Refactor ShutdownResult type and construction #2613

wvanlint · 2023-09-28T04:31:29Z

As a follow-up to #2486, this PR:

Refactors the ShutdownResult type and moves construction to Channel methods.
Cleans up typos and unused variables/imports.
Refactors check_closed_event to be able to share it for batch funding tests. The current implementation is not usable since there might be multiple ClosureReasons. Let me know if this is not worth it though.

codecov-commenter · 2023-09-28T04:43:09Z

Codecov Report

Attention: 11 lines in your changes are missing coverage. Please review.

Comparison is base (1852715) 88.92% compared to head (316a794) 88.98%.

❗ Your organization needs to install the Codecov GitHub app to enable full functionality.

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #2613      +/-   ##
==========================================
+ Coverage   88.92%   88.98%   +0.05%     
==========================================
  Files         112      112              
  Lines       87632    87650      +18     
  Branches    87632    87650      +18     
==========================================
+ Hits        77929    77992      +63     
+ Misses       7473     7422      -51     
- Partials     2230     2236       +6

Files	Coverage Δ
lightning/src/ln/functional_tests.rs	`97.25% <100.00%> (+0.01%)`	⬆️
lightning/src/ln/functional_test_utils.rs	`91.10% <97.36%> (+0.25%)`	⬆️
lightning/src/ln/channelmanager.rs	`81.28% <84.00%> (-0.03%)`	⬇️
lightning/src/ln/channel.rs	`88.38% <77.77%> (-0.01%)`	⬇️

... and 8 files with indirect coverage changes

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

wpaulino · 2023-09-28T17:28:40Z

Any reason this is marked as draft?

wvanlint · 2023-09-28T18:03:50Z

lightning/src/ln/channel.rs

@@ -5557,7 +5573,7 @@ impl<SP: Deref> Channel<SP> where
 	/// [`ChannelMonitorUpdate`] will be returned).
 	pub fn get_shutdown(&mut self, signer_provider: &SP, their_features: &InitFeatures,
 		target_feerate_sats_per_kw: Option<u32>, override_shutdown_script: Option<ShutdownScript>)
-	-> Result<(msgs::Shutdown, Option<ChannelMonitorUpdate>, Vec<(HTLCSource, PaymentHash)>), APIError>
+	-> Result<(msgs::Shutdown, Option<ChannelMonitorUpdate>, Vec<(HTLCSource, PaymentHash)>, Option<ShutdownResult>), APIError>


I wonder if these Option<ChannelMonitorUpdate>, Vec<(HTLCSource, PaymentHash)> can be collapsed into Option<ShutdownResult> although the use cases are slightly different.

Maybe? I think I'd rather keep them separate though. We'd want to handle the monitor update result on the non-shutdown case.

Right, that was my main hesitation. It seems possible to branch on the update_id of the ChannelMonitorUpdate being equal to CLOSED_CHANNEL_UPDATE_ID, but might not be something we want to unify.

lightning/src/ln/channelmanager.rs

wvanlint · 2023-09-28T18:09:47Z

Any reason this is marked as draft?

Took another look, ready for review now. :)

lightning/src/ln/channel.rs

vladimirfomene

Just a couple of comments here and there

vladimirfomene · 2023-10-03T17:39:34Z

lightning/src/ln/channelmanager.rs

@@ -2700,7 +2700,7 @@ where
 		}

 		let (monitor_update_option, mut failed_htlcs, unbroadcasted_batch_funding_txid) = shutdown_res;
-		log_debug!(self.logger, "Finishing force-closure of channel with {} HTLCs to fail", failed_htlcs.len());
+		log_debug!(self.logger, "Finishing closure of channel with {} HTLCs to fail", failed_htlcs.len());


Does force here not give more context to anyone reading the logs?

Problem is its not longer just force-closures, its all closures.

vladimirfomene · 2023-10-03T18:08:36Z

lightning/src/ln/functional_test_utils.rs

+	pub reason: Option<ClosureReason>,
+}
+
+impl Default for ExpectedCloseEvent {


Does it not make sense to derive Default instead of manually implementing it?

