Commit Graph

1671 Commits

Author SHA1 Message Date
Rusty Russell f0659d0ab0 pytest: fix false bad gossip issue in test_forward
We construct the route manually so we may not have the
channel_announcement yet.  But we can get an update from the error
packet, which can lead to:

```
2021-01-29T01:38:23.4767334Z            ValueError: 
2021-01-29T01:38:23.4767987Z            Node errors:
2021-01-29T01:38:23.4768767Z             - lightningd-1: had bad gossip messages
2021-01-29T01:38:23.4769512Z            Global errors:
2021-01-29T01:38:23.4770300Z 
2021-01-29T01:38:23.4771109Z contrib/pyln-testing/pyln/testing/fixtures.py:197: ValueError
...
2021-01-29T01:38:23.7820197Z lightningd-1: 2021-01-29T01:26:57.460Z DEBUG   gossipd: Extracted channel_update 01027217b3086ad9f3dee1fa55b94c5fd2a4b0637bec70ba727ba4151a8de5173ddc749db3502d41ab0ae164addc8fd013d2088b6a12a2f478ae0affa94d76d8845c06226e46111a0b59caaf126043eb5bbf28c34f3a5e332a1fc7b2b73cf188910f000067000001000160136459010000060000000000000000000000010000000a000000003b023380 from onionreply 100d0000007500887217b3086ad9f3dee1fa55b94c5fd2a4b0637bec70ba727ba4151a8de5173ddc749db3502d41ab0ae164addc8fd013d2088b6a12a2f478ae0affa94d76d8845c06226e46111a0b59caaf126043eb5bbf28c34f3a5e332a1fc7b2b73cf188910f000067000001000160136459010000060000000000000000000000010000000a000000003b023380
2021-01-29T01:38:23.7837450Z lightningd-1: 2021-01-29T01:26:57.461Z DEBUG   gossipd: Bad gossip order: WIRE_CHANNEL_UPDATE before announcement 103x1x1/0
```

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
2021-02-04 12:02:52 +10:30
Rusty Russell 06a54606a3 check-includes: allow redundant "config.h"
We should actually be including this (as it may define _GNU_SOURCE
etc) before any system headers.  But where we include <assert.h> we
often didn't, because check-includes would complain that the headers
included it too.

Weaken that check, and include config.h in C files before assert.h.

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
2021-02-04 12:02:36 +10:30
Rusty Russell 1a85edd207 lightningd: no longer forward failures to gossipd, let caller do it.
We fix up the test by using pay, instead of sendpay (and making pay log
the expected message).

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Changelog-Changed: sendpay no longer extracts updates from errors, the caller should do it from the `raw_message`.
2021-02-02 13:44:01 +01:00
Rusty Russell 3c5502426b lightningd: addgossip API to inject gossip messages.
Importantly, this is synchronous, so pay will be able to use it
reliably.

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
2021-02-02 13:44:01 +01:00
Michael Schmoock bc40287ade pytest: peer_connected chainable tests 2021-02-01 09:57:15 +10:30
Michael Schmoock bdf0d60fd6 chore: fix typo in openchannel hook log
Nit: The underscore in "openchannel_hook" is wrong, bcause the name of
the hook is just "openchannel". The "_hook" implied this to be part of
the name.

