Commit Graph

1791 Commits

Author SHA1 Message Date
Rusty Russell
14eddb95ba fundchannel, multifundchannel: reserve inputs for two weeks, not 12 hours.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Changelog-Changed: Plugins: `fundchannel` and `multifundchannel` will now reserve funding they use for 2 weeks instead of 12 hours.
2021-05-26 15:08:01 +09:30
Rusty Russell
95f77f119b pytest: test double-spending an opening tx input.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
2021-05-26 15:08:01 +09:30
Rusty Russell
f24dc9173d wallet: add "reserved_to_block" field to listfunds.
We already have this field in reserveinputs and unreserveinputs.

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Changelog-Added: JSON-RPC: `listfunds` has a new `reserved_to_block` field.
2021-05-26 15:08:01 +09:30
Nalin Bhardwaj
94419c7d73 common: Check chain hash in gossip_timestamp_filter
Changelog-Fixed: Validate chain hash for gossip_timestamp_filter messages
2021-05-26 10:43:50 +09:30
Rusty Russell
96bd0961a4 pay: fix corner case where we eliminate our own hint.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Changelog-Fixed: `pay`: Fix occasional crash paying an invoice with a routehint to us.
2021-05-26 08:07:38 +09:30
Rusty Russell
df594be80a pytest: add message check to test_blockheight_disagreement
Make sure there was an actual disagreement!

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
2021-05-25 16:25:17 +02:00
Rusty Russell
61d146a2e7 pytest: fix gossip flake in test_restart_many_payments
```
2021-05-23T09:18:56.9768962Z lightningd-4: 2021-05-23T08:33:31.918Z DEBUG   0266e4598d1d3c415f572a8488830b60f7e744ed9235eb0b1ba93283b315c03518-gossipd: Ignoring future channel_announcment for 103x4x0 (current block 102)
2021-05-23T09:18:56.9771608Z lightningd-4: 2021-05-23T08:33:31.918Z DEBUG   0266e4598d1d3c415f572a8488830b60f7e744ed9235eb0b1ba93283b315c03518-gossipd: Bad gossip order: WIRE_CHANNEL_UPDATE before announcement 103x4x0/0
2021-05-23T09:18:56.9774035Z lightningd-4: 2021-05-23T08:33:31.919Z DEBUG   0266e4598d1d3c415f572a8488830b60f7e744ed9235eb0b1ba93283b315c03518-gossipd: Bad gossip order: WIRE_CHANNEL_UPDATE before announcement 103x4x0/1
```

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
2021-05-25 16:25:17 +02:00
Antoine Poinsot
fe8074c8c3 Refuse to parse v2 onion addresses without deprecated_apis
Tor v2 hidden services have been deprecated for a while:
https://blog.torproject.org/v2-deprecation-timeline .

This prevents user from being able to set them in the configuration
and to connect to them while still letting us be able to parse them
for gossip.

Changelog-Deprecated: lightningd: v2 Tor addresses.  Use v3.  See https://blog.torproject.org/v2-deprecation-timeline.

Signed-off-by: Antoine Poinsot <darosior@protonmail.com>
2021-05-24 20:22:45 +09:30
Rusty Russell
d70661d32e pytest: check nonstandard onion generation.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
2021-05-24 12:52:19 +02:00
niftynei
9d9fdfdfa6 rbf-test: confirm that rbf'd inflight is opened and closed as expected
We weren't testing that a non-tip inflight would get opened correctly.
Until now, that is.
2021-05-24 12:17:48 +09:30
niftynei
d6bd6cc5cf open-rbf: we broadcast all the commitments for a channel when closed
If you drop an rbf'd channel to chain (before any updates have been
made) we should drop *all* of the inflights to chain
2021-05-24 12:17:48 +09:30
niftynei
dff9516cad rbf-tests: check every inflight has signed commitment produced
Little check to make sure that we can produce a signed commitment tx for
every inflight we've got saved.
2021-05-24 12:17:48 +09:30
niftynei
4247ec3a05 inflights: save the whole psbt to the database
Otherwise we're missing info when we go to broadcast these and can't
properly sign the transaction to close it.

