]> git.bitcoin.ninja Git - rust-lightning/log
rust-lightning
4 years agoUpdate the root Cargo.toml to point to the new subcrate
Matt Corallo [Fri, 24 Jul 2020 18:31:36 +0000 (14:31 -0400)]
Update the root Cargo.toml to point to the new subcrate

4 years agoAdd automatically generated C/C++ wrapper bindings
Matt Corallo [Thu, 10 Sep 2020 22:19:15 +0000 (18:19 -0400)]
Add automatically generated C/C++ wrapper bindings

4 years agoAdd all the manually-generated bits for the c-bindings crate
Matt Corallo [Tue, 12 May 2020 18:54:12 +0000 (14:54 -0400)]
Add all the manually-generated bits for the c-bindings crate

Including:
 * A script to automatically generate all the rest,
 * Cargo.toml and cbindgen.toml,
 * manually-written wrapper types for a few types

4 years agoAdd annotations for things which we cannot (yet) expose
Matt Corallo [Tue, 12 May 2020 17:48:07 +0000 (13:48 -0400)]
Add annotations for things which we cannot (yet) expose

4 years agoAdd tool to read a Rust crate and generate C-compatible wrappers
Matt Corallo [Tue, 12 May 2020 18:49:29 +0000 (14:49 -0400)]
Add tool to read a Rust crate and generate C-compatible wrappers

In general, it maps:
 * Traits to a struct with a void* and a list of function pointers,
   emulating what the compiler will do for a dyn trait anyway,
 * Structs as a struct with a single opaque pointer to the
   underlying type and a flag to indicate ownership. While this is
   a bit less effecient than just a direct pointer, it neatly lets
   us expose in the public interface the concept of ownership by
   setting a flag in the generated struct.
 * Unit enums as enums with each type copied over and conversion
   functions,
 * Non-unit enums have each field converted back and forth with a
   type flag and a union across all the C-mapped fields.

4 years agoRemove a few final ::-containing paths in generic args
Matt Corallo [Fri, 11 Sep 2020 01:37:08 +0000 (21:37 -0400)]
Remove a few final ::-containing paths in generic args

This cleans upa few last cases of functions/objects which our C
bindings generator doesn't know how to read.

4 years agoMerge pull request #686 from TheBlueMatt/2020-09-bump-deps
Matt Corallo [Thu, 10 Sep 2020 22:11:30 +0000 (15:11 -0700)]
Merge pull request #686 from TheBlueMatt/2020-09-bump-deps

Bump rust-bitcoin and friends

4 years agoSilence new rustc warning about useless muts 2020-09-bump-deps
Matt Corallo [Thu, 10 Sep 2020 20:32:45 +0000 (16:32 -0400)]
Silence new rustc warning about useless muts

4 years agoUse rust-bitcoin's new SigHashCache instead of SignatureHashComp's
Matt Corallo [Tue, 25 Aug 2020 22:47:24 +0000 (18:47 -0400)]
Use rust-bitcoin's new SigHashCache instead of SignatureHashComp's

Thew new API is a bit harder to misuse by taking a reference to the
transaction to require the inputs match the input being signed.

4 years agoUpdate to latest upstream rust-bitcoin
Matt Corallo [Tue, 25 Aug 2020 21:12:00 +0000 (17:12 -0400)]
Update to latest upstream rust-bitcoin

4 years agoMerge pull request #676 from TheBlueMatt/2020-08-c-bindings-cleanups-3
Matt Corallo [Wed, 26 Aug 2020 15:14:34 +0000 (08:14 -0700)]
Merge pull request #676 from TheBlueMatt/2020-08-c-bindings-cleanups-3

Pre-C-Bindings Cleanups #3

4 years agoMerge pull request #674 from TheBlueMatt/2020-08-keyif-rand-names
Matt Corallo [Wed, 26 Aug 2020 15:07:58 +0000 (08:07 -0700)]
Merge pull request #674 from TheBlueMatt/2020-08-keyif-rand-names

Simplify + clarify random-bytes-fetching from KeysInterface

4 years agoAdd a C-bindings-compatible read lock type for NetworkGraph 2020-08-c-bindings-cleanups-3
Matt Corallo [Mon, 24 Aug 2020 18:14:05 +0000 (14:14 -0400)]
Add a C-bindings-compatible read lock type for NetworkGraph

In order to calculate a route, it is likely that users need to take
a read()-lock on NetGraphMsgHandler::network_graph. This is not
possible naively from C bindings, as Rust's native RwLock is not
exposed.

