fix: prevent channel data overwrite in FS→KV migration by ovitrif · Pull Request #74 · synonymdev/ldk-node

ovitrif · 2026-03-09T22:24:29Z

Summary

Fixed FS→KV channel data migration blindly overwriting newer state (monitors and channel manager).
The migration now skips writing a channel monitor when the KV store already holds one with a newer or equal update_id, and skips the channel manager when one already exists.
Migration read/deserialization failures now fail-closed (ReadFailed) to prevent silent data loss.

Test plan

cargo fmt --check — clean
cargo build — builds successfully
cargo test --lib — all unit tests pass (including new migration tests)
Bindings regenerated (Swift/Kotlin/Python)
xcframework checksum verified in Package.swift

Release

v0.7.0-rc.33

During FS→KV store migration, the code now compares update_id before writing and skips if the KV store already has a newer monitor. This prevents stale migration data from clobbering current channel state on repeated restarts. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

Replace incomplete testing docs with comprehensive coverage of unit tests, integration tests (with macOS ulimit note), and CLN/LND/VSS backend tests using correct RUSTFLAGS syntax. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

Address PR review: change channel monitor migration error handling from fail-open (proceed with overwrite) to fail-closed (abort with ReadFailed) for both deserialization errors and non-NotFound I/O errors. This prevents silent data loss when the KV store has valid data that can't be read due to transient errors or format changes. Also moves CHANGELOG entry to Synonym Fork Additions per project convention. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

The previous bindgen run used gobley-uniffi-bindgen v0.3.7 (from gobley main branch) instead of v0.2.0 (from fix-v0.2.0 branch), which added explicit `public` visibility modifiers and return types throughout the generated Kotlin code. Regenerated with the correct version. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

Extract the channel data migration block from build_with_store_internal into a standalone apply_channel_data_migration function, making it generic over the keys manager type to enable direct unit testing. Add 7 unit tests covering all code paths: - Invalid monitor data returns ReadFailed - Empty monitors list succeeds - Fresh write to empty store - Equal update_id (existing == migrated) still writes - Existing data with same update_id gets overwritten - Corrupt existing data triggers fail-closed (ReadFailed) - Store IO error triggers fail-closed (ReadFailed) Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

ben-kaufman · 2026-03-11T04:57:05Z

Looks good, but needs to add some tests ideally.

coreyphillips · 2026-03-11T14:50:09Z

One nit. Line 1666 uses > so when existing_update_id == migrated_update_id, we still write. Could use >= to skip the redundant write. Unless I'm missing something.

Use >= instead of > when comparing existing vs migrated update_id. When they are equal, the write is redundant — the store already has equivalent state. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

ovitrif · 2026-03-11T22:10:25Z

One nit. Line 1666 uses > so when existing_update_id == migrated_update_id, we still write. Could use >= to skip the redundant write. Unless I'm missing something.

Good catch — changed to >= so equal update_id skips the redundant write. Updated tests accordingly.

See cde6037

ovitrif · 2026-03-11T22:10:33Z

Looks good, but needs to add some tests ideally.

Added in 630b364 — 6 unit tests covering all migration code paths (invalid data, empty list, fresh write, skip on equal/newer update_id, corrupt existing data, store IO error). Run with cargo test --lib test_migration.

Add existence-check guard to channel manager migration (same fail-closed pattern as monitors). If the KV store already has a channel manager, the migration is skipped to prevent rolling back HTLC states, commitment indices, and pending payments with stale FS-sourced data. Also adds AI rule requiring unit tests for new business logic, and 3 unit tests covering the channel manager guard (fresh write, skip when existing, fail on read error). Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

…t-data test Correct the comment explaining why the channel manager guard uses existence-only checks: the real blocker is that ChannelManagerReadArgs requires already-deserialized channel monitors (loaded later in build), not that infrastructure is unavailable. Add test documenting intentional behavior when corrupt data exists in the store: migration is skipped (existence-only), preserving corrupt data. This is the correct trade-off — overwriting a potentially valid channel manager with stale FS data risks fund loss. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

# Conflicts: # CHANGELOG.md # Package.swift # bindings/kotlin/ldk-node-android/lib/src/main/jniLibs/arm64-v8a/libldk_node.so # bindings/kotlin/ldk-node-android/lib/src/main/jniLibs/armeabi-v7a/libldk_node.so # bindings/kotlin/ldk-node-android/lib/src/main/jniLibs/x86_64/libldk_node.so # bindings/kotlin/ldk-node-jvm/lib/src/main/resources/darwin-aarch64/libldk_node.dylib # bindings/kotlin/ldk-node-jvm/lib/src/main/resources/darwin-x86-64/libldk_node.dylib

ben-kaufman · 2026-03-12T02:06:27Z

@ovitrif new bindings and version bump missing?

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

ovitrif · 2026-03-12T10:08:00Z

@ovitrif new bindings and version bump missing?

Yes, was planned, now it's done 🫡

ovitrif · 2026-03-13T02:18:47Z

@ben-kaufman version bump skipped because there was no release of the previous PR (#75), where the artifacts got picked up by a CI job for the apps repos.

Thus there was no reason to bump the version.
Instead, I deleted the former rc.33 packages, release, and tag; rebuilt bindings and release, so now this encapsulates the changes of both PRs, on tip of this branch.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

ovitrif · 2026-03-09T22:24:29Z

Summary

Fixed FS→KV channel data migration blindly overwriting newer state (monitors and channel manager).
The migration now skips writing a channel monitor when the KV store already holds one with a newer or equal update_id, and skips the channel manager when one already exists.
Migration read/deserialization failures now fail-closed (ReadFailed) to prevent silent data loss.