Found-by: @jasan
2021-05-24 12:17:48 +09:30
niftynei
727d7e9493 df-rbf: test that, with enough bumps, we can eventually add to mempool
We don't pass the minimum fee requirement the first few times we attempt
an RBF, so it fails. Keep going until actually replaces
2021-05-24 12:17:48 +09:30
niftynei
82fa3fa2ef df-rbf: order inflights by funding_feerate
When we re-populate from disk, we need to know what order to recreate the
inflights list in.

Fixes #4511
2021-05-24 12:17:48 +09:30
niftynei
8ae9f6ac9e df-tests: add better explainer for test logic 2021-05-24 12:17:48 +09:30
niftynei
bab5ef4aff df-test: test for a failed rbf attempt, currently crashes
An attempted + failed RBF results in a crashed/dead node
2021-05-24 12:17:48 +09:30
niftynei
e2867be609 df-tests: re-write test_funding_cancel_race for v2 commands
test_funding_cancel_race uses fundchannel_open etc; this new test does a
similar (but not exact thing, as 'aborts' dont work after an update
is confirmed) thing, using openchannel_update and openchannel_abort.
2021-05-23 17:42:09 +09:30
niftynei
2d4939fa6d df-tests: re-write the funding wallet corners test for v2 2021-05-23 17:42:09 +09:30
niftynei
c93bd5bd51 funder: test for adding utxos, calculating total available
Make sure that 100% available actually puts in 100% of what we can add
(note that this isn't 'all' -- we don't include uneconomic utxos)
2021-05-23 08:19:50 +09:30
Rusty Russell
14febd4ecc pytest: disable mpp test which is blocking merges.
I'm still diagnosing exactly why this fails, but it will need Christian's
help at this stage.

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
2021-05-23 08:13:27 +09:30
Rusty Russell
33736b860a lightningd: attach HTLC timeout to htlc itself, fix gratuitous disconnect bug.
We set the timeout on first HTLC, but didn't clear it if that HTLC failed.

It's saner to have a per-HTLC timeout (since that's what it is!) and
also our timer infra is specially coded to scale approximately infinitely so
trying to optimize this is vastly premature.

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Changelog-Fixed: Protocol: We would sometimes gratuitously disconnect 30 seconds after an HTLC failed.
2021-05-21 14:45:05 +09:30
Rusty Russell
11180e7aa1 pytest: add test for HTLC timeout on failed htlcs.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
2021-05-21 14:45:05 +09:30
Rusty Russell
e141b71675 pytest: fix flake in test_reconnect_no_update
If l2 didn't get FUNDING_LOCKED from l1 before it disconnected, it
won't be in state CHANNELD_NORMAL: it will be in DUALOPEND_AWAITING_LOCKIN.

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
2021-05-17 18:58:10 -05:00
niftynei
ce1e5bd388 df-test: make flakey test not as flakey
This would flake fairly regularly, what we really care about is
asserting that the l2 node is in CHANNELD_NORMAL state, while the l1
node hasn't progressed that far yet.
2021-05-12 11:25:41 +09:30
niftynei
e375932108 df-tests: test_coin_movement_notices, make work?
No idea how this slipped past the first time
2021-05-12 11:25:41 +09:30
niftynei
71a4a2e31c df: rework closing logic
Trying to put all the disconnect logic into the same path was a dumb
idea. If you asked to reconnect but passed in an 'unsaved' channel, we
would not call the 'reconnect' code.

Instead, we make a differentiation between "unsaved" channels
(ones that we haven't received commitment tx for) and handle the
disconnect for these separate from where we want to do a reconnect.
2021-05-12 11:25:41 +09:30
niftynei
efdc36c8e9 tests: mark test as requiring developer
hangs with EXP_DF when developer=0
2021-05-12 11:25:41 +09:30
niftynei
5cac36724e df-test: v2 has reconnects 2021-05-12 11:25:41 +09:30
niftynei
6d3fb11bc6 df-tests: patch for state == AWAITING_UNILATERAL problem
Found on CI where DEVELOPER=0 EXPERIMENTAL_DUAL_FUND=1,
as we turn off automatic reconnects when DEVELOPER=1