Changelog-None
2021-02-01 09:57:15 +10:30
Michael Schmoock 3a0b1c5b1d pytest: improve test_openchannel_hook_chaining
The current test was not checking for the output of the first plugin in
the chain.
2021-01-29 13:37:42 +10:30
Michael Schmoock 67f2939540 pytest: custommsg chainable tests 2021-01-29 13:37:42 +10:30
Michael Schmoock 8e71c7a1f1 feat: make custommsg hook chainable
Changelog-Changed: custommsg hook is now chainable
2021-01-29 13:37:42 +10:30
Christian Decker 2834aaced0 gci: Stabilize test_forward_event_notification 2021-01-29 10:29:09 +10:30
Christian Decker 62cb1c3fbc pytest: Stabilize test_forward_stats 2021-01-29 10:29:09 +10:30
Christian Decker da2e956538 pytest: Stabilize test_routing_gossip
openchannel internally generates blocks, which may cause nodes to be
out of sync and ignore "future" channel announcements, resulting in
bad gossip.
2021-01-29 10:29:09 +10:30
Christian Decker 5ecaff65ee pytest: Give each run of the hsmtool its own pty 2021-01-29 10:29:09 +10:30
Christian Decker 1463797c61 pytest: Stabilize test_funding_close_upfront
Reconnections and unsynchronized states where causing us some issues.
2021-01-29 10:29:09 +10:30
Christian Decker bbdf35c6fe pytest: Stabilize test_closing_negotiation_reconnect
The test was not considering that concurrent sendrawtx of the same tx
is not stable, and either endpoint will submit it first. Now just
checking state transitions and the mempool.
2021-01-29 10:29:09 +10:30
Christian Decker 52e82b76b6 pytest: Stabilize test_bad_onion 2021-01-29 10:29:09 +10:30
Christian Decker 3d4c111721 pytest: Stabilize test_multiple_channels
If we're quick (or the node is slow) we end up reconnecting before our
counterparty has realized the state transition, resulting in an
unexpected re-establish.
2021-01-29 10:29:09 +10:30
Christian Decker 542f3225e3 pytest: Parameterize process waits for hsmtool calls
We were sometimes waiting only 5 seconds, which is way too short on a
heavily loaded machine such as CI. Making it 30 seconds and collecting
it in a single place so we can adjust more easily.
2021-01-29 10:29:09 +10:30
Christian Decker 8cc62d76e4 pytest: Stabilize test_channel_{spendable,receivable}
They were using TIMEOUT / 2 which may be way too long (hit against
test timeout), so we use a still ludicrous 30 seconds instead.
2021-01-29 10:29:09 +10:30
Christian Decker 6384cadd69 pytest: Stabilize the negotiation tests
We also make the logic a bit nicer to read. The failure was due to
more than one status message being present if we look at the wrong
time:

```
arr = ['CLOSINGD_SIGEXCHANGE:We agreed on a closing fee of 20334 satoshi for tx:17f1e9d377840edf79d8b6f1ed0faba59bb307463461...9b98', 'CLOSINGD_SIGEXCHANGE:Waiting for another closing fee offer: ours was 20334 satoshi, theirs was 20332 satoshi,']                                                                                                   │

     def only_one(arr):
         """Many JSON RPC calls return an array; often we only expect a single entry
         """
 >       assert len(arr) == 1
 E       AssertionError
```
2021-01-29 10:29:09 +10:30
Christian Decker 04ed93f5f8 pytest: Stabilize test_funding_external_wallet_corners 2021-01-29 10:29:09 +10:30
Christian Decker 18483ca582 pytest: Disable test_v2_open if not developer
It requires `--dev-force-features` which isn't available without
`DEVELOPER=1`
2021-01-29 10:29:09 +10:30
Christian Decker 7962db821c pytest: Stabilize test_channel_state_changed_bilateral 2021-01-29 10:29:09 +10:30
Christian Decker 07f5054700 pytest: Stabilize test_setchannelfee_state
Synching with the blockchain was slower than our timeout...
2021-01-29 10:29:09 +10:30
Christian Decker 03449e3cf0 pytest: Stabilize test_gossip_persistence
We weren't waiting for the `dev_fail` transaction to hit the mempool,
throwing the results off.
2021-01-29 10:29:09 +10:30
Christian Decker ae40c10bcb pytest: Stabilize test_onchain_timeout
The timeout on the pay future was too short under valgrind.
2021-01-29 10:29:09 +10:30
Christian Decker c564f165fa pytest: Stabilize test_penalty_htlc_tx_timeout
We weren't waiting for the transactions to enter the mempool which
could cause all of our fine-tuned block counts to be off. Now just
waiting for the expected number of txs.
2021-01-29 10:29:09 +10:30
Christian Decker c0f06f2779 pytest: Simplify and stabilize test_reconnect_no_update 2021-01-29 10:29:09 +10:30
Christian Decker 2b12cac31e pytest: Skip hsm encryption test if we don't have a TTY 2021-01-29 10:29:09 +10:30
Rusty Russell 001b5d6416 offers: make it a runtime option.
The fetchinvoice and offers plugins disable themselves if the option
isn't enabled (it's enabled by default on EXPERIMENTAL_FEATURES).