Test plan

cargo fmt --check — clean
cargo build — builds successfully
cargo test --lib — all unit tests pass (including new migration tests)
Bindings regenerated (Swift/Kotlin/Python)
xcframework checksum verified in Package.swift

Release

v0.7.0-rc.33

During FS→KV store migration, the code now compares update_id before writing and skips if the KV store already has a newer monitor. This prevents stale migration data from clobbering current channel state on repeated restarts. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

Replace incomplete testing docs with comprehensive coverage of unit tests, integration tests (with macOS ulimit note), and CLN/LND/VSS backend tests using correct RUSTFLAGS syntax. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

Address PR review: change channel monitor migration error handling from fail-open (proceed with overwrite) to fail-closed (abort with ReadFailed) for both deserialization errors and non-NotFound I/O errors. This prevents silent data loss when the KV store has valid data that can't be read due to transient errors or format changes. Also moves CHANGELOG entry to Synonym Fork Additions per project convention. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

The previous bindgen run used gobley-uniffi-bindgen v0.3.7 (from gobley main branch) instead of v0.2.0 (from fix-v0.2.0 branch), which added explicit `public` visibility modifiers and return types throughout the generated Kotlin code. Regenerated with the correct version. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

Extract the channel data migration block from build_with_store_internal into a standalone apply_channel_data_migration function, making it generic over the keys manager type to enable direct unit testing. Add 7 unit tests covering all code paths: - Invalid monitor data returns ReadFailed - Empty monitors list succeeds - Fresh write to empty store - Equal update_id (existing == migrated) still writes - Existing data with same update_id gets overwritten - Corrupt existing data triggers fail-closed (ReadFailed) - Store IO error triggers fail-closed (ReadFailed) Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

ben-kaufman · 2026-03-11T04:57:05Z

Looks good, but needs to add some tests ideally.

coreyphillips · 2026-03-11T14:50:09Z

One nit. Line 1666 uses > so when existing_update_id == migrated_update_id, we still write. Could use >= to skip the redundant write. Unless I'm missing something.

Use >= instead of > when comparing existing vs migrated update_id. When they are equal, the write is redundant — the store already has equivalent state. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

ovitrif · 2026-03-11T22:10:25Z

One nit. Line 1666 uses > so when existing_update_id == migrated_update_id, we still write. Could use >= to skip the redundant write. Unless I'm missing something.

Good catch — changed to >= so equal update_id skips the redundant write. Updated tests accordingly.

See cde6037

ovitrif · 2026-03-11T22:10:33Z

Looks good, but needs to add some tests ideally.

Added in 630b364 — 6 unit tests covering all migration code paths (invalid data, empty list, fresh write, skip on equal/newer update_id, corrupt existing data, store IO error). Run with cargo test --lib test_migration.

Add existence-check guard to channel manager migration (same fail-closed pattern as monitors). If the KV store already has a channel manager, the migration is skipped to prevent rolling back HTLC states, commitment indices, and pending payments with stale FS-sourced data. Also adds AI rule requiring unit tests for new business logic, and 3 unit tests covering the channel manager guard (fresh write, skip when existing, fail on read error). Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

…t-data test Correct the comment explaining why the channel manager guard uses existence-only checks: the real blocker is that ChannelManagerReadArgs requires already-deserialized channel monitors (loaded later in build), not that infrastructure is unavailable. Add test documenting intentional behavior when corrupt data exists in the store: migration is skipped (existence-only), preserving corrupt data. This is the correct trade-off — overwriting a potentially valid channel manager with stale FS data risks fund loss. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

# Conflicts: # CHANGELOG.md # Package.swift # bindings/kotlin/ldk-node-android/lib/src/main/jniLibs/arm64-v8a/libldk_node.so # bindings/kotlin/ldk-node-android/lib/src/main/jniLibs/armeabi-v7a/libldk_node.so # bindings/kotlin/ldk-node-android/lib/src/main/jniLibs/x86_64/libldk_node.so # bindings/kotlin/ldk-node-jvm/lib/src/main/resources/darwin-aarch64/libldk_node.dylib # bindings/kotlin/ldk-node-jvm/lib/src/main/resources/darwin-x86-64/libldk_node.dylib

ben-kaufman · 2026-03-12T02:06:27Z

@ovitrif new bindings and version bump missing?

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

ovitrif · 2026-03-12T10:08:00Z

@ovitrif new bindings and version bump missing?

Yes, was planned, now it's done 🫡

ovitrif · 2026-03-13T02:18:47Z

@ben-kaufman version bump skipped because there was no release of the previous PR (#75), where the artifacts got picked up by a CI job for the apps repos.

Thus there was no reason to bump the version.
Instead, I deleted the former rc.33 packages, release, and tag; rebuilt bindings and release, so now this encapsulates the changes of both PRs, on tip of this branch.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

ovitrif · 2026-03-09T22:24:29Z

Summary

Fixed FS→KV channel data migration blindly overwriting newer state (monitors and channel manager).
The migration now skips writing a channel monitor when the KV store already holds one with a newer or equal update_id, and skips the channel manager when one already exists.
Migration read/deserialization failures now fail-closed (ReadFailed) to prevent silent data loss.