Thus, we provide a simple wrapper around the RwLockReadGuard and
expose simple accessor methods.

4 years agoRename TxCreationKeys::new to not conflict w/ auto-gen'd C bindings
Matt Corallo [Thu, 13 Aug 2020 18:45:34 +0000 (14:45 -0400)]
Rename TxCreationKeys::new to not conflict w/ auto-gen'd C bindings

The C bindings automatically create a _new() function for structs
which contain only pub fields which we know how to map. This
conflicts with the actual TxCreationKeys::new() function, so we
simply rename it to capture its nature as a derivation function.

4 years agoGive ChannelManagerReadArgs HashMap-of-monitors ownership
Matt Corallo [Fri, 7 Aug 2020 20:27:26 +0000 (16:27 -0400)]
Give ChannelManagerReadArgs HashMap-of-monitors ownership

Its somewhat awkward that ChannelManagerReadArgs requires a mutable
reference to a HashMap of ChannelMonitors, forcing the callsite to
define a scope for the HashMap which they almost certainly won't use
after deserializing the ChannelManager. Worse, to map the current
version to C bindings, we'd need to also create a HashMap binding,
which is overkill for just this one use.

Instead, we just give the ReadArgs struct ownership of the HashMap
and add a constructor which fills the HashMap for you.

4 years agoUse ln OutPoints not bitcoin ones in SpendableOutputDescriptors
Matt Corallo [Thu, 30 Jul 2020 17:19:11 +0000 (13:19 -0400)]
Use ln OutPoints not bitcoin ones in SpendableOutputDescriptors

Lightning OutPoints only have 16 bits to express the output index
instead of Bitcoin's 32 bits, implying that some outputs are
possibly not expressible as lightning OutPoints. However, such
OutPoints can never be hit within the lightning protocol, and must
be on-chain spam sent by a third party wishing to donate us money.
Still, in order to do so, the third party would need to fill nearly
an entire block with garbage, so this case should be relatively
safe.

A new comment in channelmonitor explains the reasoning a bit
further.

4 years agoUse slices to references not slices of concrete objects in pub API
Matt Corallo [Thu, 23 Jul 2020 20:10:29 +0000 (16:10 -0400)]
Use slices to references not slices of concrete objects in pub API

Because the C bindings maps objects into new structs which contain
only a pointer to the underlying (immovable) Rust type, it cannot
create a list of Rust types which are contiguous in memory. Thus,
in order to allow C clients to call certain Rust functions, we have
to use &[&Type] not &[Type]. This commit fixes this issue for the
get_route function.

4 years agoRefer to return types by the trait that they're defined via
Matt Corallo [Sun, 31 May 2020 03:22:16 +0000 (23:22 -0400)]
Refer to return types by the trait that they're defined via

Instead of using the explicit type which is being returned, refer
to them as Self::AssociatedType, to make clear to the bindings what
type of thing is being returned.

4 years agoRefer to generic types by importing them instead of a super-mod.
Matt Corallo [Sun, 31 May 2020 03:20:17 +0000 (23:20 -0400)]
Refer to generic types by importing them instead of a super-mod.

This avoids one case the bindings generation hasn't bothered to
handle by simply importing types that are referred to.

4 years agoAlways refer to Deref types with where clauses instead of direct
Matt Corallo [Sun, 31 May 2020 03:16:29 +0000 (23:16 -0400)]
Always refer to Deref types with where clauses instead of direct

This makes it a little easier to write C bindings generation as
we only have to handle one case instead of both.

4 years agoReorder struct definitions so that they are in dependency order.
Matt Corallo [Thu, 11 Jun 2020 19:34:28 +0000 (15:34 -0400)]
Reorder struct definitions so that they are in dependency order.

There are a few cases where the upcoming C bindings don't know how
to handle something which depends on something defined later in the
file. Instead of adding another pass to the C bindings generator,
it is much simpler to just reorder structs.

