b803574fcacef77d5ce8c9c we added several
additional feerate fetches which broke the `full_stack_target`
change-detection test.

This updates the hard-coded test to support the new feerate fetches
and also includes a comment on `FeeEstimator` to indicate that
users really need to be caching feerates as otherwise they'll slow
us down.

6 months agoMerge pull request #3048 from TheBlueMatt/2024-05-123-backports v0.0.123-rc1
Matt Corallo [Tue, 7 May 2024 18:12:26 +0000 (11:12 -0700)]
Merge pull request #3048 from TheBlueMatt/2024-05-123-backports

[0.0.123] Bump crate versions to rc1

6 months agoBump crate versions to rc1 2024-05-123-backports
Matt Corallo [Tue, 7 May 2024 18:01:21 +0000 (18:01 +0000)]
Bump crate versions to rc1

6 months agoMerge pull request #3042 from TheBlueMatt/2024-05-123-backports
Matt Corallo [Tue, 7 May 2024 17:47:32 +0000 (10:47 -0700)]
Merge pull request #3042 from TheBlueMatt/2024-05-123-backports

0.0.123 Backports

6 months agoMerge pull request #3045 from TheBlueMatt/2024-03-fees-are-dust
Matt Corallo [Tue, 7 May 2024 15:56:45 +0000 (15:56 +0000)]
Merge pull request #3045 from TheBlueMatt/2024-03-fees-are-dust

Include excess counterparty commitment transaction fees in dust exposure

6 months agoAdd a simple test of the new excess-fees-are-dust behavior
Matt Corallo [Sun, 5 May 2024 02:00:42 +0000 (02:00 +0000)]
Add a simple test of the new excess-fees-are-dust behavior

6 months agoUpdate default dust exposure limit and the documentation thereof
Matt Corallo [Sun, 5 May 2024 00:44:07 +0000 (00:44 +0000)]
Update default dust exposure limit and the documentation thereof

Now that we're including excess counterparty commitment transaction
fees in our dust calculation, we need to update the docs
accordingly. We do so here, describing some of the considerations
and risks that come with the new changes.

We also take this opportunity to double the default value, as users
have regularly complained that non-anchor channels fail to send
HTLCs with the default settings with some feerates.

Fixes #2922

6 months agoInclude excess commitment transaction fees in dust exposure
Matt Corallo [Tue, 30 Apr 2024 20:42:36 +0000 (20:42 +0000)]
Include excess commitment transaction fees in dust exposure

Transaction fees on counterparty commitment transactions are
ultimately not our money and thus are really "dust" from our PoV -
they're funds that may be ours during off-chain updates but are not
ours once we go on-chain.

Thus, here, we count any such fees in excess of our own fee
estimates towards dust exposure. We don't bother to make an
inbound/outbound channel distinction here as in most cases users
will use `MaxDustExposure::FeeRateMultiplier` which will scale
with the fee we set on outbound channels anyway.

Note that this also enables the dust exposure checks on anchor
channels during feerate updates. We'd previously elided these as
increases in the channel feerates do not change the HTLC dust
exposure, but now do for the fee dust exposure.

6 months agoMerge pull request #3039 from valentinewallace/2024-04-invoice-amt-msats-overflow
Matt Corallo [Mon, 6 May 2024 19:21:13 +0000 (19:21 +0000)]
Merge pull request #3039 from valentinewallace/2024-04-invoice-amt-msats-overflow

Fix overflow in invoice amount setter.

