This flake has been difficult to reproduce, so let's dump the gossip
store to aid in debugging. See issue #5410.

Changelog-None

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>

Here's the before-vs-after comparison (ignoring whitespace changes):

```diff
 --- /tmp/before	2022-07-20 21:52:44.336641810 +0930
 +++ /tmp/after	2022-07-20 21:55:54.355487769 +0930
 @@ -1,7 +1,7 @@
 -LIGHTNING-CLI(1)                                          lightning-cli                                         LIGHTNING-CLI(1)
 +LIGHTNING-CLI(1)                                                                                                LIGHTNING-CLI(1)
  
  NAME
 -       lightning-cli - Control lightning daemon
 +       lightning-cli -- Control lightning daemon
  
  SYNOPSIS
         lightning-cli [OPTIONS] command
 @@ -14,10 +14,7 @@
  
          --conf=PATH Sets configuration file (default: lightning-dir/config ).
  
 -        --network=network
 -        --mainnet
 -        --testnet
 -        --signet Sets network explicitly.
 +       --network=network --mainnet --testnet --signet Sets network explicitly.
  
          --rpc-file=FILE Named pipe to use to talk to lightning daemon: default is lightning-rpc in the lightning directory.
  
 @@ -43,8 +40,8 @@
  
          --version/-V Print version number to standard output and exit.
  
 -        allow-deprecated-apis=BOOL  Enable  deprecated options. It defaults to true, but you should set it to false when testing
 -       to ensure that an upgrade won't break your configuration.
 +       allow-deprecated-apis=BOOL Enable deprecated options. It defaults to true, but you should set it to false when testing to
 +       ensure that an upgrade won't break your configuration.
  
  COMMANDS
         lightning-cli simply uses the JSON RPC interface to talk to lightningd, and prints the results. Thus the commands  avail‐
 @@ -67,7 +64,7 @@
  
         lightning-cli help
  
 -              1.     Fund a 10k sat channel using uncomfirmed outputs
 +       2.     Fund a 10k sat channel using uncomfirmed outputs
  
         lightning-cli --keywords fundchannel id=028f...ae7d amount=10000sat minconf=0
  
 @@ -84,4 +81,4 @@
         Note: the modules in the ccan/ directory have their own licenses, but the rest of the code is covered  by  the  BSD-style
         MIT license.
  
 -                                                                                                                LIGHTNING-CLI(1)
 +Core Lightning v0.11.0.1-350-gac2e137                                                                           LIGHTNING-CLI(1)
```

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Fixes: #5437

There are no definition lists in Markdown, and lists get mangled if
they follow immediately: they need a line between them.

So use bullets for options, and use an indent so the text gets in the
line below.