4 years agoMerge pull request #667 from valentinewallace/remove-channels-chanmon
Matt Corallo [Tue, 25 Aug 2020 21:07:27 +0000 (14:07 -0700)]
Merge pull request #667 from valentinewallace/remove-channels-chanmon

Remove Channel's ChannelMonitor copy

4 years agoRemove Channel's ChannelMonitor copy
Valentine Wallace [Thu, 13 Aug 2020 18:35:39 +0000 (14:35 -0400)]
Remove Channel's ChannelMonitor copy

4 years agoAdd commitment transaction broadcast as a ChannelMonitor event
Valentine Wallace [Mon, 24 Aug 2020 21:27:49 +0000 (17:27 -0400)]
Add commitment transaction broadcast as a ChannelMonitor event

To do this, we replace get_and_clear_pending_htlcs_updated with
get_and_clear_pending_monitor_events, and which still transmits HTLCUpdates
as before, but now also transmits a new MonitorEvent::CommitmentTxBroadcasted
event when a channel's commitment transaction is broadcasted.

4 years agoMerge pull request #675 from dspicher/master
Matt Corallo [Tue, 25 Aug 2020 01:18:38 +0000 (18:18 -0700)]
Merge pull request #675 from dspicher/master

Refactor route tests

4 years agoRefactor route tests
Dominik Spicher [Mon, 24 Aug 2020 21:06:07 +0000 (23:06 +0200)]
Refactor route tests

* Splits up the monolithic test into smaller unit tests

* Factors out helpers for graph setup

* Changes `id_to_feature_flags` to be a function, there was no
reason why it had to be a macro

* Activates a previously commented-out test that checks for
the failure case in `disable_node_test`

4 years agoSimplify + clarify random-bytes-fetching from KeysInterface 2020-08-keyif-rand-names
Matt Corallo [Sun, 23 Aug 2020 21:06:33 +0000 (17:06 -0400)]
Simplify + clarify random-bytes-fetching from KeysInterface

Due to a desire to be able to override temporary channel IDs and
onion keys, KeysInterface had two separate fetch-random-32-bytes
interfaces - an onion-key specific version which fetched 2 random
32 byte strings and a temporary-channel-id specific version.

It turns out, we never actually need to override both at once (as
creating a new channel and sending an outbound payment are always
separate top-level calls), so there's no reason to add two
functions to the interface when both really do the same thing.

4 years agoMerge pull request #665 from lightning-signer/pub-messages
Matt Corallo [Sun, 23 Aug 2020 21:28:29 +0000 (14:28 -0700)]
Merge pull request #665 from lightning-signer/pub-messages

Make message fields public

4 years agoMake message fields public
Devrandom [Mon, 10 Aug 2020 09:26:46 +0000 (11:26 +0200)]
Make message fields public

4 years agoMerge pull request #671 from valentinewallace/bump-msrv
Matt Corallo [Wed, 19 Aug 2020 20:04:41 +0000 (13:04 -0700)]
Merge pull request #671 from valentinewallace/bump-msrv

Bump MSRV to 1.30.0

4 years agoBump MSRV to 1.30.0
Valentine Wallace [Wed, 19 Aug 2020 19:35:37 +0000 (15:35 -0400)]
Bump MSRV to 1.30.0

We wanted to bump to 1.29 to continue to support mrustc bootstrapping, but on 1.29
there's a bug preventing us from compiling the lightning package only, thus parts
of lightning-net-tokio cause a compilation error.

The advantage of bumping the MSRV is an improved borrow checker which should
enable improved code quality, and not having jump through weird hoops sometimes
to get 1.22 working.

4 years agoMerge pull request #669 from joemphilips/fix_capacity-is-always-zero-bug_in_list_channels
Matt Corallo [Thu, 13 Aug 2020 21:06:39 +0000 (14:06 -0700)]
Merge pull request #669 from joemphilips/fix_capacity-is-always-zero-bug_in_list_channels

Fix bug in Channel

4 years agoMerge pull request #668 from lightning-signer/nit
Matt Corallo [Thu, 13 Aug 2020 17:08:21 +0000 (10:08 -0700)]
Merge pull request #668 from lightning-signer/nit

fix typo

4 years agoFix bug in Channel
joe.miyamoto [Thu, 13 Aug 2020 08:10:24 +0000 (17:10 +0900)]
Fix bug in Channel

Before this commit, `fn get_inbound_outbound_available_balance_msat` always returns 0.
It is because using `cmp::min` instead of `cmp::max` .