Test plan

cargo fmt --check — clean
cargo build — builds successfully
cargo test --lib — all unit tests pass (including new migration tests)
Bindings regenerated (Swift/Kotlin/Python)
xcframework checksum verified in Package.swift

Release

v0.7.0-rc.33

During FS→KV store migration, the code now compares update_id before writing and skips if the KV store already has a newer monitor. This prevents stale migration data from clobbering current channel state on repeated restarts. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

Replace incomplete testing docs with comprehensive coverage of unit tests, integration tests (with macOS ulimit note), and CLN/LND/VSS backend tests using correct RUSTFLAGS syntax. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

Address PR review: change channel monitor migration error handling from fail-open (proceed with overwrite) to fail-closed (abort with ReadFailed) for both deserialization errors and non-NotFound I/O errors. This prevents silent data loss when the KV store has valid data that can't be read due to transient errors or format changes. Also moves CHANGELOG entry to Synonym Fork Additions per project convention. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

The previous bindgen run used gobley-uniffi-bindgen v0.3.7 (from gobley main branch) instead of v0.2.0 (from fix-v0.2.0 branch), which added explicit `public` visibility modifiers and return types throughout the generated Kotlin code. Regenerated with the correct version. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

Extract the channel data migration block from build_with_store_internal into a standalone apply_channel_data_migration function, making it generic over the keys manager type to enable direct unit testing. Add 7 unit tests covering all code paths: - Invalid monitor data returns ReadFailed - Empty monitors list succeeds - Fresh write to empty store - Equal update_id (existing == migrated) still writes - Existing data with same update_id gets overwritten - Corrupt existing data triggers fail-closed (ReadFailed) - Store IO error triggers fail-closed (ReadFailed) Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

ben-kaufman · 2026-03-11T04:57:05Z

Looks good, but needs to add some tests ideally.

coreyphillips · 2026-03-11T14:50:09Z

One nit. Line 1666 uses > so when existing_update_id == migrated_update_id, we still write. Could use >= to skip the redundant write. Unless I'm missing something.

Use >= instead of > when comparing existing vs migrated update_id. When they are equal, the write is redundant — the store already has equivalent state. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

ovitrif · 2026-03-11T22:10:25Z

One nit. Line 1666 uses > so when existing_update_id == migrated_update_id, we still write. Could use >= to skip the redundant write. Unless I'm missing something.

Good catch — changed to >= so equal update_id skips the redundant write. Updated tests accordingly.

See cde6037

ovitrif · 2026-03-11T22:10:33Z

Looks good, but needs to add some tests ideally.

Added in 630b364 — 6 unit tests covering all migration code paths (invalid data, empty list, fresh write, skip on equal/newer update_id, corrupt existing data, store IO error). Run with cargo test --lib test_migration.

Add existence-check guard to channel manager migration (same fail-closed pattern as monitors). If the KV store already has a channel manager, the migration is skipped to prevent rolling back HTLC states, commitment indices, and pending payments with stale FS-sourced data. Also adds AI rule requiring unit tests for new business logic, and 3 unit tests covering the channel manager guard (fresh write, skip when existing, fail on read error). Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

…t-data test Correct the comment explaining why the channel manager guard uses existence-only checks: the real blocker is that ChannelManagerReadArgs requires already-deserialized channel monitors (loaded later in build), not that infrastructure is unavailable. Add test documenting intentional behavior when corrupt data exists in the store: migration is skipped (existence-only), preserving corrupt data. This is the correct trade-off — overwriting a potentially valid channel manager with stale FS data risks fund loss. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

# Conflicts: # CHANGELOG.md # Package.swift # bindings/kotlin/ldk-node-android/lib/src/main/jniLibs/arm64-v8a/libldk_node.so # bindings/kotlin/ldk-node-android/lib/src/main/jniLibs/armeabi-v7a/libldk_node.so # bindings/kotlin/ldk-node-android/lib/src/main/jniLibs/x86_64/libldk_node.so # bindings/kotlin/ldk-node-jvm/lib/src/main/resources/darwin-aarch64/libldk_node.dylib # bindings/kotlin/ldk-node-jvm/lib/src/main/resources/darwin-x86-64/libldk_node.dylib

ben-kaufman · 2026-03-12T02:06:27Z

@ovitrif new bindings and version bump missing?

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

ovitrif · 2026-03-12T10:08:00Z

@ovitrif new bindings and version bump missing?

Yes, was planned, now it's done 🫡

ovitrif · 2026-03-13T02:18:47Z

@ben-kaufman version bump skipped because there was no release of the previous PR (#75), where the artifacts got picked up by a CI job for the apps repos.

Thus there was no reason to bump the version.
Instead, I deleted the former rc.33 packages, release, and tag; rebuilt bindings and release, so now this encapsulates the changes of both PRs, on tip of this branch.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

ovitrif · 2026-03-09T22:24:29Z

Summary

Fixed FS→KV channel data migration blindly overwriting newer state (monitors and channel manager).
The migration now skips writing a channel monitor when the KV store already holds one with a newer or equal update_id, and skips the channel manager when one already exists.
Migration read/deserialization failures now fail-closed (ReadFailed) to prevent silent data loss.

