Changelog-Changed: `fundchannel_cancel` will now succeed even when executed while a `fundchannel_complete` is ongoing; in that case, it will be considered as cancelling the funding *after* the `fundchannel_complete` succeeds.
Let me introduce the concept of "Sequential Consistency":
All operations on parallel processes form a single total order agreed upon by all processes.
So for example, suppose we have parallel invocations of `fundchannel_complete` and `fundchannel_cancel`:
+--[fundchannel_complete]-->
|
--[fundchannel_start]-+
|
+--[fundchannel_cancel]---->
What "Sequential Consistency" means is that the above parallel operations can be serialized as a single total order as:
--[fundchannel_start]--[fundchannel_complete]--[fundchannel_cancel]-->
Or:
--[fundchannel_start]--[fundchannel_cancel]--[fundchannel_complete]-->
In the first case, `fundchannel_complete` succeeds, and the `fundchannel_cancel` invocation also succeeds, sending an `error` to the peer to make them forget the chanel.
In the second case, `fundchannel_cancel` succeeds, and the succeeding `fundchannel_complete` invocation fails, since the funding is already cancelled and there is nothing to complete.
Note that in both cases, `fundchannel_cancel` **always** succeeds.
Unfortunately, prior to this commit, `fundchannel_cancel` could fail with a `Try fundchannel_cancel again` error if the `fundchannel_complete` is ongoing when the `fundchannel_cancel` is initiated.
This violates Sequential Consistency, as there is no single total order that would have caused `fundchannel_cancel` to fail.
This commit is a minimal patch which just reschedules `fundchannel_cancel` to occur after any `fundchannel_complete` that is ongoing.
We passed below the floor when the user specified `1000perkb`.
Matt Whitlock says :
I was withdrawing with feerate=1000perkb, which should be the minimum-allowed fee rate. Indeed, bitcoin-cli getmempoolinfo reports:
{
"loaded": true,
"size": 15097,
"bytes": 9207924,
"usage": 32831760,
"maxmempool": 64000000,
"mempoolminfee": 0.00001000,
"minrelaytxfee": 0.00001000
}
Changelog-fixed: rpc: The `feerate` parameters now correctly handle the standardness minimum when passed as `perkb`.
Signed-off-by: Antoine Poinsot <darosior@protonmail.com>
Reported-by: Matt Whitlock
We erase peer data after the last channel close transaction for that
peer is 100 blocks deep. We were failing to finish the migration because
the peer_id lookup on these was failing.
Now we ignore any channel with a null peer_id.
Fixes#3768
We use a database snapshot with 3 channels -- two of which have HTLCs
dangling and one is an initial open channel tx in the 'old' tx hex
format in last_tx and confirm that they are successfully updated to PSBT
format on start.
Spec is wrong (it says it should be compulsory), and Eclair doesn't set it
at all, leading to an error when they send their announcement_signatures.
Fixes: #3703
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Changelog-changed: large-channels: negotiate successfully with Eclair nodes.
There are various places where our tests failed with
--enable-expimental-features. And our plugin test overlapped an
existing feature.
We make our expected_feature functions more generic, and use them
everywhere.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Much nicer for grepping, since `{ "foo": { "bar": [7] } }` is turned into
`foo.bar[0]=7`.
Changelog-Added: cli: New `--flat` mode for easy grepping.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
-H removes the top-level if there's only one, and 'format-hint'
breaks this heuristic, so we end up with:
```
help=command=autocleaninvoice [cycle_seconds] [expired_by]
category=plugin
description=Set up autoclean of expired invoices.
verbose=Perform cleanup every {cycle_seconds} (default 3600), or disable autoclean if 0. Clean up expired invoices that have expired for {expired_by} seconds (default 86400).
command=check command_to_check
```
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
We've been seeing some Travis timeouts under VALGRIND, with the
10 second timeout here: use TIMEOUT as per standard.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
It's possible for our peer to publish a commitment tx that has already
updated our balance for an htlc before we've completed removing it from
our commitment tx (aka before we've updated our balance). This used to
crash, now we just update our balance (and the channel balance logs!)
and keep going.
If they've removed anything from our balance, we'll end up counting it
as chain_fees below. Not ideal but fine... probably.
This moves the notification for our coin spends from when it's
successfully submited to the mempool to when they're confirmed in a
block.
We also add an 'informational' notice tagged as `spend_track` which
can be used to track which transaction a wallet output was spent in.
Previously we were annotating every movement with the blockheight of
lightningd at notification time. Which is lossy in terms of info, and
won't be helpful for reorg reconciliation. Here we switch over to
logging chain moves iff they've been confirmed.
Next PR will fix this up for withdrawals, which are currently tagged
with a blockheight of zero, since we log on successful send.
