The spec doesn't say to do this, but it makes sense, otherwise
they'll never be able to mutually close the channel.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
We were always prefixing the `message` field with the internal type
prefix 0x0407, followed by the length prefix. Neither is needed since
the type being constant is of no interest to the plugin and the length
being implicit due to the JSON-encoding.
Reported-by: Ilya Evdokimov
Changelog-Fixed: plugin: The `custommsg` hook no longer includes the internal type prefix and length prefix in its `payload`
Changelog-Deprecated: plugin: The `message` field on the `custommsg` hook is deprecated in favor of the `payload` field, which skips the internal prefix.
They need to specify fees to get a channel before this, but it's possible.
The test revealed no surprises (other than the last change).
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Saves a great deal of confusion for regtest users.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Changelog-Changed: JSON-RPC: If bitcoind won't give a fee estimate in regtest, use minimum.
This was likely missed because we don't run the tests under valgrind anymore
due to time constraints. I do run them on a semi-regular basis, which is why
I found this.
This avoids spamming the logs. We also remove the duplicate debug
logs on self-disable (plugin_kill logs it for us).
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
We currently log every kill at INFO level, even if it's during shutdown.
Change those to debug, but lift those where we got a malformed response.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
We use waitpid() manually for subdaemons, so we need to step
around that (otherwise we could simply ignore them).
We could destroy subdaemons only once they've exited, but
that works badly with the sd->conn, which will be freed
when error (i.e. close) is detected, so the current code
is probably the best compromise.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
If a plugin died due to connection close, we'd always say
"Plugin exited before completing handshake.", which was often
wrong.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
If not all nodes are up-to-date with the new blocks, they can reject
announcements:
```
lightningd-4: 2021-02-23T02:02:47.832Z DEBUG 022d223620a359a47ff7f7ac447c85c46c923da53389221a0054c11c1e3ca31d59-gossipd: Ignoring future channel_announcment for 124x1x0 (current block 123)
lightningd-4: 2021-02-23T02:02:47.848Z DEBUG lightningd: Adding block 133: 2d950451211398de9c10bf9df7eb53b385390eca31e306bc8fc1387b53d9f9a2
lightningd-4: 2021-02-23T02:02:47.865Z DEBUG 022d223620a359a47ff7f7ac447c85c46c923da53389221a0054c11c1e3ca31d59-gossipd: Bad gossip order: WIRE_CHANNEL_UPDATE before announcement 124x1x0/0
lightningd-4: 2021-02-23T02:02:47.866Z DEBUG 022d223620a359a47ff7f7ac447c85c46c923da53389221a0054c11c1e3ca31d59-gossipd: Bad gossip order: WIRE_CHANNEL_UPDATE before announcement 124x1x0/1
```
Technically, this change is not sufficient either, since *gossipd* might
not know about new block yet. But it makes this case less likely.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
We get a label clash: easy, just re-serve:
```
2021-02-18T04:29:37.474Z **BROKEN** plugin-offers: Failed invoice_request lnr1qgsqvgnwgcg35z6ee2h3yczraddm72xrfua9uve2rlrm9deu7xyfzrcyyqwtp0rmsgquvuacqcl5cdfzwzmu3v8tqgvpqs8e80dlmxm7ey4xwrqdsqqqqqqqqqqqqqqqq2pqqfqpqynzqzx9rylzy40ernj4jzc3p2dwy3n8x6lqeaywwk725ghx4kx63pcfxgg2z3nsn80jzge06nt3ks8pr6rvnujq48376lpmrr3cq04nurpy783eyr0awh5773lrlmjek07rjf0nx4g9235ulkcs7jp2h5gumjyquhadh846da3jptxm9g0qz5lne4hjhag for offer 1cb0bc7b8201c673b8063f4c352270b7c8b0eb02181040f93bdbfd9b7ec92a67: Got JSON error: {\"code\":900,\"message\":\"Duplicate label\",\"data\":{\"label\":\"1cb0bc7b8201c673b8063f4c352270b7c8b0eb02181040f93bdbfd9b7ec92a67-08c5193e2255f91ce5590b110a9ae2466736be0cf48e75bcaa22e6ad8da88709-1\",\"bolt12\":\"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\",\"payment_hash\":\"396250395aa046f4c58162a5ae31952d01bd0c8e5213f32e748ec428a9379cd2\",\"msatoshi\":7700446,\"amount_msat\":\"7700446msat\",\"status\":\"unpaid\",\"description\":\"Weekly coffee for rusty!\",\"expires_at\":1614832137,\"local_offer_id\":\"1cb0bc7b8201c673b8063f4c352270b7c8b0eb02181040f93bdbfd9b7ec92a67\"}}
```
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
We used to only set it for single-use offers (where it's required),
but it's still interesting for multi-use offers, so let's keep it
there.
We also put this field in the documentation.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
We assume if they set this to 0 (which nobody did previously), they're
using it as a modern flag and use it to indicate when they're
finished. Otherwise, we count how many blocks they've sent and use
that to determine whether they've finished.