Test plan

cargo fmt --check — clean
cargo build — builds successfully
cargo test --lib — all unit tests pass (including new migration tests)
Bindings regenerated (Swift/Kotlin/Python)
xcframework checksum verified in Package.swift

Release

v0.7.0-rc.33

During FS→KV store migration, the code now compares update_id before writing and skips if the KV store already has a newer monitor. This prevents stale migration data from clobbering current channel state on repeated restarts. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

Replace incomplete testing docs with comprehensive coverage of unit tests, integration tests (with macOS ulimit note), and CLN/LND/VSS backend tests using correct RUSTFLAGS syntax. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

Address PR review: change channel monitor migration error handling from fail-open (proceed with overwrite) to fail-closed (abort with ReadFailed) for both deserialization errors and non-NotFound I/O errors. This prevents silent data loss when the KV store has valid data that can't be read due to transient errors or format changes. Also moves CHANGELOG entry to Synonym Fork Additions per project convention. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

The previous bindgen run used gobley-uniffi-bindgen v0.3.7 (from gobley main branch) instead of v0.2.0 (from fix-v0.2.0 branch), which added explicit `public` visibility modifiers and return types throughout the generated Kotlin code. Regenerated with the correct version. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

Extract the channel data migration block from build_with_store_internal into a standalone apply_channel_data_migration function, making it generic over the keys manager type to enable direct unit testing. Add 7 unit tests covering all code paths: - Invalid monitor data returns ReadFailed - Empty monitors list succeeds - Fresh write to empty store - Equal update_id (existing == migrated) still writes - Existing data with same update_id gets overwritten - Corrupt existing data triggers fail-closed (ReadFailed) - Store IO error triggers fail-closed (ReadFailed) Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

ben-kaufman · 2026-03-11T04:57:05Z

Looks good, but needs to add some tests ideally.

coreyphillips · 2026-03-11T14:50:09Z

One nit. Line 1666 uses > so when existing_update_id == migrated_update_id, we still write. Could use >= to skip the redundant write. Unless I'm missing something.

Use >= instead of > when comparing existing vs migrated update_id. When they are equal, the write is redundant — the store already has equivalent state. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

ovitrif · 2026-03-11T22:10:25Z

One nit. Line 1666 uses > so when existing_update_id == migrated_update_id, we still write. Could use >= to skip the redundant write. Unless I'm missing something.

Good catch — changed to >= so equal update_id skips the redundant write. Updated tests accordingly.

See cde6037

ovitrif · 2026-03-11T22:10:33Z

Looks good, but needs to add some tests ideally.

Added in 630b364 — 6 unit tests covering all migration code paths (invalid data, empty list, fresh write, skip on equal/newer update_id, corrupt existing data, store IO error). Run with cargo test --lib test_migration.

Add existence-check guard to channel manager migration (same fail-closed pattern as monitors). If the KV store already has a channel manager, the migration is skipped to prevent rolling back HTLC states, commitment indices, and pending payments with stale FS-sourced data. Also adds AI rule requiring unit tests for new business logic, and 3 unit tests covering the channel manager guard (fresh write, skip when existing, fail on read error). Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

…t-data test Correct the comment explaining why the channel manager guard uses existence-only checks: the real blocker is that ChannelManagerReadArgs requires already-deserialized channel monitors (loaded later in build), not that infrastructure is unavailable. Add test documenting intentional behavior when corrupt data exists in the store: migration is skipped (existence-only), preserving corrupt data. This is the correct trade-off — overwriting a potentially valid channel manager with stale FS data risks fund loss. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

# Conflicts: # CHANGELOG.md # Package.swift # bindings/kotlin/ldk-node-android/lib/src/main/jniLibs/arm64-v8a/libldk_node.so # bindings/kotlin/ldk-node-android/lib/src/main/jniLibs/armeabi-v7a/libldk_node.so # bindings/kotlin/ldk-node-android/lib/src/main/jniLibs/x86_64/libldk_node.so # bindings/kotlin/ldk-node-jvm/lib/src/main/resources/darwin-aarch64/libldk_node.dylib # bindings/kotlin/ldk-node-jvm/lib/src/main/resources/darwin-x86-64/libldk_node.dylib

ben-kaufman · 2026-03-12T02:06:27Z

@ovitrif new bindings and version bump missing?

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

ovitrif · 2026-03-12T10:08:00Z

@ovitrif new bindings and version bump missing?

Yes, was planned, now it's done 🫡

ovitrif · 2026-03-13T02:18:47Z

@ben-kaufman version bump skipped because there was no release of the previous PR (#75), where the artifacts got picked up by a CI job for the apps repos.

Thus there was no reason to bump the version.
Instead, I deleted the former rc.33 packages, release, and tag; rebuilt bindings and release, so now this encapsulates the changes of both PRs, on tip of this branch.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

ovitrif · 2026-03-09T22:24:29Z

Summary

Fixed FS→KV channel data migration blindly overwriting newer state (monitors and channel manager).
The migration now skips writing a channel monitor when the KV store already holds one with a newer or equal update_id, and skips the channel manager when one already exists.
Migration read/deserialization failures now fail-closed (ReadFailed) to prevent silent data loss.