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Changelog-Added: `experimental-offers` enables fetch, payment and creation of (early draft) offers.
2021-01-13 14:45:36 +01:00
Rusty Russell 27c006f7aa libplugin: make init return a string.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Changelog-Added: libplugin: init can return a non-NULL string to disable the plugin.
2021-01-13 14:45:36 +01:00
Rusty Russell 529ae0d766 plugins: allow plugins to disable themselves at startup.
By returning 'disable: <reason>' inside getmanifest or init result.

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Changelog-Added: plugins: plugins can now disable themselves by returning `disable`, even if marked important.
2021-01-13 14:45:36 +01:00
Rusty Russell fc3e679c97 lightningd: control onion messages by `experimental-onion-messages` option.
Note that this also changes so the feature is not represented in channels,
reflecting the recent drafts.

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Changelog-Added: `experimental-onion-messages` enables send, receive and relay of onion messages.
2021-01-13 14:45:36 +01:00
niftynei 7e324177de df-tests: simultaneous openchannel_init (while in progress)
Reject a peer's request to open a channel while we're already in progress
2021-01-10 13:44:04 +01:00
niftynei 2c9ce2566c df-tests: test_lightningd_still_loading 2021-01-10 13:44:04 +01:00
niftynei a357e75c5d df-tests: test_disconnect_half_signed + test_reconnect_signed 2021-01-10 13:44:04 +01:00
niftynei fec1d9b707 df-tests: test_reconnect_openingd 2021-01-10 13:44:04 +01:00
niftynei 7eb6213fca df-tests: test_reconnect_sender_add1, others
We're still ignoring multifundchannel failure tests, but now we use the
'dual-fund' flag instead of a blanket true.
2021-01-10 13:44:04 +01:00
niftynei d9c68a708a df-tests: test_reconnect_receiver_add 2021-01-10 13:44:04 +01:00
niftynei 96c42262ab df-tests: test_peerinfo 2021-01-10 13:44:04 +01:00
niftynei 135cd178f7 df-tests: test_channel_persistence 2021-01-10 13:44:04 +01:00
niftynei a82bfa83ff df-tests: have the df_accepter plugin keep track of attempts
Test connection/reconnection handling for v2 opens. We needed to fixup
the accepter plugin so that we were freeing up inputs on
disconnect/failure.
2021-01-10 13:44:04 +01:00
niftynei 88b4b5b2bb df-tests: test_disconnect_opener 2021-01-10 13:44:04 +01:00
niftynei 3e330d7186 df-tests: test_forward_local_failed_stats 2021-01-10 13:44:04 +01:00
niftynei 973224dd14 df-tests: if we're using fundbalancechannel, set up plugin
nodes need to be using the accepter plugin, so they'll match funding.
we used to use push_msat for this, but v2 gets rid of it.

*sad trombone*
2021-01-10 13:44:04 +01:00
niftynei f822a93cf4 df-tests: add `setacceptmaxfunding` to df_accepter plugin
Method for setting a ceiling on the fund matching capabilities of
df_accepter. Setting it to zero means we don't fund the channel anymore.
2021-01-10 13:44:04 +01:00
niftynei e964c31369 df-tests: test_payment_success|failed_persistence 2021-01-10 13:44:04 +01:00
niftynei 09da171381 df-tests: test_io_logging
Move pid collection down to when dualopend is definitely dead
2021-01-10 13:44:04 +01:00
niftynei 609d509ca5 df-tests: test_channel_state_changed_unilateral 2021-01-10 13:44:04 +01:00
niftynei ac88c72f1e df-tests: test_channel_state_changed_bilateral 2021-01-10 13:44:04 +01:00