We save location where transaction was started, in case we try to nest.
There's now no error case; db_exec_mayfail() is the only one.
This means the tests need to override fatal() if they want to intercept
these errors.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
we should never be doing two startups at once, but why take chances? Plus,
we can then assert that all db calls are in transactions.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Otherwise we find ourselves outside a commitment. This is a bandaid
until we remove nested commitments again at the end of this series.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Nesting is provided by only actually performing the outermost
transaction and simulating the nested ones. This still allows us to
ensure on lower levels that we are in the context of a transaction
without having to resort to keeping explicitly track of it in the
calling code.
Signed-off-by: Christian Decker <decker.christian@gmail.com>
In addition we also set some of the test values to a pattern instead
of just `memset`ting it to 0, which may hide some crossed lines.
Signed-off-by: Christian Decker <decker.christian@gmail.com>
We use these quite often and it is cumbersome having to do these
simple conversions inline, so just expose pseudo-sqlite3 methods to
bind and extract from/to a stmt.
Signed-off-by: Christian Decker <decker.christian@gmail.com>
Technically it's the caller that'll own the statement, but it is nice
to have db_exec_prepared dispose of it.
Signed-off-by: Christian Decker <decker.christian@gmail.com>
This makes executing a query/command a two step process, but allows us
to use the native binding and avoid having to build queries as SQL
strings. Two major advantages are that we are no longer vulnerable to
SQL injections and that we do not have to hex-encode binary fields
like private keys, hashes, and routing onions, halving the storage
requirements for those.
Signed-off-by: Christian Decker <decker.christian@gmail.com>
This was causing me some trouble by making it look like the last query
failed, when it really was an old one. No need to drag failures around
for longer than needed.
Signed-off-by: Christian Decker <decker.christian@gmail.com>
Also added a small warning to one of the used enums not to reorder or
insert values. They'd break the update path.
Signed-off-by: Christian Decker <decker.christian@gmail.com>
Also, we split the more sophisticated json_add helpers to avoid pulling in
everything into lightning-cli, and unify the routines to print struct
short_channel_id (it's ':', not '/' too).
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
They happen to advance at the same pace but mixing them may have
unforeseen consequences, and I have done so a few times already so
this explicitly separates them.
Signed-off-by: Christian Decker <decker.christian@gmail.com>
And store in peer->last_tx/peer->last_sig like all other places,
that way we broadcast it if we need to.
Note: the removal of tmpctx in funder_channel() is needed because we
use txs[0], which was allocated off tmpctx.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Automatically exiting the DB transaction upon any failure is strange
since it'll kill any later attempt to commit. The commit itself should
be used to verify that everything was ok.
Not the nicest code, but it allows us to store the bip32_max_index so
that we don't forget our addresses upon restart. We could have done
the same by retrieving the max index from our index, but then we'd
forget addresses that don't have an associated output. Conversion
to/from string is so that we can store arbitrary one off values in the
DB in the future, independent of type.
Rusty Russell
Christian Decker