Test plan

cargo fmt --check — clean
cargo build — builds successfully
cargo test --lib — all unit tests pass (including new migration tests)
Bindings regenerated (Swift/Kotlin/Python)
xcframework checksum verified in Package.swift

Release

v0.7.0-rc.33

During FS→KV store migration, the code now compares update_id before writing and skips if the KV store already has a newer monitor. This prevents stale migration data from clobbering current channel state on repeated restarts. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

Replace incomplete testing docs with comprehensive coverage of unit tests, integration tests (with macOS ulimit note), and CLN/LND/VSS backend tests using correct RUSTFLAGS syntax. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

Address PR review: change channel monitor migration error handling from fail-open (proceed with overwrite) to fail-closed (abort with ReadFailed) for both deserialization errors and non-NotFound I/O errors. This prevents silent data loss when the KV store has valid data that can't be read due to transient errors or format changes. Also moves CHANGELOG entry to Synonym Fork Additions per project convention. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

The previous bindgen run used gobley-uniffi-bindgen v0.3.7 (from gobley main branch) instead of v0.2.0 (from fix-v0.2.0 branch), which added explicit `public` visibility modifiers and return types throughout the generated Kotlin code. Regenerated with the correct version. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

Extract the channel data migration block from build_with_store_internal into a standalone apply_channel_data_migration function, making it generic over the keys manager type to enable direct unit testing. Add 7 unit tests covering all code paths: - Invalid monitor data returns ReadFailed - Empty monitors list succeeds - Fresh write to empty store - Equal update_id (existing == migrated) still writes - Existing data with same update_id gets overwritten - Corrupt existing data triggers fail-closed (ReadFailed) - Store IO error triggers fail-closed (ReadFailed) Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

ben-kaufman · 2026-03-11T04:57:05Z

Looks good, but needs to add some tests ideally.

coreyphillips · 2026-03-11T14:50:09Z

One nit. Line 1666 uses > so when existing_update_id == migrated_update_id, we still write. Could use >= to skip the redundant write. Unless I'm missing something.

Use >= instead of > when comparing existing vs migrated update_id. When they are equal, the write is redundant — the store already has equivalent state. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

ovitrif · 2026-03-11T22:10:25Z

One nit. Line 1666 uses > so when existing_update_id == migrated_update_id, we still write. Could use >= to skip the redundant write. Unless I'm missing something.

Good catch — changed to >= so equal update_id skips the redundant write. Updated tests accordingly.

See cde6037

ovitrif · 2026-03-11T22:10:33Z

Looks good, but needs to add some tests ideally.

Added in 630b364 — 6 unit tests covering all migration code paths (invalid data, empty list, fresh write, skip on equal/newer update_id, corrupt existing data, store IO error). Run with cargo test --lib test_migration.

Add existence-check guard to channel manager migration (same fail-closed pattern as monitors). If the KV store already has a channel manager, the migration is skipped to prevent rolling back HTLC states, commitment indices, and pending payments with stale FS-sourced data. Also adds AI rule requiring unit tests for new business logic, and 3 unit tests covering the channel manager guard (fresh write, skip when existing, fail on read error). Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

…t-data test Correct the comment explaining why the channel manager guard uses existence-only checks: the real blocker is that ChannelManagerReadArgs requires already-deserialized channel monitors (loaded later in build), not that infrastructure is unavailable. Add test documenting intentional behavior when corrupt data exists in the store: migration is skipped (existence-only), preserving corrupt data. This is the correct trade-off — overwriting a potentially valid channel manager with stale FS data risks fund loss. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

# Conflicts: # CHANGELOG.md # Package.swift # bindings/kotlin/ldk-node-android/lib/src/main/jniLibs/arm64-v8a/libldk_node.so # bindings/kotlin/ldk-node-android/lib/src/main/jniLibs/armeabi-v7a/libldk_node.so # bindings/kotlin/ldk-node-android/lib/src/main/jniLibs/x86_64/libldk_node.so # bindings/kotlin/ldk-node-jvm/lib/src/main/resources/darwin-aarch64/libldk_node.dylib # bindings/kotlin/ldk-node-jvm/lib/src/main/resources/darwin-x86-64/libldk_node.dylib

ben-kaufman · 2026-03-12T02:06:27Z

@ovitrif new bindings and version bump missing?

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

ovitrif · 2026-03-12T10:08:00Z

@ovitrif new bindings and version bump missing?

Yes, was planned, now it's done 🫡

ovitrif · 2026-03-13T02:18:47Z

@ben-kaufman version bump skipped because there was no release of the previous PR (#75), where the artifacts got picked up by a CI job for the apps repos.

Thus there was no reason to bump the version.
Instead, I deleted the former rc.33 packages, release, and tag; rebuilt bindings and release, so now this encapsulates the changes of both PRs, on tip of this branch.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

ovitrif · 2026-03-09T22:24:29Z

Summary

Fixed FS→KV channel data migration blindly overwriting newer state (monitors and channel manager).
The migration now skips writing a channel monitor when the KV store already holds one with a newer or equal update_id, and skips the channel manager when one already exists.
Migration read/deserialization failures now fail-closed (ReadFailed) to prevent silent data loss.

Test plan

cargo fmt --check — clean
cargo build — builds successfully
cargo test --lib — all unit tests pass (including new migration tests)
Bindings regenerated (Swift/Kotlin/Python)
xcframework checksum verified in Package.swift

Release

v0.7.0-rc.33

During FS→KV store migration, the code now compares update_id before writing and skips if the KV store already has a newer monitor. This prevents stale migration data from clobbering current channel state on repeated restarts. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

Replace incomplete testing docs with comprehensive coverage of unit tests, integration tests (with macOS ulimit note), and CLN/LND/VSS backend tests using correct RUSTFLAGS syntax. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

Address PR review: change channel monitor migration error handling from fail-open (proceed with overwrite) to fail-closed (abort with ReadFailed) for both deserialization errors and non-NotFound I/O errors. This prevents silent data loss when the KV store has valid data that can't be read due to transient errors or format changes. Also moves CHANGELOG entry to Synonym Fork Additions per project convention. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