See: https://github.com/lightningnetwork/lightning-rfc/pull/826
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Changelog-Changed: Protocol: we use `sync_complete` for gossip range query replies, with detection for older spec nodes.
This ensures that after the migration in the previous commit we never
insert a new htlc with a null value.
Fixes: #4363
Reported-by: Zoltán Gálli <@gallizoltan>
Changelog-Fixed: db: Fixed an access to a NULL-field in the `channel_htlcs` table and resulting warning.
wally offers up `wally_clone_psbt` but it's a bit clunky (requires
checking return value, starting/stopping the wally_allocation context)
Helper method wraps this all up nice + neat!
Nested `with` exception checks don't work; fix flakes i'm seeing with
valgrind causing failures because blockchain not up to date when
`fundchannel` called.
We weren't waiting for the channel to get to normal state:
E pyln.client.lightning.RpcError: RPC call failed: method: pay, payload: {'bolt11': 'lnbcrt20m1pszfpezpp549um3vuyt52rgea32g7u55a5fv29yv7t94cmr2f0vjgcc33m3dvqdqzdgxqyjw5qcqp9sp59vu4tat2n53ylzrgxa95s5wu2s885a9llud64c0f6gjqts7h6tks9qy9qsq7u0j3w2h3sxxp9axpjxkz525znjn0t92gnrgk7y6plyq39zw9994g88xxjx52egk4965dp5qt2w08hk009eq9hm8nykwmxe7r705k8gpkqu9mw'}, error: {'code': 210, 'message': 'Ran out of routes to try after 1 attempt: see `paystatus`', 'attempts': [{'status': 'failed', 'failreason': 'Cannot attempt payment, we have no channel to which we can add an HTLC', 'partid': 1, 'amount': '2000000000msat'}]}
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
A new target `check-gen-updated` to verify that all derived/generated
files that were modified were also checked in. This is used on CI to
check, and is not added to `check` since it'll complain on dirty
trees, i.e., before the devs check in their changes.
Suggested-by: Rusty Russell <@rustyrussell>
The leak exists if we `tal_free` the result array onto another parent,
but the `ctx` we allocated on is still valid. This leads to a
temporary gap in the ownership tree which is then reported as the
following error:
```text
- Node /tmp/ltests-ufn3ox3p/test_htlc_out_timeout_1/lightning-1/ has memory leaks: [
{
"backtrace": [
"ccan/ccan/tal/tal.c:442 (tal_alloc_)",
"ccan/ccan/tal/tal.c:471 (tal_alloc_arr_)",
"ccan/ccan/tal/tal.c:799 (tal_dup_)",
"ccan/ccan/tal/str/str.c:18 (tal_strdup_)",
"wallet/wallet.c:1652 (wallet_state_change_get)",
"lightningd/peer_control.c:869 (json_]add_channel)",
"lightningd/peer_control.c:1319 (json_add_peer)",
"lightningd/peer_control.c:1348 (json_listpeers)",
"lightningd/jsonrpc.c:643 (command_exec)",
"lightningd/jsonrpc.c:753 (rpc_command_hook_callback)",
"lightningd/plugin_hook.c:288 (plugin_hook_call_)",
"lightningd/jsonrpc.c:808 (plugin_hook_call_rpc_command)",
"lightningd/jsonrpc.c:888 (parse_request)",
"lightningd/jsonrpc.c:979 (read_json)",
"ccan/ccan/io/io.c:59 (next_plan)",
"ccan/ccan/io/io.c:435 (io_do_always)",
"ccan/ccan/io/poll.c:300 (handle_always)",
"ccan/ccan/io/poll.c:377 (io_loop)",
"lightningd/io_loop_with_timers.c:24 (io_loop_with_timers)",
"lightningd/lightningd.c:1016 (main)"
],
"label": "wallet/wallet.c:1652:char[]",
"parents": [
"common/json_stream.c:29:struct json_stream",
"ccan/ccan/io/io.c:91:struct io_conn",
"lightningd/lightningd.c:116:struct lightningd"
],
"value": "0x556b0856ab68"
},
```
Changelog-None
LightningNode.join_nodes and LightningNode.openchannel internally
generate blocks, which causes nodes to be out of sync, and ignore
"future" announcements, because they haven't seen that block yet.
It was using a trick to only shut down the first node, and forgetting
about the others. This could lead to processes not being stopped
correctly and to test failures because the directory isn't cleaned up
correctly.
Now we use the executor to shut as many nodes as possible in parallel.
We split into quick smoke-tests, normal-tests, and valgrind-tests. The
first provides a quick-abort preventing overusing resources on what
will be a failed run anyway.
We'd like to have some more structure to the tests, with smoke-tests
going first, followed by valgrind and normal tests running in
parallel. So to cut down on the copy-pasta we extract the scripts from
the yaml, into their own files from where they can be easily invoked.