On node start we replay onchaind's transactions from the database/from
our loaded htlc table. To keep things tidy, we shouldn't notify the
ledger about these, so we wrap pretty much everything in a flag that
tells us whether or not this is a replay.
There's a very small corner case where dust transactions will get missed
if the node crashes after the htlc has been added to the database but
before we've successfully notified onchaind about it.
Notably, most of the obtrusive updates to onchaind wrappings are due to
the fact that we record dust (ignored outputs) before we receive
confirmation of its confirmation.
Mostly we update existing tests to account for channel balances. In a
few places, new tests were needed as there wasn't an existing pathway
that tested the chain-fees for a few penalty cases
For cheats, we do a little bit of weird accounting. First we 'update'
our on-ledger balance to be the entirety of the channel's balance. Then,
as outputs get resolved, we record the fees and outputs as withdrawals
from this amount.
It's possible that they might successfully 'cheat', in which case we
record those as 'penalty' but debits (not credits).
The previous patch changed the gossip_store, but in a trivial way.
The next patch will implement upgrading, so this is the test.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
And the percentage of the initial amount, not the constently increasing
one !
Changelog-Fixed: pay: we now respect maxfeepercent, even for tiny amounts.
Signed-off-by: Antoine Poinsot <darosior@protonmail.com>
The previous implementation was a bit lazy: in particular, since we didn't
remember the disabled plugins, we would load them on rescan.
Changelog-Changed: config: the `plugin-disable` option works even if specified before the plugin is found.
That's more convenient for most callers, which don't need a fmt.
Fixed-by: Darosior <darosior@protonmail.com>
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
This is what I expected from plugin_kill, and now all the callers do the
equivalent anywat, it's easy.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
This means we now clean up options in startup plugins (that was only
done by dynamic code!), and now they both share the 60 second timeout
instead of 20 seconds for dynamic.
For the dynamic case though, it's 60 seconds to both complete
getmanifest and init, which seems fair.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
We modify the slow_init() so it doesn't go too slowly for this test.
This demonstrates a crash, where we currently try to fail a command
multiple times.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
This actually passes fine, but it's an interesting case to test.
Fixed-by: Darosior <darosior@protonmail.com>
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Telling `lightningd` to pass a `-datadir` to `bitcoin-cli` so it doesn't go
snooping where it doesn't belong (i.e., the user's home directory and config).
Changelog-None
Suggested-by: Simon Vrouwe <@SimonVrouwe>
Signed-off-by: Christian Decker <@cdecker>
I noticed the following in logs for tests/test_connection.py::test_feerate_stress:
```
DEBUG 022d223620a359a47ff7f7ac447c85c46c923da53389221a0054c11c1e3ca31d59-chan#1: Failing HTLC 18446744073709551615 due to peer death
DEBUG 022d223620a359a47ff7f7ac447c85c46c923da53389221a0054c11c1e3ca31d59-chan#1: local_routing_failure: 8194 (WIRE_TEMPORARY_NODE_FAILURE)
```
This is because it reports the (transient) node_failure error, because
our channel_failure message is incomplete. Fix this wart up.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
If we don't wait for close tx to reach mempool, it might not get to
depth 100, and we don't get 'onchaind complete, forgetting peer'.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
We didn't wait until l2 processed the final state of HTLC #2, so
it might not include it when it drops onchain, leading to us only
getting 3 (not 4) sendrawtx calls.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Previously we've used the term 'funder' to refer to the peer
paying the fees for a transaction; v2 of openchannel will make
this no longer true. Instead we rename this to 'opener', or the
peer sending the 'open_channel' message, since this will be universally
true in a dual-funding world.
They were looking for specific amounts which proved to be rather flaky. Now
they look for specific outputs being available in the `listfunds` result after
everything was settled.
The new `keysend` plugin modifies the node features that we send to
peers. This commit breaks out the 'expected_features' we use for tests
to encompass this differentiation.
The documentation was wrong, and I copied my mistake to `libplugin` where it
was then ignored instead of ORed into the node's featurebits. This fixes both.
As discussed with Christian, prepending the length to the payload returned
is awkward, but it's the only way to set a legacy payload. As this will
be soon deprecated, simplify the external API.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
We weren't actually waiting until l3 got the channel_update from l2,
so it might not be able to create the routehint.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
When route returns a result which is too expensive, we try to figure out which
hop is most expensive to exclude it for next time.
If it's a single-hop route, we don't count it, since the first hop is free.
That's not usually a problem, since single-hop routes can't exceed our limits
(they're always "free"!).
But if we are using a routehint, the total cost could exceed our limits,
even if the start of the routehint is a single hop away.
This reproduces that test case.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
This makes testing easier, and makes sense: lightningd might not
*know* about other connected channels, depending on gossip, but if the
user specifies it we should obey it.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Changelog-Changed: JSON: `invoice` `exposeprivatechannels` now includes explicitly named channels even if they seem like dead-ends.