The previous bindgen run used gobley-uniffi-bindgen v0.3.7 (from gobley main branch) instead of v0.2.0 (from fix-v0.2.0 branch), which added explicit `public` visibility modifiers and return types throughout the generated Kotlin code. Regenerated with the correct version. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

Extract the channel data migration block from build_with_store_internal into a standalone apply_channel_data_migration function, making it generic over the keys manager type to enable direct unit testing. Add 7 unit tests covering all code paths: - Invalid monitor data returns ReadFailed - Empty monitors list succeeds - Fresh write to empty store - Equal update_id (existing == migrated) still writes - Existing data with same update_id gets overwritten - Corrupt existing data triggers fail-closed (ReadFailed) - Store IO error triggers fail-closed (ReadFailed) Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

ben-kaufman · 2026-03-11T04:57:05Z

Looks good, but needs to add some tests ideally.

coreyphillips · 2026-03-11T14:50:09Z

One nit. Line 1666 uses > so when existing_update_id == migrated_update_id, we still write. Could use >= to skip the redundant write. Unless I'm missing something.

Use >= instead of > when comparing existing vs migrated update_id. When they are equal, the write is redundant — the store already has equivalent state. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

ovitrif · 2026-03-11T22:10:25Z

One nit. Line 1666 uses > so when existing_update_id == migrated_update_id, we still write. Could use >= to skip the redundant write. Unless I'm missing something.

Good catch — changed to >= so equal update_id skips the redundant write. Updated tests accordingly.

See cde6037

ovitrif · 2026-03-11T22:10:33Z

Looks good, but needs to add some tests ideally.

Added in 630b364 — 6 unit tests covering all migration code paths (invalid data, empty list, fresh write, skip on equal/newer update_id, corrupt existing data, store IO error). Run with cargo test --lib test_migration.

Add existence-check guard to channel manager migration (same fail-closed pattern as monitors). If the KV store already has a channel manager, the migration is skipped to prevent rolling back HTLC states, commitment indices, and pending payments with stale FS-sourced data. Also adds AI rule requiring unit tests for new business logic, and 3 unit tests covering the channel manager guard (fresh write, skip when existing, fail on read error). Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

…t-data test Correct the comment explaining why the channel manager guard uses existence-only checks: the real blocker is that ChannelManagerReadArgs requires already-deserialized channel monitors (loaded later in build), not that infrastructure is unavailable. Add test documenting intentional behavior when corrupt data exists in the store: migration is skipped (existence-only), preserving corrupt data. This is the correct trade-off — overwriting a potentially valid channel manager with stale FS data risks fund loss. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

# Conflicts: # CHANGELOG.md # Package.swift # bindings/kotlin/ldk-node-android/lib/src/main/jniLibs/arm64-v8a/libldk_node.so # bindings/kotlin/ldk-node-android/lib/src/main/jniLibs/armeabi-v7a/libldk_node.so # bindings/kotlin/ldk-node-android/lib/src/main/jniLibs/x86_64/libldk_node.so # bindings/kotlin/ldk-node-jvm/lib/src/main/resources/darwin-aarch64/libldk_node.dylib # bindings/kotlin/ldk-node-jvm/lib/src/main/resources/darwin-x86-64/libldk_node.dylib

ben-kaufman · 2026-03-12T02:06:27Z

@ovitrif new bindings and version bump missing?

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

ovitrif · 2026-03-12T10:08:00Z

@ovitrif new bindings and version bump missing?

Yes, was planned, now it's done 🫡

ovitrif · 2026-03-13T02:18:47Z

@ben-kaufman version bump skipped because there was no release of the previous PR (#75), where the artifacts got picked up by a CI job for the apps repos.

Thus there was no reason to bump the version.
Instead, I deleted the former rc.33 packages, release, and tag; rebuilt bindings and release, so now this encapsulates the changes of both PRs, on tip of this branch.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

ovitrif · 2026-03-09T22:24:29Z

Summary

Fixed FS→KV channel data migration blindly overwriting newer state (monitors and channel manager).
The migration now skips writing a channel monitor when the KV store already holds one with a newer or equal update_id, and skips the channel manager when one already exists.
Migration read/deserialization failures now fail-closed (ReadFailed) to prevent silent data loss.

Test plan

cargo fmt --check — clean
cargo build — builds successfully
cargo test --lib — all unit tests pass (including new migration tests)
Bindings regenerated (Swift/Kotlin/Python)
xcframework checksum verified in Package.swift

Release

v0.7.0-rc.33

During FS→KV store migration, the code now compares update_id before writing and skips if the KV store already has a newer monitor. This prevents stale migration data from clobbering current channel state on repeated restarts. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

Replace incomplete testing docs with comprehensive coverage of unit tests, integration tests (with macOS ulimit note), and CLN/LND/VSS backend tests using correct RUSTFLAGS syntax. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

Address PR review: change channel monitor migration error handling from fail-open (proceed with overwrite) to fail-closed (abort with ReadFailed) for both deserialization errors and non-NotFound I/O errors. This prevents silent data loss when the KV store has valid data that can't be read due to transient errors or format changes. Also moves CHANGELOG entry to Synonym Fork Additions per project convention. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

The previous bindgen run used gobley-uniffi-bindgen v0.3.7 (from gobley main branch) instead of v0.2.0 (from fix-v0.2.0 branch), which added explicit `public` visibility modifiers and return types throughout the generated Kotlin code. Regenerated with the correct version. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

Extract the channel data migration block from build_with_store_internal into a standalone apply_channel_data_migration function, making it generic over the keys manager type to enable direct unit testing. Add 7 unit tests covering all code paths: - Invalid monitor data returns ReadFailed - Empty monitors list succeeds - Fresh write to empty store - Equal update_id (existing == migrated) still writes - Existing data with same update_id gets overwritten - Corrupt existing data triggers fail-closed (ReadFailed) - Store IO error triggers fail-closed (ReadFailed) Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

ben-kaufman · 2026-03-11T04:57:05Z

Looks good, but needs to add some tests ideally.

coreyphillips · 2026-03-11T14:50:09Z

One nit. Line 1666 uses > so when existing_update_id == migrated_update_id, we still write. Could use >= to skip the redundant write. Unless I'm missing something.