Here's a before-and-after example:

```diff
 --- /tmp/after	2022-07-20 21:55:54.355487769 +0930
 +++ /tmp/after2	2022-07-20 21:58:17.305642576 +0930
 @@ -10,38 +10,71 @@
         lightning-cli sends commands to the lightning daemon.

  OPTIONS
 -       --lightning-dir=DIR Set the directory for the lightning daemon we're talking to; defaults to $HOME/.lightning.
 +       • --lightning-dir=DIR

 -       --conf=PATH Sets configuration file (default: lightning-dir/config ).
 +         Set the directory for the lightning daemon we're talking to; defaults to $HOME/.lightning.

 -       --network=network --mainnet --testnet --signet Sets network explicitly.
 +       • --conf=PATH

 -       --rpc-file=FILE Named pipe to use to talk to lightning daemon: default is lightning-rpc in the lightning directory.
 +         Sets configuration file (default: lightning-dir/config ).

 -       --keywords/-k Use format key=value for parameters in any order
 +       • --network=network

 -       --order/-o Follow strictly the order of parameters for the command
 +       • --mainnet

 -       --json/-J Return result in JSON format (default unless help command, or result contains a format-hint field).
 +       • --testnet

 -       --raw/-R Return raw JSON directly as lightningd replies; this can be faster for large requests.
 +       • --signet

 -       --human-readable/-H Return result in human-readable output.
 +         Sets network explicitly.

 -       --flat/-F  Return  JSON result in flattened one-per-line output, e.g. { "help": [ { "command": "check" } ] } would become
 +       • --rpc-file=FILE
 +
 +         Named pipe to use to talk to lightning daemon: default is lightning-rpc in the lightning directory.
 +
 +       • --keywords/-k
 +
 +         Use format key=value for parameters in any order
 +
 +       • --order/-o
 +
 +         Follow strictly the order of parameters for the command
 +
 +       • --json/-J
 +
 +         Return result in JSON format (default unless help command, or result contains a format-hint field).
 +
 +       • --raw/-R
 +
 +         Return raw JSON directly as lightningd replies; this can be faster for large requests.
 +
 +       • --human-readable/-H
 +
 +         Return result in human-readable output.
 +
 +       • --flat/-F
 +
 +         Return  JSON  result  in  flattened  one-per-line  output,  e.g.  {  "help":  [ { "command": "check" } ] } would become
         help[0].command=check.  This is useful for simple scripts which want to find a  specific  output  field  without  parsing
         JSON.

 -       --notifications/-N=LEVEL  If  LEVEL is 'none', then never print out notifications.  Otherwise, print out notifications of
 -       LEVEL or above (one of io, debug, info (the default), unusual or broken: they are prefixed with # .
 +       • --notifications/-N=LEVEL
 +
 +         If  LEVEL  is 'none', then never print out notifications.  Otherwise, print out notifications of LEVEL or above (one of
 +         io, debug, info (the default), unusual or broken: they are prefixed with # .
 +
 +       • --help/-h
 +
 +         Pretty-print summary of options to standard output and exit.  The format can be changed using -F, -R, -J, -H etc.
 +
 +       • --version/-V

 -       --help/-h Pretty-print summary of options to standard output and exit.  The format can be changed using -F,  -R,  -J,  -H
 -       etc.
 +         Print version number to standard output and exit.

 -       --version/-V Print version number to standard output and exit.
 +       • allow-deprecated-apis=BOOL

 -       allow-deprecated-apis=BOOL Enable deprecated options. It defaults to true, but you should set it to false when testing to
 -       ensure that an upgrade won't break your configuration.
 +         Enable deprecated options. It defaults to true, but you should set it to false when testing to ensure that  an  upgrade
 +         won't break your configuration.

  COMMANDS
         lightning-cli simply uses the JSON RPC interface to talk to lightningd, and prints the results. Thus the commands  avail‐
 @@ -60,13 +93,13 @@
         this is not encouraged.

  EXAMPLES
 -       1.     List commands
 +       1.     List commands:

 -       lightning-cli help
 +              • lightning-cli help

 -       2.     Fund a 10k sat channel using uncomfirmed outputs
 +       2.     Fund a 10k sat channel using uncomfirmed outputs:

 -       lightning-cli --keywords fundchannel id=028f...ae7d amount=10000sat minconf=0
 +              • lightning-cli --keywords fundchannel id=028f...ae7d amount=10000sat minconf=0

  BUGS
         This manpage documents how it should work, not how it does work. The pretty printing of results isn't pretty.
```

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>

You can't start a list without a paragraph separator.

```diff
 --- /tmp/before	2022-07-20 22:02:23.485372596 +0930
 +++ /tmp/after	2022-07-20 22:02:33.745528456 +0930
 @@ -21,12 +21,16 @@
         On startup of the daemon, no autoclean is set up.
  
  RETURN VALUE
 -       On success, an object is returned,  containing:  -  enabled  (boolean):
 -       whether invoice autocleaning is active
 +       On success, an object is returned, containing:
  
 -       If enabled is true: - expired_by (u64): how long an invoice must be ex‐
 -       pired (seconds) before we delete it - cycle_seconds (u64): how long  an
 -       invoice must be expired (seconds) before we delete it
 +       • enabled (boolean): whether invoice autocleaning is active
 +
 +       If enabled is true:
 +
 +       • expired_by (u64): how long an invoice must be expired  (seconds)  be‐
 +         fore we delete it
 +       • cycle_seconds  (u64):  how  long an invoice must be expired (seconds)
 +         before we delete it
  
  AUTHOR
         ZmnSCPxj <ZmnSCPxj@protonmail.com> is mainly responsible.
```

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>

….py in SHASUMS)

1. If the tool changes, you need to regenerate since the output may
   change.

2. This didn't just filter that out, ignored all but the first
   dependency, which made bisecting the bookkeeper plugin a nightmare:
   it didn't regenerate the .po file, causing random crashes.