6 months agoMerge pull request #3035 from TheBlueMatt/2024-04-upstream-123-bindings
Matt Corallo [Mon, 6 May 2024 18:54:11 +0000 (18:54 +0000)]
Merge pull request #3035 from TheBlueMatt/2024-04-upstream-123-bindings

6 months agoMerge pull request #3018 from jkczyz/2024-04-optional-description
Matt Corallo [Mon, 6 May 2024 18:51:30 +0000 (18:51 +0000)]
Merge pull request #3018 from jkczyz/2024-04-optional-description

Optional description for `Offer` and `Refund`

6 months agoMerge pull request #3029 from TheBlueMatt/2024-04-fix-batch-funding-failures
Matt Corallo [Mon, 6 May 2024 18:50:57 +0000 (18:50 +0000)]
Merge pull request #3029 from TheBlueMatt/2024-04-fix-batch-funding-failures

Add error handling for channels which fail to be created in `funding_transaction_generated_intern`

6 months agoMerge pull request #3033 from TheBlueMatt/2024-04-notify-bp-on-blocks
Matt Corallo [Mon, 6 May 2024 18:49:16 +0000 (18:49 +0000)]
Merge pull request #3033 from TheBlueMatt/2024-04-notify-bp-on-blocks

Wake `background-processor` from `ChainMonitor` on new blocks

6 months agoMerge pull request #3032 from valentinewallace/2024-04-pico-btc-overflow
Matt Corallo [Mon, 6 May 2024 18:46:52 +0000 (18:46 +0000)]
Merge pull request #3032 from valentinewallace/2024-04-pico-btc-overflow

Fix overflow in lightning-invoice `amount_pico_btc`

6 months agoMerge pull request #3031 from TheBlueMatt/2024-04-new-rustc-warning
Matt Corallo [Mon, 6 May 2024 18:46:24 +0000 (18:46 +0000)]
Merge pull request #3031 from TheBlueMatt/2024-04-new-rustc-warning

Fix new rustc `#[macro_export]` warning in doctests

6 months agoMerge pull request #3017 from jkczyz/2024-04-optional-pubkey
Matt Corallo [Mon, 6 May 2024 18:45:25 +0000 (18:45 +0000)]
Merge pull request #3017 from jkczyz/2024-04-optional-pubkey

Sending to `Offer` without `signing_pubkey`

6 months agoMerge pull request #3028 from jkczyz/2024-04-offer-id-followups
Matt Corallo [Mon, 6 May 2024 18:44:44 +0000 (18:44 +0000)]
Merge pull request #3028 from jkczyz/2024-04-offer-id-followups

Follow-ups to #2970

6 months agoFix overflow in invoice amount setter.
Valentine Wallace [Tue, 30 Apr 2024 17:01:03 +0000 (13:01 -0400)]
Fix overflow in invoice amount setter.

7 months agoMerge pull request #3035 from TheBlueMatt/2024-04-upstream-123-bindings
Matt Corallo [Sat, 4 May 2024 00:55:25 +0000 (17:55 -0700)]
Merge pull request #3035 from TheBlueMatt/2024-04-upstream-123-bindings

7 months agoMerge pull request #3018 from jkczyz/2024-04-optional-description
Matt Corallo [Fri, 3 May 2024 19:55:17 +0000 (12:55 -0700)]
Merge pull request #3018 from jkczyz/2024-04-optional-description

Optional description for `Offer` and `Refund`

7 months agoMerge pull request #3029 from TheBlueMatt/2024-04-fix-batch-funding-failures
Matt Corallo [Fri, 3 May 2024 19:17:58 +0000 (12:17 -0700)]
Merge pull request #3029 from TheBlueMatt/2024-04-fix-batch-funding-failures

Add error handling for channels which fail to be created in `funding_transaction_generated_intern`

7 months agoOnly require description when offer has an amount
Jeffrey Czyz [Thu, 2 May 2024 16:43:04 +0000 (11:43 -0500)]
Only require description when offer has an amount

The spec was changed to allow excluding an offer description if the
offer doesn't have an amount. However, it is still required when the
amount is set.

7 months agoMerge pull request #3034 from TheBlueMatt/2024-04-2982-123-fix
valentinewallace [Thu, 2 May 2024 18:01:06 +0000 (14:01 -0400)]
Merge pull request #3034 from TheBlueMatt/2024-04-2982-123-fix

[0.0.123] Don't attempt to resume channels if we already exchanged funding

7 months agoSwap `UserConfig::default()` for `test_default_channel_config`
Matt Corallo [Wed, 1 May 2024 21:56:04 +0000 (21:56 +0000)]
Swap `UserConfig::default()` for `test_default_channel_config`

As LDK changes, `UserConfig::default()` may imply marginally
different behavior, whereas `test_default_channel_config` is
intended to tweak defaults to provide a stable behavior for test
contexts.