Use >= instead of > when comparing existing vs migrated update_id. When they are equal, the write is redundant — the store already has equivalent state. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

ovitrif · 2026-03-11T22:10:25Z

One nit. Line 1666 uses > so when existing_update_id == migrated_update_id, we still write. Could use >= to skip the redundant write. Unless I'm missing something.

Good catch — changed to >= so equal update_id skips the redundant write. Updated tests accordingly.

See cde6037

ovitrif · 2026-03-11T22:10:33Z

Looks good, but needs to add some tests ideally.

Added in 630b364 — 6 unit tests covering all migration code paths (invalid data, empty list, fresh write, skip on equal/newer update_id, corrupt existing data, store IO error). Run with cargo test --lib test_migration.

Add existence-check guard to channel manager migration (same fail-closed pattern as monitors). If the KV store already has a channel manager, the migration is skipped to prevent rolling back HTLC states, commitment indices, and pending payments with stale FS-sourced data. Also adds AI rule requiring unit tests for new business logic, and 3 unit tests covering the channel manager guard (fresh write, skip when existing, fail on read error). Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

…t-data test Correct the comment explaining why the channel manager guard uses existence-only checks: the real blocker is that ChannelManagerReadArgs requires already-deserialized channel monitors (loaded later in build), not that infrastructure is unavailable. Add test documenting intentional behavior when corrupt data exists in the store: migration is skipped (existence-only), preserving corrupt data. This is the correct trade-off — overwriting a potentially valid channel manager with stale FS data risks fund loss. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

# Conflicts: # CHANGELOG.md # Package.swift # bindings/kotlin/ldk-node-android/lib/src/main/jniLibs/arm64-v8a/libldk_node.so # bindings/kotlin/ldk-node-android/lib/src/main/jniLibs/armeabi-v7a/libldk_node.so # bindings/kotlin/ldk-node-android/lib/src/main/jniLibs/x86_64/libldk_node.so # bindings/kotlin/ldk-node-jvm/lib/src/main/resources/darwin-aarch64/libldk_node.dylib # bindings/kotlin/ldk-node-jvm/lib/src/main/resources/darwin-x86-64/libldk_node.dylib

ben-kaufman · 2026-03-12T02:06:27Z

@ovitrif new bindings and version bump missing?

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

ovitrif · 2026-03-12T10:08:00Z

@ovitrif new bindings and version bump missing?

Yes, was planned, now it's done 🫡

ovitrif · 2026-03-13T02:18:47Z

@ben-kaufman version bump skipped because there was no release of the previous PR (#75), where the artifacts got picked up by a CI job for the apps repos.

Thus there was no reason to bump the version.
Instead, I deleted the former rc.33 packages, release, and tag; rebuilt bindings and release, so now this encapsulates the changes of both PRs, on tip of this branch.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

ovitrif · 2026-03-09T22:24:29Z

Summary

Fixed FS→KV channel data migration blindly overwriting newer state (monitors and channel manager).
The migration now skips writing a channel monitor when the KV store already holds one with a newer or equal update_id, and skips the channel manager when one already exists.
Migration read/deserialization failures now fail-closed (ReadFailed) to prevent silent data loss.

Test plan

cargo fmt --check — clean
cargo build — builds successfully
cargo test --lib — all unit tests pass (including new migration tests)
Bindings regenerated (Swift/Kotlin/Python)
xcframework checksum verified in Package.swift

Release

v0.7.0-rc.33

During FS→KV store migration, the code now compares update_id before writing and skips if the KV store already has a newer monitor. This prevents stale migration data from clobbering current channel state on repeated restarts. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

Replace incomplete testing docs with comprehensive coverage of unit tests, integration tests (with macOS ulimit note), and CLN/LND/VSS backend tests using correct RUSTFLAGS syntax. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

Address PR review: change channel monitor migration error handling from fail-open (proceed with overwrite) to fail-closed (abort with ReadFailed) for both deserialization errors and non-NotFound I/O errors. This prevents silent data loss when the KV store has valid data that can't be read due to transient errors or format changes. Also moves CHANGELOG entry to Synonym Fork Additions per project convention. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

The previous bindgen run used gobley-uniffi-bindgen v0.3.7 (from gobley main branch) instead of v0.2.0 (from fix-v0.2.0 branch), which added explicit `public` visibility modifiers and return types throughout the generated Kotlin code. Regenerated with the correct version. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

Extract the channel data migration block from build_with_store_internal into a standalone apply_channel_data_migration function, making it generic over the keys manager type to enable direct unit testing. Add 7 unit tests covering all code paths: - Invalid monitor data returns ReadFailed - Empty monitors list succeeds - Fresh write to empty store - Equal update_id (existing == migrated) still writes - Existing data with same update_id gets overwritten - Corrupt existing data triggers fail-closed (ReadFailed) - Store IO error triggers fail-closed (ReadFailed) Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

ben-kaufman · 2026-03-11T04:57:05Z

Looks good, but needs to add some tests ideally.

coreyphillips · 2026-03-11T14:50:09Z

One nit. Line 1666 uses > so when existing_update_id == migrated_update_id, we still write. Could use >= to skip the redundant write. Unless I'm missing something.