If we want this, try $(filter-out tools/fromschema.py) instead.  But I
don't think we want that.

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>

If there's only a single underscore, lowdown ignores it, but if there are multiple
(see min_final_cltv_expiry) it decides we're trying to highlight part of the word.

Reported-by: @wtogami
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>

Should help diagnose #5572
which hit the invalid csum on a >64MB entry, if it happens again.

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>

Requires us to update to latest lnproto which is now using the most up
to date python-bitcoinlib, as well as updating our python lock files
(which pin the grpcio deps, because of locking problems h/t @cdecker)

Few extremely minor updates to the ubuntu dockerfile and ci builds

Reproduce crash for #5557!

If we record the channel open because bookkeeper was added after the
channel open request started but the channel confirms later, we end up
with re-recording any associated push or leased fees (paid or rcvd).

In the case where you've paid for these fees, your channel balance goes
negative and the node crashes the next time you call `listbalances`.

Reported-by: @chrisguida

We were double counting channel lease fees because we were double firing
the channel open event sequence (so to speak). If we don't report
balances for unopened channels, we don't have this problem?

Changelog-Changed: Plugins: `balance_snapshot` notification does not send balances for channels that aren't locked-in/opened yet

Clean up for #5557.

If you've got duplicate 'lease_fee' entries, we delete them!

Make sure that we're not issuing duplicate lease_fee events!

test_bookkeeping_closing_trimmed_htlcs fails to find 'all outputs
resolved' occassionally, seems like it's because the
OUR_DELAYED_TO_WALLET doesn't make it into the mempool before we start
mining blocks?

So here make sure there's something in the mempool before before we
start making new blocks.

The gossip_store version byte was unaccounted for in the initial traversal
of gossip_store_end. This lead to an offset and a bogus message length
field. As a result, an early portion of the gossip_store could have been
skipped, potentially leading to gossip propagation issues downstream.

Fixes #5572 #5565

Changelog-fixed: proper gossip_store operation may resolve some previous gossip propagation issues

Apparently we had two private channel announcements (the !private assert
failed).  While this shouldn't happen, don't crash because of it.

Fixes: #5578
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Changelog-Fixed: Plugins: topology plugin could crash when it sees duplicate private channel announcements.

1) dualopen has fd to connectd
2) channeld needs to take over
3) dualopen passes fd that leads to a connectd over for channeld to use
4) lightningd must receive the fd transfer request and process
5) dualopen shuts down and closes everything it owns

4 & 5 end up in a race. If 5 happens before 4, channeld ends up with an invalid fd for connectd — leaving it in a position to not receive messages.

Lingering for a second makes 4 win the race. Since the daemon is closing anyway, waiting for a second should be alright.

Changelog-Fixed: Fixed a condition for newly created channels that could trigger a need for reconnect.

Currently discovered IPs are only announced when we don't have any
usable addresses detected or configured already. However, the cli
command `getinfo` still showed theses unannounced addr as if they
were announced.

Changelog-Fixed: peer_control: getinfo showing unannounced addresses.

This one got badly messed up over time. I know we usually don't fix
these to have easier git-bisect. I can remove this commit if required.

Changelog-Fixed: peer_control: getinfo shows the correct port on discovered IPs

We added a submod dep (lowdown) that requires config to run first.

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>

This is usually fine, but without this, commando (another branch!) has
a race:

1. A command has multiple parts.
2. We start sending them out.
3. We get a response, which completes the cmd.
4. We go to send the next one out, but it's been freed.

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>

This can happen, and in fact does below in our test_autoclean_once
test where we update the datastore, and return from the cmd.

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>