This commit changes a few uses of `UserConfig::default()` to
`test_default_channel_config` in cases that will fail over the
coming commits due to dust changes.

7 months agoCorrect indentation in `get_pending_htlc_stats`
Matt Corallo [Wed, 27 Mar 2024 02:21:40 +0000 (02:21 +0000)]
Correct indentation in `get_pending_htlc_stats`

The previous commit left indentation in `get_pending_htlc_stats`
deliberately wrong, to ease reviewability. This commit fixes the
indentation.

7 months agoUnify `get_{inbound,outbound}_pending_htlc_stats`
Matt Corallo [Wed, 27 Mar 2024 02:14:50 +0000 (02:14 +0000)]
Unify `get_{inbound,outbound}_pending_htlc_stats`

In most cases we already call both in a pair, and in fact always
consolidate some of the returned values across both accessors, so
there's not much reason to have them be separate methods.

Here we merge them.

7 months agoDon't attempt to resume channels if we already exchanged funding
Matt Corallo [Wed, 1 May 2024 14:32:47 +0000 (14:32 +0000)]
Don't attempt to resume channels if we already exchanged funding

ddf75afd16 introduced the ability to re-exchange our `ChannelOpen`
after a peer disconnects if we didn't complete funding on our end.
It did not implement nor consider what would happen if we
re-connected after we created our own funding transactions, and
currently it panics (and even if it did not it would replay the
`FundingTransactionGenerated` event to users).

This addresses this in a temporary fashion to fix the immediate
panic for 0.0.123, allowing us to fix the issue more fully later.

7 months agoMerge pull request #3033 from TheBlueMatt/2024-04-notify-bp-on-blocks
Matt Corallo [Wed, 1 May 2024 21:21:28 +0000 (14:21 -0700)]
Merge pull request #3033 from TheBlueMatt/2024-04-notify-bp-on-blocks

Wake `background-processor` from `ChainMonitor` on new blocks

7 months agoWake `background-processor` from `ChainMonitor` on new blocks 2024-04-notify-bp-on-blocks
Matt Corallo [Tue, 30 Apr 2024 21:30:26 +0000 (21:30 +0000)]
Wake `background-processor` from `ChainMonitor` on new blocks

When we receive a new block we may generate
`Event::SpendableOutputs` in `ChannelMonitor`s which then need to
be processed by the background processor. While it will do so
eventually when its normal loop goes around, this may cause user
tests to be delayed in finding events, so we should notify the BP
immediately to wake it on new blocks.

We implement that here, unconditionally notifying the
`background-processor` whenever we receive a new block or confirmed
transactions.

7 months agoUpdate documentation on `[u8; 32]` wrappers for clarity 2024-04-upstream-123-bindings
Matt Corallo [Wed, 1 May 2024 16:04:13 +0000 (16:04 +0000)]
Update documentation on `[u8; 32]` wrappers for clarity

7 months agoMove `[u8; 32]` wrapper types to a common module
Matt Corallo [Wed, 1 May 2024 15:49:59 +0000 (15:49 +0000)]
Move `[u8; 32]` wrapper types to a common module

The `PaymentHash`, `PaymentSecret`, `PaymentPreimage`, and
`ChannelId` types are all small wrappers around `[u8; 32]` and are
used throughout the codebase but were defined in the top-level
`ln/mod.rs` file and the relatively sparsely-populated
`ln/channel_id.rs` file.

Here we move them to a common `types` module and go ahead and
update all our in-crate `use` statements to refer to the new
module for bindings. We do, however, leave a `pub use` alias for
the old paths to avoid upgrade hassle for users.