This is what actually lets us pay blinded invoices.
Unfortunately, our internal logic assumes every hop in a path has a
next `short_channel_id`, so we have to use a dummy. This is
sufficient for testing, however.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Sending update_fee immediately after channel establishment seems to
upset LND, so work around it by deferring it. The reason we increase
the fee after establishment is because now we might need to close the
channel in a hurry due to htlcs, but until there are htlcs that's
unnecessary.
Fixes: #3596
Changelog-Changed: Added workaround for lnd rejecting our commitment_signed when we send an update_fee after channel confirmed.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
We now track all pending RPC passthrough calls, and terminate them with an
error if the plugin dies.
Changelog-Fixed: JSON-RPC: Pending RPC method calls are now terminated if the handling plugin exits prematurely.
The spec states that invoices with an amount, but lacking a multiplier, should
be interpreted as integer Bitcoin amounts:
`amount`: optional number in that currency, followed by an optional
`multiplier` letter. The unit encoded here is the 'social' convention of a
payment unit -- in the case of Bitcoin the unit is 'bitcoin' NOT satoshis.
Suggested-by: Stefano Pellegrini <@St333p>
Signed-off-by: Christian Decker <@cdecker>
Changelog-Fixed: invoice: The invoice parser assumed that an amount without a multiplier was denominated in msatoshi instead of bitcoins.
We had them split according the separate use-cases:
- testing
- doc-gen
- wire-gen
But that was causing new contributors to miss some dependencies when they
first got hacking. So this consolidates all of our own dependencies in a root
requirements.txt, with the notable exception of `pyln-client`, `pyln-testing`
and `pyln-proto` which are distributed as PyPI modules and therefore have
their own dependencies that need to be tracked in the module root.
Closes#3518
Note that now we check capacity once we've figured out which peer, which
broke a test (we returned "unknown peer" instead of "capacity exceeded"),
so we rework that too.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Shows what features we use in various contexts, including those added
by plugins in getmanifest.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Changelog-Added: Plugin: `feature_set` object added to `init`
A CONSERVATIVE/3 target for them.
Some noisy changes to the tests as we had to update the estimatesmartfee
mock.
Changelog-Changed: We now use a higher feerate for resolving onchain HTLCs and for penalty transactions
Changelog-Fixed: Passing 0 as minconf to withdraw allows you to use unconfirmed transaction outputs, even if explicitly passed as the `utxos` parameter
This is to prepare for dynamic features, including making plugins first
class citizens at setting them.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
This restriction was removed from the spec as of
86c2ebcc5973a4133d3ce4d80ae1c203061a1646.
We also fix up some strange formatting in that part of the documentation.
Changelog-changed: We now announce multiple addresses of the same type, if given.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Changelog-Removed: JSON: `fundchannel` and `fundchannel_start` `satoshi` parameter removed (renamed to `amount` in 0.7.3).
This adapts our fee estimations requests to the Bitcoin backend to the
new semantic, and batch the requests.
This makes our request for fees much simpler, and leaves some more
flexibility for a plugin to do something smart (it could still lie before
but now it's explicit, at least.) as we don't explicitly request
estimation for a specific mode and a target.
Changelog-Changed: We now batch the requests for fee estimation to our Bitcoin backend.
Changelog-Changed: We now get more fine-grained fee estimation from our Bitcoin backend.
We kept track of an URGENT, a NORMAL, and a SLOW feerate. They were used
for opening (NORMAL), mutual (NORMAL), UNILATERAL (URGENT) transactions
as well as minimum and maximum estimations, and onchain resolution.
We now keep track of more fine-grained feerates:
- `opening` used for funding and also misc transactions
- `mutual_close` used for the mutual close transaction
- `unilateral_close` used for unilateral close (commitment transactions)
- `delayed_to_us` used for resolving our output from our unilateral close
- `htlc_resolution` used for resolving onchain HTLCs
- `penalty` used for resolving revoked transactions
We don't modify our requests to our Bitcoin backend, as the next commit
will batch them !
Changelog-deprecated: The "urgent", "slow", and "normal" field of the `feerates` command are now deprecated.
Changelog-added: The fields "opening", "mutual_close", "unilateral_close", "delayed_to_us", "htlc_resolution" and "penalty" have been added to the `feerates` command.
pytest captures the output by monkey patching out `sys.stdout`. This may
conflict with our use of `sys.stdout` when configuring logging, resulting in
the "Write to closed file" issue that is spamming the logs. By making the
logging configuration a fixture hopefully we always use the correct
stdout (after pytest has monkey-patched it).
Some tests may not spawn a node at all, so make sure that our assumption that
the directory exists in the fixture cleanup is correct by creating the
directory.
