Since we are the only writer, we don't need one.
Name (time in s) Min Max Mean StdDev Median
sqlite: test_spam_listcommands (before) 2.1193 2.4524 2.2343 0.1341 2.2229
sqlite: test_spam_listcommands (after) 2.0140 2.2349 2.1001 0.0893 2.0644
Postgres: test_spam_listcommands (before) 6.5572 6.8440 6.7067 0.1032 6.6967
Postgres: test_spam_listcommands (after) 4.4237 5.0024 4.6495 0.2278 4.6717
A nice 31% speedup!
Changelog-Changed: Postgres: significant speedup on read-only operations (e.g. 30% on empty SELECTs)
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
We always start a transaction before processing, but there are cases where
we don't need to. Switch to doing it on-demand.
This doesn't make a big difference for sqlite3, but it can for Postgres because
of the latency: 12% or so. Every bit helps!
30 runs, min-max(mean+/-stddev):
Postgres before: 8.842773-9.769030(9.19531+/-0.21)
Postgres after: 8.007967-8.321856(8.14172+/-0.066)
sqlite3 before: 7.486042-8.371831(8.15544+/-0.19)
sqlite3 after: 7.973411-8.576135(8.3025+/-0.12)
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Basically, `devtools/reduce-includes.sh */*.c`.
Build time from make clean (RUST=0) (includes building external libs):
Before:
real 0m38.944000-40.416000(40.1131+/-0.4)s
user 3m6.790000-17.159000(15.0571+/-2.8)s
sys 0m35.304000-37.336000(36.8942+/-0.57)s
After:
real 0m37.872000-39.974000(39.5466+/-0.59)s
user 3m1.211000-14.968000(12.4556+/-3.9)s
sys 0m35.008000-36.830000(36.4143+/-0.5)s
Build time after touch config.vars (RUST=0):
Before:
real 0m19.831000-21.862000(21.5528+/-0.58)s
user 2m15.361000-30.731000(28.4798+/-4.4)s
sys 0m21.056000-22.339000(22.0346+/-0.35)s
After:
real 0m18.384000-21.307000(20.8605+/-0.92)s
user 2m5.585000-26.843000(23.6017+/-6.7)s
sys 0m19.650000-22.003000(21.4943+/-0.69)s
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Each header should only include the other headers it needs to compile;
`devtools/reduce-includes.sh */*.h` does this. The C files then need
additional includes if they don't compile.
And remove the entirely useless wire/onion_wire.h, which only serves to include wire/onion_wiregen.h.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
And gracefully fail for this case.
There's no such thing for Postgres, but that's because dbs need to be
set up by the admin.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
The character classification functions in <ctype.h> are designed to
classify characters returned by <stdio.h> getchar() and friends, which
return characters as signed integers in the range 0 to 255 or EOF. The
behavior of the ctype functions is undefined if they are passed a value
outside of that range, which may happen if they are passed a char-typed
value and the system's char type is signed.
<ccan/str/str.h> defines some inline utility functions that perform the
necessary cast to coerce a char-typed argument into the allowed value
range. Call these wrappers instead of the bare ctype functions when
classifying char-typed characters.
Changelog-None
Now you can grep for 'sqlite3 version' and see where we would like
to update.
Debian 11 (Bullseye) and Ubuntu 20.04 (Focal) ship with SQLite 3.31.1.
RHEL 9 ships with 3.34.1. Fedora 38+ uses SQLite 3.40+.
Unfortunately, RHEL8 ships with 3.26.0, and is still on maintenance Support
(security fixes, no new features): runs until May 31, 2029.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
`bitcoin_tx_with_psbt` would somewhat opaquely steal the passed `psbt` value.
This caused a bug where code made a `bitcoin_tx` using a psbt without realizing the value was stolen. Because the resulting `bitcoin_tx` was placed in tmpctx it was not immediately clear that using `psbt` afterwards was an error until the tmpctx was cleared — creating a valgrind backtrace far from the actual issue.
Switching to the routine to using TAKES and adding documentation in the header, makes it explicitly clear which operation the user is doing — helping prevent future regressions of this kind.
Changelog-None
If we connected out, remember that address. We always remember the last
address, but that may be an incoming address. This is explicitly the last
outgoing address which worked.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
It's a u64, we should pass by copy. This is a big sweeping change,
but mainly mechanical (change one, compile, fix breakage, repeat).
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Put an assertion inside db.c, and run every command we do (in testing) through
a `check` variant.
I inserted a deliberate bug (made addpsbtoutput call wallet_get_newindex()
before returning when running `check`, and indeed, backtrace as expected.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Update the lightningd <-> channeld interface with lots of new commands to needed to facilitate spicing.
Implement the channeld splicing protocol leveraging the interactivetx protocol.
Implement lightningd’s channel_control to support channeld in its splicing efforts.
Changelog-Added: Added the features to enable splicing & resizing of active channels.
This is almost always true already; fix up the few non-standard ones.
This is enforced with an assert, and I ran the entire test suite to
double-check.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
This avoids the mess where we override db_fatal for teqsts, and keeps it
generic.
Also allows us to get rid of one #if DEVELOPER, and an ugly global for
bookkeeper.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
It is possible for db_column_bytes() to return 0 and for
db_column_blob() to return NULL even when db_column_is_null() returns
false. We need to short circuit in this case.
Detected by UBSan:
db/bindings.c:479:12: runtime error: null pointer passed as argument 2, which is declared to never be null
/usr/include/string.h:44:28: note: nonnull attribute specified here
#0 0x95f117 in db_col_arr_ db/bindings.c:479:2
#1 0x95ef85 in db_col_channel_type db/bindings.c:459:32
#2 0x852c03 in wallet_stmt2channel wallet/wallet.c:1483:9
#3 0x81f396 in wallet_channels_load_active wallet/wallet.c:1749:23
#4 0x81f03d in wallet_init_channels wallet/wallet.c:1765:9
#5 0x72f1f9 in load_channels_from_wallet lightningd/peer_control.c:2257:7
#6 0x672856 in main lightningd/lightningd.c:1121:25
This was always the intent, but now we have to reconstruct from the
disparate fields.
This means `option_anchor_outputs` is now redundant.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
I couldn't figure out why my new SQL query was returning 0 rows,
and it was because we were ignoring errors.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Avoids the following when postgres returns no query result:
==63458== Conditional jump or move depends on uninitialised value(s)
==63458== at 0x226A1F: db_postgres_step (db_postgres.c:156)
==63458== by 0x22535B: db_step (utils.c:155)
==63458== by 0x1E089A: db_data_version_get (exec.c:49)
==63458== by 0x194F6F: db_setup (db.c:1029)
==63458== by 0x199A2F: wallet_new (wallet.c:101)
==63458== by 0x154B70: main (lightningd.c:1035)
Changelog-None
Although it's deprecated already (because it stores as string), it's
better to make the name explicit. And create a new helper which stores as BIGINT.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Because it used internal routines, it didn't pass operations through the
db hook! So make it use the generic routines, with the twist that they
are not translated.
And when we use this in a migration hook, we're actually in a
transaction.
This, in turn, introduces an issue: we need to be outside a transaction
to "PRAGMA foreign_keys = OFF", but completing the transaction when
there is a db hook actually enters the io loop, freeing the tmpctx!
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
The initial snapshots on an already-running lightningd are expected to
be unbalanced, but this shouldn't cause users to long for the green,
green grass of home.
This controls the Art of Noise.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