7 months agoAdd bindings no-export tags to doc test types
Matt Corallo [Wed, 1 May 2024 15:19:20 +0000 (15:19 +0000)]
Add bindings no-export tags to doc test types

We don't actually intend these to be public as they're just for
docs but the bindings don't currently parse `#[doc(hidden)]` as
"no-export" so we add manual no-export tags as well.

7 months agoAdd derives where they make sense and the bindings require them
Matt Corallo [Wed, 1 May 2024 15:18:29 +0000 (15:18 +0000)]
Add derives where they make sense and the bindings require them

7 months agoRename and expose message-specific `NextHop`
Matt Corallo [Tue, 23 Apr 2024 15:56:13 +0000 (15:56 +0000)]
Rename and expose message-specific `NextHop`

`onion::message::messenger::PeeledOnion` is a public enum which
included the private enum `NextHop`, which is not acceptable. Thus,
we here expose `NextHop` but rename it `NextMessageHop` to make
clear that it is specific to messages.

7 months agoMerge pull request #3032 from valentinewallace/2024-04-pico-btc-overflow
Matt Corallo [Wed, 1 May 2024 13:08:45 +0000 (06:08 -0700)]
Merge pull request #3032 from valentinewallace/2024-04-pico-btc-overflow

Fix overflow in lightning-invoice `amount_pico_btc`

7 months agoMerge pull request #3031 from TheBlueMatt/2024-04-new-rustc-warning
valentinewallace [Tue, 30 Apr 2024 17:16:11 +0000 (13:16 -0400)]
Merge pull request #3031 from TheBlueMatt/2024-04-new-rustc-warning

Fix new rustc `#[macro_export]` warning in doctests

7 months agoFix new rustc `#[macro_export]` warning in doctests 2024-04-new-rustc-warning
Matt Corallo [Tue, 30 Apr 2024 13:45:45 +0000 (13:45 +0000)]
Fix new rustc `#[macro_export]` warning in doctests

rustc now warns any time a `#[macro_export]` is used inside a
function as it generates surprising results. Because doctests are
run inside implicit test functions this means any use of
`#[macro_export]` inside a doctest will now warn. We do this in
`lightning-custom-message` to demonstrated how to use the crate,
which now fails to compile.

Here we fix this by adding an `fn main() {}` to the doctest, which
causes doctests to be compiled as freestanding code rather than
inside a test function.

Note that while discussing this upstream it came up that rustc is
also planning on changing the way doctests are compiled to compile
an entire crate's doctests in one go rather than in separate
crates, causing doctests to have a shared namespace which may
generate future surprising outcomes.

7 months agoFix overflow in lightning-invoice amount_pico_btc.
Valentine Wallace [Tue, 30 Apr 2024 14:05:35 +0000 (10:05 -0400)]
Fix overflow in lightning-invoice amount_pico_btc.

7 months agoSend peers error messages for failures due to invalid batch funding 2024-04-fix-batch-funding-failures
Matt Corallo [Mon, 29 Apr 2024 21:05:52 +0000 (21:05 +0000)]
Send peers error messages for failures due to invalid batch funding

If we fail to fund a batch open we'll force-close all channels in
the batch, however would previously fail to send error messages to
our peers for any channels we were due to test after the one that
failed.

This commit fixes that issue, sending the required error messages
to allow our peers to clean up their state.

7 months agoClose channels when `find_funding_output` fails to find an output
Matt Corallo [Mon, 29 Apr 2024 18:12:59 +0000 (18:12 +0000)]
Close channels when `find_funding_output` fails to find an output

In `funding_transaction_generated_intern`, if `find_funding_output`
fails (e.g. due to the require output not being present in the
provided funding transaction) we'd previously not generated a
`ChannelClosed` event which leaves users possibly in a confused
state.

Here we fix this, also fixing the relevant tests to check for the
new event.

Fixes #2843.