This test has been modified to make the error happen every time, and
then fixed.

lightningd-2: 2021-05-07T20:12:03.790Z DEBUG   0266e4598d1d3c415f572a8488830b60f7e744ed9235eb0b1ba93283b315c03518-chan#1: Peer has reconnected, state AWAITING_UNILATERAL
lightningd-2: 2021-05-07T20:12:03.812Z **BROKEN** lightningd: FATAL SIGNAL 6 (version e8b3f78)
lightningd-2: 2021-05-07T20:12:03.812Z **BROKEN** lightningd: backtrace: common/daemon.c:44 (send_backtrace) 0x56384ee072e9
lightningd-2: 2021-05-07T20:12:03.813Z **BROKEN** lightningd: backtrace: common/daemon.c:52 (crashdump) 0x56384ee0733b

----------------------------- Captured stderr call -----------------------------
lightningd: lightningd/peer_control.c:1100: peer_connected_hook_final: Assertion `channel->state == DUALOPEND_OPEN_INIT || channel->state == DUALOPEND_AWAITING_LOCKIN' failed.
lightningd: FATAL SIGNAL 6 (version e8b3f78)
0x56384ee072a1 send_backtrace
	common/daemon.c:39
0x56384ee0733b crashdump
	common/daemon.c:52
0x7f88486a020f ???
	???:0
0x7f88486a018b ???
	???:0
0x7f884867f858 ???
	???:0
0x7f884867f728 ???
	???:0
0x7f8848690f35 ???
	???:0
0x56384eddc94e peer_connected_hook_final
	lightningd/peer_control.c:1100
0x56384edea2ed plugin_hook_call_
	lightningd/plugin_hook.c:275
0x56384eddfeb8 plugin_hook_call_peer_connected
	lightningd/peer_control.c:1156
0x56384eddfeb8 peer_connected
	lightningd/peer_control.c:1209
0x56384edc30cd connectd_msg
	lightningd/connect_control.c:332
0x56384edebe6f sd_msg_read
	lightningd/subd.c:509
0x56384edebfb1 read_fds
	lightningd/subd.c:310
0x56384ee483b0 next_plan
	ccan/ccan/io/io.c:59
0x56384ee4885b do_plan
	ccan/ccan/io/io.c:407
0x56384ee488f8 io_ready
	ccan/ccan/io/io.c:417
0x56384ee4a23c io_loop
	ccan/ccan/io/poll.c:445
0x56384edcabda io_loop_with_timers
	lightningd/io_loop_with_timers.c:24
0x56384edce826 main
	lightningd/lightningd.c:1111
0x7f88486810b2 ???
	???:0
0x56384edb52ad ???
	???:0
0xffffffffffffffff ???
	???:0
2021-05-12 11:25:41 +09:30
niftynei
558abe288a tests: explicitly mark tests to run both as v1+v2
By default, tests only run as v1 unless marked as v2.

These tests we want to run as both v1+v2

Includes fixes to have tests pass
2021-05-12 11:25:41 +09:30
niftynei
dc758f616b tests: mark as v1 or v2 openchannel required 2021-05-12 11:25:41 +09:30
niftynei
3a2d602922 tests: add marker for v1/v2 channel opens
Tests that will only run when !EXPERIMENTAL_DUAL_FUND:

	@pytest.marker.openchannel('v1')
	def test_...()

Tests that will only run when EXPERIMENTAL_DUAL_FUND:

	@pytest.marker.openchannel('v2')
	def test_...()
2021-05-12 11:25:41 +09:30
niftynei
d0bbf07655 tests: not DEVELOPER -> mark.developer
Nicer syntaxtic sugar for marking pytests as 'developer required'
2021-05-12 11:25:41 +09:30
niftynei
b4e24ac8ba df: anchor outputs are on if EXP_DF
So we should treat it the same as EXPERIMENTAL_FEATURES
2021-05-12 11:25:41 +09:30
niftynei
484d6bde83 tests: move EXP_DF into the testing utils 2021-05-12 11:25:41 +09:30
niftynei
7a488cc11e df-tests: remove df_accepter plugin
The `funder` plugin, does the same stuff as this temporary test plugin,
so we move over to using that instead.
2021-05-12 11:25:41 +09:30
Rusty Russell
d8e68893f5 bcli: become less aggressive with onchain fee levels.
Users are more upset recently with the cost of unilateral closes
than they are the risk of being cheated.  While we complete our
anchor implementation so we can use low fees there, let's
get less aggressive (we already have 34 or 18 blocks to close
in the worst case).