Definitely makes sense! Derived instead. :)

vladimirfomene · 2023-10-03T19:32:30Z

lightning/src/ln/channel.rs

@@ -5626,11 +5642,19 @@ impl<SP: Deref> Channel<SP> where
 		};

 		// From here on out, we may not fail!
+		let shutdown_result;
+		let unbroadcasted_batch_funding_txid = self.context.unbroadcasted_batch_funding_txid();
 		self.context.target_closing_feerate_sats_per_kw = target_feerate_sats_per_kw;
 		if self.context.channel_state & !STATE_FLAGS < ChannelState::FundingSent as u32 {


Instead of defining shutdown_result on L5645 won't it be better to have this if/else expression return a ShutdownResult and then capture it in the shutdown_result variable.

Done, capturing shutdown_result from if-else expressions in a couple places now.

TheBlueMatt · 2023-10-05T22:56:45Z

Can you write more full commit messages that describe what we're doing (not just "refactor", but "refactor to do X") and why we're doing it?

wvanlint · 2023-10-11T00:39:09Z

Can you write more full commit messages that describe what we're doing (not just "refactor", but "refactor to do X") and why we're doing it?

I've rewritten the commit subjects and added explanations to the bodies following https://cbea.ms/git-commit/, let me know if anything can be reworded there though!

TheBlueMatt

LGTM, but I'd rather expand a few assertions and checks cause this is all pretty delicate code.

TheBlueMatt · 2023-10-15T17:58:28Z

lightning/src/ln/functional_test_utils.rs

+				channel_capacity_sats,
+				..
+			} if (
+				expected_event.channel_id.map(|expected| *channel_id == expected).unwrap_or(true) &&


Shouldn't these all be unwrap_or(false)? We don't want to match if the caller specified a channel id to match but the event didn't have it.

The .map function is called on the expected event field, so the unwrap_or(true) call is meant to match if the caller left expectations unspecified. Does that sound correct? A ChannelClosed event will always have a channel_id.

TheBlueMatt · 2023-10-15T18:03:21Z

lightning/src/ln/functional_test_utils.rs

+pub fn check_closed_event(node: &Node, _events_count: usize, expected_reason: ClosureReason, is_check_discard_funding: bool,
 	expected_counterparty_node_ids: &[PublicKey], expected_channel_capacity: u64) {


?

Suggested change

pub fn check_closed_event(node: &Node, _events_count: usize, expected_reason: ClosureReason, is_check_discard_funding: bool,

expected_counterparty_node_ids: &[PublicKey], expected_channel_capacity: u64) {

pub fn check_closed_event(node: &Node, events_count: usize, expected_reason: ClosureReason, is_check_discard_funding: bool,

expected_counterparty_node_ids: &[PublicKey], expected_channel_capacity: u64) {

assert_eq!(events_count, expected_counterparty_node_ids.len());

Needed to account for the DiscardFunding events too but verified the events_count here.

TheBlueMatt · 2023-10-15T18:07:24Z

lightning/src/ln/channel.rs

-	Option<Txid>
-);
+/// The result of a shutdown that should be handled.
+pub(crate) struct ShutdownResult {


Suggested change

pub(crate) struct ShutdownResult {

#[must_use]

pub(crate) struct ShutdownResult {

TheBlueMatt · 2023-10-15T18:10:27Z

lightning/src/ln/channelmanager.rs

 							chan.get_shutdown(&self.signer_provider, their_features, target_feerate_sats_per_1000_weight, override_shutdown_script)?;
 						failed_htlcs = htlcs;
+						shutdown_result = local_shutdown_result;


Let's debug_assert our assumptions.

Suggested change

shutdown_result = local_shutdown_result;

shutdown_result = local_shutdown_result;

debug_assert_eq!(shutdown_result.is_some(), chan.is_shutdown());

Added this assertion to all the cooperative closing paths.

TheBlueMatt · 2023-10-18T22:00:00Z

Ugh, sorry, this needs rebase due to a silent conflict with #2658.

The check_closed_event function verified closure events against multiple counterparty nodes, but only a single closure reason and channel capacity. This commit introduces a check_closed_events function to verify events against descriptions of each expected event, and refactors check_closed_event in function of check_closed_events.

This refactors ShutdownResult as follows: - Makes ShutdownResult a struct instead of a tuple to represent individual results that need to be handled. This recently also includes funding batch closure propagations. - Makes Channel solely responsible for constructing ShutdownResult as it should own all channel-specific logic.

wvanlint · 2023-10-19T16:40:55Z

Ugh, sorry, this needs rebase due to a silent conflict with #2658.