7 months agoMerge pull request #3017 from jkczyz/2024-04-optional-pubkey
Matt Corallo [Mon, 29 Apr 2024 19:05:30 +0000 (12:05 -0700)]
Merge pull request #3017 from jkczyz/2024-04-optional-pubkey

Sending to `Offer` without `signing_pubkey`

7 months agoMerge pull request #3028 from jkczyz/2024-04-offer-id-followups
Matt Corallo [Mon, 29 Apr 2024 18:26:44 +0000 (11:26 -0700)]
Merge pull request #3028 from jkczyz/2024-04-offer-id-followups

Follow-ups to #2970

7 months agoRemove InvoiceRequestFields::features
Jeffrey Czyz [Fri, 26 Apr 2024 17:13:12 +0000 (12:13 -0500)]
Remove InvoiceRequestFields::features

InvoiceRequestFeatures may contain a large, odd bit. Including this in
InvoiceRequestFields, which is in each BlindedPath of a Bolt12Invoice,
could cause the invoice's onion message to exceed the maximum size. The
features are already checked before sending an invoice.

7 months agoRemove InvoiceRequestFields::amount_msats
Jeffrey Czyz [Fri, 26 Apr 2024 17:03:07 +0000 (12:03 -0500)]
Remove InvoiceRequestFields::amount_msats

Event::PaymentClaimable and Event::PaymentClaimed already contain the
paid amount, so there's no need to included the requested amount in
InvoiceRequestFields.

7 months agoRemove PaymentContext from OnionPayload
Jeffrey Czyz [Fri, 26 Apr 2024 16:37:24 +0000 (11:37 -0500)]
Remove PaymentContext from OnionPayload

PaymentContext is already stored in ClaimablePayment via PaymentPurpose,
so there is no need to repeat it in each ClaimableHTLC via OnionPayload.
This avoids cloning the PaymentContext each time an HTLC is received,
other than for the first HTLC for a payment.

7 months agoDon't require Refund::description in API
Jeffrey Czyz [Wed, 24 Apr 2024 18:59:01 +0000 (13:59 -0500)]
Don't require Refund::description in API

Refunds currently require a description, though this may change to be
optional. Remove the description requirement from the API, setting and
empty string by default.

7 months agoDon't require Offer::description in API
Jeffrey Czyz [Wed, 24 Apr 2024 17:40:22 +0000 (12:40 -0500)]
Don't require Offer::description in API

Offers currently require a description, though this may change to be
optional. Remove the description requirement from the API, setting and
empty string by default.

7 months agoBolt12Invoice for Offer without signing_pubkey
Jeffrey Czyz [Tue, 23 Apr 2024 23:35:05 +0000 (18:35 -0500)]
Bolt12Invoice for Offer without signing_pubkey

When parsing a Bolt12Invoice use both the Offer's signing_pubkey and
paths to determine if it is for an Offer or a Refund. Previously, an
Offer was required to have a signing_pubkey. But now that it is
optional, the Offers paths can be used to make the determination.
Additionally, check that the invoice matches one of the blinded node ids
from the paths' last hops.

7 months agoAdd InvoiceRequestTlvStream::paths
Jeffrey Czyz [Tue, 23 Apr 2024 23:30:59 +0000 (18:30 -0500)]
Add InvoiceRequestTlvStream::paths

Instead of reusing OfferTlvStream::paths, add a dedicated paths TLV to
InvoiceRequestTlvStream such that it can be used in Refund. This allows
for an Offer without a signing_pubkey and still be able to differentiate
whether an invoice is for an offer or a refund.

7 months agoAllow parsing Offer without signing_pubkey
Jeffrey Czyz [Wed, 24 Apr 2024 14:03:56 +0000 (09:03 -0500)]
Allow parsing Offer without signing_pubkey

If an offer has at least one path, it may omit the signing pubkey and
use the blinded node id of the last hop of a path to sign an invoice.
Allow parsing such offers but not yet creating them.