4 years agofix typo
Devrandom [Thu, 13 Aug 2020 07:58:55 +0000 (09:58 +0200)]
fix typo

4 years agoMerge pull request #662 from lightning-signer/compat5
Matt Corallo [Tue, 11 Aug 2020 16:10:05 +0000 (09:10 -0700)]
Merge pull request #662 from lightning-signer/compat5

Export various fields and structures

4 years agocheck the input shape in LocalCommitmentTransaction.new_missing_local_sig
Devrandom [Tue, 11 Aug 2020 07:23:15 +0000 (09:23 +0200)]
check the input shape in LocalCommitmentTransaction.new_missing_local_sig

4 years agomake LocalCommitmentTransaction constructor public
Devrandom [Mon, 20 Jul 2020 12:26:06 +0000 (05:26 -0700)]
make LocalCommitmentTransaction constructor public

Allows calling of InMemoryChannelKeys methods

4 years agomake SimpleManyChannelMonitor.monitors public
Devrandom [Sat, 4 Jul 2020 08:18:54 +0000 (01:18 -0700)]
make SimpleManyChannelMonitor.monitors public

4 years agoexport "unsafe_revoked_tx_signing" feature
Devrandom [Fri, 3 Jul 2020 20:41:21 +0000 (13:41 -0700)]
export "unsafe_revoked_tx_signing" feature

Allows unsafe signing in dev code, such as functional testing of
justice transactions outside our crate.

4 years agoMerge pull request #666 from TheBlueMatt/2020-08-license-change
Matt Corallo [Tue, 11 Aug 2020 01:30:38 +0000 (18:30 -0700)]
Merge pull request #666 from TheBlueMatt/2020-08-license-change

Relicense as dual Apache-2.0 + MIT

4 years agoRelicense as dual Apache-2.0 + MIT 2020-08-license-change
Matt Corallo [Mon, 10 Aug 2020 19:00:09 +0000 (15:00 -0400)]
Relicense as dual Apache-2.0 + MIT

This changes the LICENSE file and adds license headers to most files
to relicense under dual Apache-2.0 and MIT. This is helpful in that
we retain the patent grant issued under Apache-2.0-licensed work,
avoiding some sticky patent issues, while still allowing users who
are more comfortable with the simpler MIT license to use that.

See https://github.com/rust-bitcoin/rust-lightning/issues/659 for
relicensing statements from code authors.

4 years agoMerge pull request #664 from lightning-signer/tx-creation-keys
Matt Corallo [Mon, 10 Aug 2020 20:25:03 +0000 (13:25 -0700)]
Merge pull request #664 from lightning-signer/tx-creation-keys

Wrap transaction creation keys

4 years agonit
Devrandom [Mon, 10 Aug 2020 20:09:03 +0000 (22:09 +0200)]
nit

4 years agoExpose TxCreationKeys in LocalCommitmentTransaction via a method
Devrandom [Sun, 9 Aug 2020 13:56:10 +0000 (15:56 +0200)]
Expose TxCreationKeys in LocalCommitmentTransaction via a method

This makes it obvious to signer implementers that the pre-derived keys are a local cache and should not be trusted in a validating signer.

4 years agoMake TxCreationKeys public and wrap it in PreCalculatedTxCreationKeys
Devrandom [Sun, 9 Aug 2020 13:45:23 +0000 (15:45 +0200)]
Make TxCreationKeys public and wrap it in PreCalculatedTxCreationKeys

Allows calling of InMemoryChannelKeys methods.

The wrapping makes it obvious to signer implementers that the pre-derived keys are a local cache and should not be trusted in a validating signer.

