Normal constraints are clamps on min/max caused by failed payments:
min for the channels that succeeded, max for the channel which failed.
Impressions are the results of successful payments, which alter both
min and max (negatively in the forward direction, positively in the
reverse).
impression: n
1. An effect, feeling, or image retained as a consequence of experience.
2. A vague notion, remembrance, or belief.
3. A mark produced on a surface by pressure.
Unlike constraints, this is the result of our own effect on the network: they're related
but different enough to get their own API and terminology.
The name conveys both we made an impression on the channel, and that
the results are a bit vague (due to other changes since then, which we
won't know about).
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Changelog-Added: JSON-RPC: `askrene` layers now contain "impressions" representing the effects of successful payments we made through channels.
Pass the computed VERSION as a build-arg, mirroring the docker-release workflow. Without it the container's VERSION is set-but-empty, `VERSION ?=` in the Makefile won't overwrite it, the git-describe fallback never runs, and `make install-program` fails with "git is required for generating version information".
Also add --no-push, so a release captain can do a full test build without publishing elementsproject/lightningd:$VERSION and :latest (the target always runs `docker buildx build --push`, since multi-arch images cannot be --load'ed into the classic docker store). The build only populates the builder's cache; a subsequent run without the flag pushes from cache quickly.
Co-Authored-By: Claude <claude@users.noreply.github.com>
The sign target detection uses pattern matching that requires spaces on
both sides of "sign" in the TARGETS variable. When only "sign" is passed
as an argument, TARGETS becomes " sign" (leading space only), causing
the condition to fail.
Add an additional pattern match to detect "sign" with a leading space
but no trailing space, which occurs when "sign" is the only target or
the last target in the list.
This regression seems to have been exposed by GitHub Actions runner
updates (Ubuntu 22.04 → 24.04), which changed argument passing behavior.
Changelog-None: Fixes the release signing issue in CI only.
These are useful for the common pattern of "append these bytes to this tal array".
As a bonus, we do memcheck() on all these callers, for extra checking
under valgrind.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
update-mocks searches for prototypes, looking first in the local
directory, then in */*.h, and takes the first. In the case of
plugins/bkpr/test/run-currencyrate_str.c referring to
jsonrpc_request_start_, this means it could get either the one in
libplugin (correct) or the one in lightningd (wrong!), depending on
directory order randomness.
Hack it a bit harder, to look one up from the local dir before */*.h.
This is redundant in most cases, but not for three-deep-nested dirs
like this.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
It would need to run under uv. I noticed it failing while debugging the 32-bit CI build,
but it fails everywhere.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
There's a new BOLT 12 test, which checks that the ->currencies array isn't empty.
We were treating it as missing, which is wrong. So allocate empty arrays when
they appear, instead of setting them to NULL.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
I ran `uv run make update-mocks` and many mocks vanished. But that
broke MacOS (and maybe other archs) which doesn't support function
sections and no longer linked.
If this proves problematic, we might end up generating mocks properly:
1. Use #include "mocks-sphinx-xor_cipher_stream_gen.c"
2. Generate those separate files as necessary.
3. Don't commit them, build them locally.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Enables STRICT tables in developer mode, but old databases (~2019) may
have BLOB values in TEXT columns. Migration converts BLOB faildetail
to TEXT with UTF-8 validation, NULLs invalid data.
STRICT is only applied to fresh databases; existing databases being
upgraded skip STRICT to avoid type affinity issues with legacy data.
Also adds security pragmas in developer mode: trusted_schema=OFF,
cell_size_check=ON.
Fixes#5390.
Changelog-Added: Database: STRICT tables and security pragmas in developer mode
Changelog-Fixed: Database migration for old BLOB-typed faildetail values
We also put this in the store_ended message, too: so you can
tell if the equivalent_offset there really refers to this new
entry (or if two or more rewrites have happened).
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
The Reckless search command was only returning a result if you
searched a perfect match, which is not too helpful. This updates the
command so that partial search matches return a result.
Before:
reckless search bolt
Search exhausted all sources
reckless search bol
Search exhausted all sources
reckless search bolt12-pris
Search exhausted all sources
After:
reckless search bolt
Plugins matching 'bolt':
bolt12-prism (https://github.com/lightningd/plugins)
reckless search bol
Plugins matching 'bol':
bolt12-prism (https://github.com/lightningd/plugins)
reckless search bolt12-pris
Plugins matching 'bolt12-pris':
bolt12-prism (https://github.com/lightningd/plugins)
Changelog-Fixed: reckless search now returns partial matches instead of requiring exact plugin names.
This means:
1. downgrade changes (we no longer fail due to node biases).
2. various deprecations no longer are
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
- Using environment variable `SOURCE_DATE_EPOCH` with fixed value will enforce a consistent timestamp for Fedora build.
. Similar to Ubuntu fix in commit 490fb0fc3b
- Locked cargo version
- Add `no-cache` to Fedora build
Changelog-Fixed: Core lightning builds for Fedora on all systems are deterministic.
In fact, you *must* use mnemonic to successfully recover a modern node!
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Changelog-Changed: JSON-RPC: `recover` takes a 12-word mnemonic for nodes created by v25.12 or later.
We cannot use the codex32 or raw hex for recovery of 25.12 nodes,
since they will then use the incorrect derivation for all paths, and
be unable to spend (or even find!) their funds.
So implement `getsecret` to replace `getcodexsecret`.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Changelog-Changed: `lightning-hsmtool`: `getsecret` replaces `getcodexsecret` for modern nodes (gives mnemonic).
Changelog-Deprecated: `lightning-hsmtool`: `getcodexsecret`. Use `getsecret`.
This commit introduces a modern coverage infrastructure for Core Lightning:
- Migrate from ad-hoc coverage script to integrated Makefile targets
- Add LLVM source-based coverage support with per-test profraw organization
- Integrate coverage collection into pytest framework via TailableProc
- Add GitHub Actions workflow for nightly coverage reports
- Add Taskfile.yml for convenient task automation
- Add codecov.yml for Codecov integration
- Add comprehensive coverage documentation in COVERAGE.md
- Update contributor workflow docs with new coverage script path
- Add coverage data files to .gitignore (*.profraw, *.profdata)
- Remove obsolete contrib/clang-coverage-report.sh
- Remove obsolete tests/conftest.py (now using pyln-testing markers)
- Update pyproject.toml to include pyln-testing in main dependencies
The new infrastructure automatically collects coverage data when CLN_COVERAGE_DIR
is set, organizing profraw files by test name for granular analysis.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
If they had a channel bias, and ran xpay, it will update the bias
to a v2 bias (with a timestamp). We must downgrade that, or the
older version won't load!
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