7 months agoOptional OfferContents::signing_pubkey
Jeffrey Czyz [Tue, 23 Apr 2024 22:38:13 +0000 (17:38 -0500)]
Optional OfferContents::signing_pubkey

If an Offer contains a path, the blinded_node_id of the path's final hop
can be used as the signing pubkey. Make Offer::signing_pubkey and
OfferContents::signing_pubkey return an Option to support this. Upcoming
commits will implement this behavior.

7 months agoMerge pull request #3003 from TheBlueMatt/2024-04-2964-followups
Gursharan Singh [Thu, 25 Apr 2024 20:19:09 +0000 (13:19 -0700)]
Merge pull request #3003 from TheBlueMatt/2024-04-2964-followups

#2964 Followups

7 months agoMerge pull request #3021 from TheBlueMatt/2024-04-drop-blocked-completed-updates
Matt Corallo [Thu, 25 Apr 2024 18:33:18 +0000 (11:33 -0700)]
Merge pull request #3021 from TheBlueMatt/2024-04-drop-blocked-completed-updates

Drop completed blocked `ChannelMonitorUpdate`s on startup

7 months agoMerge pull request #3004 from TheBlueMatt/2024-04-2761-followups
Matt Corallo [Thu, 25 Apr 2024 17:52:12 +0000 (10:52 -0700)]
Merge pull request #3004 from TheBlueMatt/2024-04-2761-followups

#2761 Followups

7 months agoDrop completed blocked `ChannelMonitorUpdate`s on startup 2024-04-drop-blocked-completed-updates
Matt Corallo [Thu, 25 Apr 2024 14:30:05 +0000 (14:30 +0000)]
Drop completed blocked `ChannelMonitorUpdate`s on startup

If a user receives a payment preimage for an outbound payment, the
`PaymentSent` event will block any eventual RAA
`ChannelMonitorUpdate` from the same channel, assuming it comes in
before the event can be processed. If this blocking kicks in, but
the flow eventually completes with the RAA `ChannelMonitorUpdate`
being persisted, but the `ChannelManager` is only persisted prior
to the event being handled, on startup we'll have a fully
up-to-date `ChannelMonitor` but a pending, blocked
`ChannelMonitorUpdate`. When the `PaymentSent` event is replayed
we'll end up trying to apply a redundant `ChannelMonitorUpdate`
which will panic.

See the test added in this commit for an implementation of this
situation.

In this commit we fix this issue by simply dropping blocked
`ChannelMonitorUpdate`s the same as we do pending ones.

7 months agoMerge pull request #3015 from tnull/2024-04-sweeper-error-on-persistence-failure
Elias Rohrer [Wed, 24 Apr 2024 14:35:51 +0000 (16:35 +0200)]
Merge pull request #3015 from tnull/2024-04-sweeper-error-on-persistence-failure

Make `OutputSweeper::track_spendable_outputs` fallible

7 months agoMake `OutputSweeper::track_spendable_outputs` fallible
Elias Rohrer [Wed, 24 Apr 2024 07:30:16 +0000 (09:30 +0200)]
Make `OutputSweeper::track_spendable_outputs` fallible

.. as otherwise we might only log an error and continue if we fail to
persist the sweeper state.

7 months agoMerge pull request #2981 from dunxen/2024-03-PR2419fixups
Matt Corallo [Tue, 23 Apr 2024 13:48:40 +0000 (06:48 -0700)]
Merge pull request #2981 from dunxen/2024-03-PR2419fixups

Fixups for #2419

7 months agoAdd intermediate `ConstructedTransaction`
Duncan Dean [Fri, 12 Apr 2024 13:37:17 +0000 (15:37 +0200)]
Add intermediate `ConstructedTransaction`

7 months agoAdd lower bound segwit input weight estimates
Duncan Dean [Tue, 9 Apr 2024 17:07:45 +0000 (19:07 +0200)]
Add lower bound segwit input weight estimates

7 months agoAdd descriptions to test cases
Duncan Dean [Mon, 8 Apr 2024 15:44:10 +0000 (17:44 +0200)]
Add descriptions to test cases