4 years agoMerge pull request #640 from valentinewallace/test-holding-cell-edge-case
Matt Corallo [Sat, 8 Aug 2020 21:20:09 +0000 (14:20 -0700)]
Merge pull request #640 from valentinewallace/test-holding-cell-edge-case

Fail back HTLCs that fail to be freed from the holding cell

4 years agoHolding cell: if we fail to free an HTLC, fail it backwards
Valentine Wallace [Wed, 6 May 2020 22:15:43 +0000 (18:15 -0400)]
Holding cell: if we fail to free an HTLC, fail it backwards

Plus add a test.

4 years agoMerge pull request #663 from TheBlueMatt/2020-08-codecov
Matt Corallo [Mon, 3 Aug 2020 21:25:56 +0000 (14:25 -0700)]
Merge pull request #663 from TheBlueMatt/2020-08-codecov

Set codecov upload token

4 years agoSet codecov upload token 2020-08-codecov
Matt Corallo [Mon, 3 Aug 2020 20:48:58 +0000 (16:48 -0400)]
Set codecov upload token

Docs seem to indicate this should only be required for private
repos, but we have builds failing claiming this needs to be
specified, so just set it manually.

4 years agoMerge pull request #658 from lightning-signer/self-delay
Matt Corallo [Wed, 29 Jul 2020 19:08:13 +0000 (12:08 -0700)]
Merge pull request #658 from lightning-signer/self-delay

ChannelKeys - provide to_self_delay alongside the remote channel pubkeys

4 years agoChannelKeys - provide to_self_delay alongside the remote channel pubkeys
Devrandom [Thu, 23 Jul 2020 13:54:49 +0000 (06:54 -0700)]
ChannelKeys - provide to_self_delay alongside the remote channel pubkeys

In the phase 2 signer, we will construct the commitment transaction inside the signer.
In preparation, provide needed channel related data.

4 years agoMerge pull request #651 from naumenkogs/2020-06-routing-data-improvements
Matt Corallo [Mon, 27 Jul 2020 17:18:13 +0000 (10:18 -0700)]
Merge pull request #651 from naumenkogs/2020-06-routing-data-improvements

Routing improvements

4 years agoCheck htlc_maximum_msat on channel update
Gleb Naumenko [Mon, 29 Jun 2020 18:28:15 +0000 (21:28 +0300)]
Check htlc_maximum_msat on channel update

4 years agoStore channel capacity if available
Gleb Naumenko [Sun, 28 Jun 2020 12:18:33 +0000 (15:18 +0300)]
Store channel capacity if available

4 years agoAdd htlc_maximum_msat field
Gleb Naumenko [Sun, 28 Jun 2020 11:43:10 +0000 (14:43 +0300)]
Add htlc_maximum_msat field

4 years agoUpdate comment for test_msg_hole
Gleb Naumenko [Fri, 24 Jul 2020 07:19:45 +0000 (10:19 +0300)]
Update comment for test_msg_hole

4 years agoMerge pull request #655 from lightning-signer/per-commitment
Matt Corallo [Wed, 22 Jul 2020 20:11:45 +0000 (13:11 -0700)]
Merge pull request #655 from lightning-signer/per-commitment

ChannelKeys - separate commitment revocation from getting the per-commitment point

4 years agoChannelKeys - separate commitment revocation from getting the per-commitment point
Devrandom [Sat, 11 Jul 2020 10:19:43 +0000 (03:19 -0700)]
ChannelKeys - separate commitment revocation from getting the per-commitment point

The commitment secret is sensitive - it can be used by an attacker to
steal funds if the node also signs the same transaction. Therefore,
only release the secret from ChannelKeys when we are revoking a
transaction.

4 years agoImprove routing announcement test
Gleb Naumenko [Mon, 29 Jun 2020 17:48:01 +0000 (20:48 +0300)]
Improve routing announcement test

4 years agoFix formatting
Gleb Naumenko [Thu, 25 Jun 2020 14:01:42 +0000 (17:01 +0300)]
Fix formatting

4 years agoUse constant for max msats
Gleb Naumenko [Tue, 16 Jun 2020 15:58:06 +0000 (18:58 +0300)]
Use constant for max msats

4 years agoUpdate get_route comment to reflect new parameters
Gleb Naumenko [Tue, 16 Jun 2020 10:17:43 +0000 (13:17 +0300)]
Update get_route comment to reflect new parameters

4 years agoMerge pull request #644 from joemphilips/improve_error_message
Matt Corallo [Wed, 22 Jul 2020 03:04:44 +0000 (20:04 -0700)]
Merge pull request #644 from joemphilips/improve_error_message

Improve error message.