No problem! Rebased. :)

TheBlueMatt

Thanks!

wvanlint mentioned this pull request Sep 28, 2023

Batch funding #2486

Merged

wvanlint commented Sep 28, 2023

View reviewed changes

wvanlint marked this pull request as ready for review September 28, 2023 18:09

benthecarman reviewed Sep 28, 2023

View reviewed changes

lightning/src/ln/channel.rs Show resolved Hide resolved

wvanlint force-pushed the batch_funding_fix_up branch from 9c2ee48 to 6960cb8 Compare October 2, 2023 17:53

vladimirfomene reviewed Oct 3, 2023

View reviewed changes

wvanlint force-pushed the batch_funding_fix_up branch from 6960cb8 to 39e0b7d Compare October 11, 2023 00:22

wvanlint requested review from TheBlueMatt and wpaulino October 11, 2023 00:39

wpaulino previously approved these changes Oct 12, 2023

View reviewed changes

TheBlueMatt reviewed Oct 15, 2023

View reviewed changes

wvanlint dismissed wpaulino’s stale review via aa8a485 October 17, 2023 21:16

wvanlint force-pushed the batch_funding_fix_up branch from 39e0b7d to aa8a485 Compare October 17, 2023 21:16

wvanlint requested a review from TheBlueMatt October 17, 2023 23:40

wvanlint added 3 commits October 18, 2023 20:46

Clean up typos and unused variables/imports

46dab8f

wvanlint force-pushed the batch_funding_fix_up branch from aa8a485 to 316a794 Compare October 19, 2023 03:52

wpaulino approved these changes Oct 20, 2023

View reviewed changes

TheBlueMatt approved these changes Oct 29, 2023

View reviewed changes

TheBlueMatt merged commit 8f308f9 into lightningdevkit:main Oct 29, 2023

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Refactor ShutdownResult type and construction #2613

Refactor ShutdownResult type and construction #2613

wvanlint commented Sep 28, 2023 •

edited

Loading

codecov-commenter commented Sep 28, 2023 •

edited

Loading

wpaulino commented Sep 28, 2023

wvanlint Sep 28, 2023

wpaulino Sep 28, 2023

wvanlint Oct 11, 2023

wvanlint commented Sep 28, 2023

vladimirfomene left a comment

vladimirfomene Oct 3, 2023

TheBlueMatt Oct 5, 2023

vladimirfomene Oct 3, 2023

wvanlint Oct 11, 2023 •

edited

Loading

vladimirfomene Oct 3, 2023

wvanlint Oct 11, 2023

TheBlueMatt commented Oct 5, 2023

wvanlint commented Oct 11, 2023 •

edited

Loading

TheBlueMatt left a comment

TheBlueMatt Oct 15, 2023

wvanlint Oct 17, 2023

TheBlueMatt Oct 15, 2023

wvanlint Oct 17, 2023

TheBlueMatt Oct 15, 2023

wvanlint Oct 17, 2023

TheBlueMatt Oct 15, 2023

wvanlint Oct 17, 2023

TheBlueMatt commented Oct 18, 2023

wvanlint commented Oct 19, 2023

TheBlueMatt left a comment

		pub fn check_closed_event(node: &Node, _events_count: usize, expected_reason: ClosureReason, is_check_discard_funding: bool,
		expected_counterparty_node_ids: &[PublicKey], expected_channel_capacity: u64) {

	pub(crate) struct ShutdownResult {
	#[must_use]
	pub(crate) struct ShutdownResult {

	shutdown_result = local_shutdown_result;
	shutdown_result = local_shutdown_result;
	debug_assert_eq!(shutdown_result.is_some(), chan.is_shutdown());

Refactor ShutdownResult type and construction #2613

Refactor ShutdownResult type and construction #2613

Conversation

wvanlint commented Sep 28, 2023 • edited Loading

codecov-commenter commented Sep 28, 2023 • edited Loading

Codecov Report

wpaulino commented Sep 28, 2023

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

wvanlint commented Sep 28, 2023

vladimirfomene left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

wvanlint Oct 11, 2023 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

TheBlueMatt commented Oct 5, 2023

wvanlint commented Oct 11, 2023 • edited Loading

TheBlueMatt left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

TheBlueMatt commented Oct 18, 2023

wvanlint commented Oct 19, 2023

TheBlueMatt left a comment

Choose a reason for hiding this comment

wvanlint commented Sep 28, 2023 •

edited

Loading

codecov-commenter commented Sep 28, 2023 •

edited

Loading

wvanlint Oct 11, 2023 •

edited

Loading

wvanlint commented Oct 11, 2023 •

edited

Loading