The changes are:

- Commit transactions were "2 CONSERVATIVE" now "6 ECONOMICAL".
- HTLC resolution txs were "3 CONSERVATIVE" now "6 ECONOMICAL".
- Penalty txs were "3 CONSERVATIVE" now "12 ECONOMICAL".
- Normal txs were "4 ECONOMICAL" now "12 ECONOMICAL".

There can be no perfect levels, but we have had understandable
complaints recently about how high our default fee levels are.

Changelog-Changed: Protocol: channel feerates reduced to bitcoind's "6 block ECONOMICAL" rate.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
2021-05-11 11:25:16 +09:30
Rusty Russell
9825f32874 lightningd: implement --log-timestamps=false.
Fixes: #4494
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Changelog-Added: config: New option `log-timestamps` allow disabling of timestamp prefix in logs.
2021-05-05 17:19:19 -05:00
Christian Decker
c6fd849aa3 pay: Add notification for pay_success 2021-05-03 11:20:15 +09:30
Christian Decker
98aa3c3da7 plugin: Make unannounced notification topics no longer fatal
Since plugins will start sending them soon, and they are likely to get
it wrong sometimes, be a bit more lenient, warn them in the logs
instead and then make sure it doesn't accidentally work anyway.
2021-05-03 11:20:15 +09:30
Christian Decker
f963a6a551 libplugin: Add notification topics to plugin_main 2021-05-03 11:20:15 +09:30
Christian Decker
62e3358a5b plugin: Wrap custom notifications in a dict with additional origin
This should allow us to differentiate the origin of the notification,
and further prevent plugins from spoofing native notifications.
2021-05-03 11:20:15 +09:30
Christian Decker
083b41f090 plugin: Add a list of notification topics registered by plugin
We will eventually start emitting and dispatching custom notifications
from plugins just like we dispatch internal notifications. In order to
get reasonable error messages we need to make sure that the topics
plugins are asking for were correctly registered. When doing this we
don't really care about whether the plugin that registered the
notification is still alive or not (it might have died, but
subscribers should stay up and running), so we keep a list of all
topics attached to the `struct plugins` which gathers global plugin
information.
2021-05-03 11:20:15 +09:30
niftynei
29155c2fe8 tests: add test for funder options 2021-05-03 11:06:10 +09:30
niftynei
cd5970243a funder: add a plugin, funder. policies for dual-funding
Behold! An immaculately concepted plugin for configuring your node to do
amazing things*

*fund channel open requests

Changelog-Added:  Plugins: Add `funder` plugin, which allows you to setup a policy for funding v2 channel open requests. Requres --experimental-dual-fund option
2021-05-03 11:06:10 +09:30
niftynei
7c76363e20 openchannel2: add channel_max_msat to openchannel2 hook payload
Changelog-Added: Plugins: add a `channel_max_msat` value to the `openchannel2` hook. Tells you the total max funding this channel is allowed to have.
2021-05-03 11:06:10 +09:30
niftynei
69cc9201de plugin-tests: cleanup outdated openchannel2 hook fields
We removed/changed the fields on the openchannel2 hook but never
updated this test (which doesn't run with the CI for #reasons)
2021-05-03 11:06:10 +09:30
niftynei
0e753224ae tests: remove EXPERIMENTAL_FEATURE flag from openchannel2 hooks
These are now included in all builds
2021-05-03 11:06:10 +09:30
Christian Decker
9470ea3032 pytest: Fix racy test_unreachable_routehint
It was failing likely because we were racing on the log
processing. Using `wait_for_log` instead should address that.
2021-04-24 13:58:57 +09:30