4 years agoMerge pull request #623 from LNP-BP/feat-msgsizelimit
Matt Corallo [Wed, 22 Jul 2020 03:03:37 +0000 (20:03 -0700)]
Merge pull request #623 from LNP-BP/feat-msgsizelimit

Making message size limit an exportable constant

4 years agoslightly refactor internal_announcement_signatures
joe.miyamoto [Mon, 13 Jul 2020 04:24:40 +0000 (13:24 +0900)]
slightly refactor internal_announcement_signatures

For making debugging easy.
If the user gives a different node_secret for transport
layer (`PeerManager`) and for routing msg, internal_announcement_signatures
is the first place it causes an error.
By giving a detailed error message, user will be able to
fix the bug quickly.

4 years agoImprove error message.
joe.miyamoto [Mon, 13 Jul 2020 04:16:32 +0000 (13:16 +0900)]
Improve error message.

... for ChannelError and APIMisuseError
Before this commit, When rl returns error, we don't know
The actual parameter which caused the error.
By returning parameterised `String` instead of predefined `&'static str`,
We can give a caller improved error message.

TestLogger now has two additional methods
1. `assert_log_contains` which checks the logged messsage
  has how many entry which includes the specified string as a substring.
2. `aasert_log_regex` mostly the same with `assert_log_contains`
  but it is more flexible that caller specifies regex which has
  to be satisfied instead of just a substring.
For regex, tests now includes `regex` as dev-dependency.

4 years agoMoving LN_MAX_MSG_LEN const to the actual use place
Dr. Maxim Orlovsky [Tue, 21 Jul 2020 12:14:01 +0000 (14:14 +0200)]
Moving LN_MAX_MSG_LEN const to the actual use place

4 years agoRemoving spec quotation from LN_MAX_MSG_LEN definition
Dr. Maxim Orlovsky [Tue, 21 Jul 2020 12:11:46 +0000 (14:11 +0200)]
Removing spec quotation from LN_MAX_MSG_LEN definition

4 years agoRemoving duplicated code in message decryption test cases
Dr. Maxim Orlovsky [Tue, 21 Jul 2020 12:00:02 +0000 (14:00 +0200)]
Removing duplicated code in message decryption test cases

4 years agoTest cases for message encryption/decryption size limits
Dr. Maxim Orlovsky [Tue, 21 Jul 2020 11:48:37 +0000 (13:48 +0200)]
Test cases for message encryption/decryption size limits

4 years agoFixing panic messages for decryption procedures
Dr. Maxim Orlovsky [Tue, 21 Jul 2020 11:39:39 +0000 (13:39 +0200)]
Fixing panic messages for decryption procedures

4 years agoMaking message size limit an exportable constant
Dr. Maxim Orlovsky [Sun, 17 May 2020 14:02:56 +0000 (16:02 +0200)]
Making message size limit an exportable constant

4 years agoMerge pull request #656 from erasmospunk/typo
Matt Corallo [Thu, 16 Jul 2020 17:00:32 +0000 (10:00 -0700)]
Merge pull request #656 from erasmospunk/typo

Fix typo

4 years agoFix typo
John L. Jegutanis [Wed, 15 Jul 2020 13:43:18 +0000 (16:43 +0300)]
Fix typo

4 years agoMerge pull request #652 from lightning-signer/commitment-secret
Matt Corallo [Thu, 2 Jul 2020 18:31:31 +0000 (11:31 -0700)]
Merge pull request #652 from lightning-signer/commitment-secret

ChannelKeys provides individual commitment secrets

4 years agoChannelKeys provides individual commitment secrets
Devrandom [Wed, 1 Jul 2020 10:31:53 +0000 (03:31 -0700)]
ChannelKeys provides individual commitment secrets

4 years agoMerge pull request #650 from TheBlueMatt/2020-06-fix-build
Matt Corallo [Sun, 28 Jun 2020 18:01:43 +0000 (18:01 +0000)]
Merge pull request #650 from TheBlueMatt/2020-06-fix-build

Fix silent merge conflict between bcd65be and 940d7ac

4 years agoFix unused-parentheses warnings newer rustcs have added 2020-06-fix-build
Matt Corallo [Sun, 28 Jun 2020 17:21:45 +0000 (13:21 -0400)]
Fix unused-parentheses warnings newer rustcs have added

4 years agoFix silent merge conflict between bcd65bed7bd9a and 940d7ac8d897589
Matt Corallo [Sun, 28 Jun 2020 17:19:40 +0000 (13:19 -0400)]
Fix silent merge conflict between bcd65bed7bd9a and 940d7ac8d897589

4 years agoMerge pull request #647 from valentinewallace/test-remote-fee-spike-buffer-violation
Matt Corallo [Sat, 27 Jun 2020 00:41:35 +0000 (00:41 +0000)]
Merge pull request #647 from valentinewallace/test-remote-fee-spike-buffer-violation

Test remote fee spike buffer violation

4 years agoMerge pull request #638 from TheBlueMatt/2020-06-c-bindings-cleanups-2
Matt Corallo [Wed, 24 Jun 2020 00:16:28 +0000 (00:16 +0000)]
Merge pull request #638 from TheBlueMatt/2020-06-c-bindings-cleanups-2

Pre-C bindings cleanups (2)

4 years agoUse usize for transaction-position-in-block values 2020-06-c-bindings-cleanups-2
Matt Corallo [Sat, 13 Jun 2020 01:01:32 +0000 (21:01 -0400)]
Use usize for transaction-position-in-block values

We use them largely as indexes into a Vec<Transaction> so there's
little reason for them to be u32s. Instead, use them as usize
everywhere.

We also take this opportunity to add range checks before
short_channel_id calculation, as we could otherwise end up with a
bogus short_channel_id due to an output index out of range.

4 years agoDrop unused import in fuzz router.rs
Matt Corallo [Thu, 11 Jun 2020 20:05:41 +0000 (16:05 -0400)]
Drop unused import in fuzz router.rs

4 years agoTake &NetworkGraph as input in get_route, not a NetGraphMsgHandler
Matt Corallo [Thu, 11 Jun 2020 19:40:28 +0000 (15:40 -0400)]
Take &NetworkGraph as input in get_route, not a NetGraphMsgHandler

This was just an oversight when route calculation was split up into
parts - it makes no sense for get_route to require that we have a
full route message handler, only a network graph (which can always
be accessed from a NetGraphMsgHandler anyway).

4 years agoMake ChainWatchInterface::filter_block return only idxes, not refs
Matt Corallo [Sat, 13 Jun 2020 21:33:21 +0000 (17:33 -0400)]
Make ChainWatchInterface::filter_block return only idxes, not refs

Instead of making the filter_block fn in the ChainWatchInterface
trait return both a list of indexes of transaction positions within
the block and references to the transactions themselves, return
only the list of indexes and then build the reference list at the
callsite.

While this may be slightly less effecient from a memory locality
perspective, it shouldn't be materially different.

This should make it more practical to generate bindings for
filter_block as it no longer needs to reference Rust Transaction
objects that are contained in a Rust Block object (which we'd
otherwise just pass over the FFI in fully-serialized form).

4 years agoAvoid references to primitives and add NetworkGraph::new()
Matt Corallo [Thu, 11 Jun 2020 19:33:20 +0000 (15:33 -0400)]
Avoid references to primitives and add NetworkGraph::new()

non-mut references to primitives are only excess overhead, so
there's not much reason to ever have them. As a nice bonus, it also
is one less thing to worry about when generating C bindings

4 years agoAvoid use std and use std::fmt and fmt:: instead in network_graph
Matt Corallo [Thu, 11 Jun 2020 19:32:23 +0000 (15:32 -0400)]
Avoid use std and use std::fmt and fmt:: instead in network_graph

This is more consistent with the way we use std::cmp over the
codebase and avoids `use std`, which is only actually needed to
support older rustcs, so feels a bit awkward.

4 years agoAdd test for remote fee spike buffer violations.
Valentine Wallace [Mon, 22 Jun 2020 19:30:10 +0000 (15:30 -0400)]
Add test for remote fee spike buffer violations.

HTLCs that violate the remote's fee spike buffer should be failed.

4 years agoLog fee spike buffer violations.
Valentine Wallace [Mon, 22 Jun 2020 19:29:29 +0000 (15:29 -0400)]
Log fee spike buffer violations.

4 years agoReduce RwLock usage in public interface of NetworkGraph
Matt Corallo [Thu, 11 Jun 2020 19:23:35 +0000 (15:23 -0400)]
Reduce RwLock usage in public interface of NetworkGraph

This isn't a big difference in the API, but it avoids needing to
wrap a given NetworkGraph in a RwLock before passing it, which
makes it much easier to generate C bindings for.

4 years agoDrop a useless import of std::result::Result since its always there
Matt Corallo [Sun, 31 May 2020 03:23:47 +0000 (23:23 -0400)]
Drop a useless import of std::result::Result since its always there

4 years agoDrop some unnecessary lifetime specifiers in return type definitions
Matt Corallo [Sun, 31 May 2020 03:18:35 +0000 (23:18 -0400)]
Drop some unnecessary lifetime specifiers in return type definitions

In general, we don't need an explicit lifetime when doing something
like:
fn get_thing(&self) -> &Thing { &self.thing }.

This also makes it easier to reason about what's going on in the
bindings generation.

4 years agoChannelMonitor::get_funding_txo returns both the txid and scriptPK
Matt Corallo [Wed, 10 Jun 2020 03:00:30 +0000 (23:00 -0400)]
ChannelMonitor::get_funding_txo returns both the txid and scriptPK

... instead of only the txid.

This is another instance of it not being possible to fully
re-implement SimpleManyChannelMonitor using only public methods. In
this case you couldn't properly register outpoints for monitoring
so that the funding transaction would be matched.

4 years agoMerge pull request #641 from jkczyz/2020-06-deterministic-test-logging
Matt Corallo [Sun, 21 Jun 2020 19:44:56 +0000 (19:44 +0000)]
Merge pull request #641 from jkczyz/2020-06-deterministic-test-logging

Make test output deterministic

4 years agoMake test output deterministic
Jeffrey Czyz [Wed, 17 Jun 2020 15:29:30 +0000 (08:29 -0700)]
Make test output deterministic

Tests use sources of randomness to produce seeds, preimages, secrets,
and ephemeral data. However, this makes comparing logs between different
test runs difficult. Remove uses of random number generators and the
current time in favor of fixed values in order to make the test output
deterministic.

4 years agoMerge pull request #639 from valentinewallace/fix-feerate-new-channel
Matt Corallo [Tue, 16 Jun 2020 19:51:24 +0000 (19:51 +0000)]
Merge pull request #639 from valentinewallace/fix-feerate-new-channel

Use our actual feerate in open_channel messages, not a new one

4 years agoSwitch all feerate u64's to u32's.
Valentine Wallace [Mon, 15 Jun 2020 21:28:01 +0000 (17:28 -0400)]
Switch all feerate u64's to u32's.

The protocol only allows a u32, so if we received or sent something
larger it would be an issue (though it's unlikely).

4 years agoUse our actual feerate in open_channel messages, not a new one
Valentine Wallace [Sat, 13 Jun 2020 20:46:25 +0000 (16:46 -0400)]
Use our actual feerate in open_channel messages, not a new one

When we were sending an open_channel messages we were asking the
feerate estimator for a new value instead of using the one we had.
If the feerate estimator gave a different value than the one it did
when we created the Channel struct, we'd start out-of-sync with our
counterparty and blow up on funding_signed. Even worse, the
ConfirmationTarget used was different, so its highly likely they
would disagree.

Also remove newly unused fee estimator parameter from get_open-channel
API.

Co-authored-by: Matt Corallo <git@bluematt.me>
Co-authored-by: Valentine Wallace <vwallace@protonmail.com>
4 years agoMerge pull request #577 from valentinewallace/fix-onchain-fee-check-htlcs
Matt Corallo [Mon, 15 Jun 2020 20:17:03 +0000 (20:17 +0000)]
Merge pull request #577 from valentinewallace/fix-onchain-fee-check-htlcs

Incl tx fee when calcing inbound+outbound HTLC limits on channels