Use >= instead of > when comparing existing vs migrated update_id. When they are equal, the write is redundant — the store already has equivalent state. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

ovitrif · 2026-03-11T22:10:25Z

One nit. Line 1666 uses > so when existing_update_id == migrated_update_id, we still write. Could use >= to skip the redundant write. Unless I'm missing something.

Good catch — changed to >= so equal update_id skips the redundant write. Updated tests accordingly.

See cde6037

ovitrif · 2026-03-11T22:10:33Z

Looks good, but needs to add some tests ideally.

Added in 630b364 — 6 unit tests covering all migration code paths (invalid data, empty list, fresh write, skip on equal/newer update_id, corrupt existing data, store IO error). Run with cargo test --lib test_migration.

Add existence-check guard to channel manager migration (same fail-closed pattern as monitors). If the KV store already has a channel manager, the migration is skipped to prevent rolling back HTLC states, commitment indices, and pending payments with stale FS-sourced data. Also adds AI rule requiring unit tests for new business logic, and 3 unit tests covering the channel manager guard (fresh write, skip when existing, fail on read error). Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

…t-data test Correct the comment explaining why the channel manager guard uses existence-only checks: the real blocker is that ChannelManagerReadArgs requires already-deserialized channel monitors (loaded later in build), not that infrastructure is unavailable. Add test documenting intentional behavior when corrupt data exists in the store: migration is skipped (existence-only), preserving corrupt data. This is the correct trade-off — overwriting a potentially valid channel manager with stale FS data risks fund loss. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

# Conflicts: # CHANGELOG.md # Package.swift # bindings/kotlin/ldk-node-android/lib/src/main/jniLibs/arm64-v8a/libldk_node.so # bindings/kotlin/ldk-node-android/lib/src/main/jniLibs/armeabi-v7a/libldk_node.so # bindings/kotlin/ldk-node-android/lib/src/main/jniLibs/x86_64/libldk_node.so # bindings/kotlin/ldk-node-jvm/lib/src/main/resources/darwin-aarch64/libldk_node.dylib # bindings/kotlin/ldk-node-jvm/lib/src/main/resources/darwin-x86-64/libldk_node.dylib

ben-kaufman · 2026-03-12T02:06:27Z

@ovitrif new bindings and version bump missing?

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

ovitrif · 2026-03-12T10:08:00Z

@ovitrif new bindings and version bump missing?

Yes, was planned, now it's done 🫡

ovitrif · 2026-03-13T02:18:47Z

@ben-kaufman version bump skipped because there was no release of the previous PR (#75), where the artifacts got picked up by a CI job for the apps repos.

Thus there was no reason to bump the version.
Instead, I deleted the former rc.33 packages, release, and tag; rebuilt bindings and release, so now this encapsulates the changes of both PRs, on tip of this branch.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

ovitrifand others added 4 commits March 9, 2026 22:22

chore: bump version to 0.7.0-rc.33
530f449
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

chore: regenerate bindings for v0.7.0-rc.33
fdab2be
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

This comment was marked as resolved.
Sign in to view

ovitrif requested review from ben-kaufman and coreyphillips and removed request for ben-kaufman and coreyphillips March 10, 2026 18:31

ovitrifand others added 2 commits March 10, 2026 21:03

chore: regenerate bindings for v0.7.0-rc.33
e998973
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

ovitrif marked this pull request as draft March 10, 2026 22:01

ovitrif changed the title ~~fix: prevent channel monitor migration from overwriting newer state~~fix: prevent channel monitor overwrite in migrationsMar 10, 2026

ovitrif marked this pull request as ready for review March 11, 2026 00:46

ovitrif requested review from ben-kaufman and coreyphillips March 11, 2026 03:03

coreyphillips approved these changes Mar 11, 2026
View reviewed changes

fix: skip redundant monitor write when update_id is equal
cde6037
Use >= instead of > when comparing existing vs migrated update_id. When they are equal, the write is redundant — the store already has equivalent state. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

claudeBot reviewed Mar 11, 2026
View reviewed changes

Comment threadsrc/builder.rs

ovitrif marked this pull request as draft March 11, 2026 23:27

ovitrifand others added 2 commits March 12, 2026 00:29

ovitrif marked this pull request as ready for review March 12, 2026 01:17

chore: reword changelog and regenerate bindings for v0.7.0-rc.33
1531af7
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

ovitrif requested a review from coreyphillips March 12, 2026 09:04

ovitrif changed the title ~~fix: prevent channel monitor overwrite in migrations~~fix: prevent channel data overwrite in migrationsMar 13, 2026

ben-kaufman reviewed Mar 16, 2026
View reviewed changes

Comment threadsrc/builder.rs Outdated

ben-kaufman approved these changes Mar 16, 2026
View reviewed changes

ovitrif changed the title ~~fix: prevent channel data overwrite in migrations~~fix: prevent channel data overwrite in FS→KV migrationMar 16, 2026

fix: simplify channel manager migration comment
ffc0ec5
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

coreyphillips approved these changes Mar 16, 2026
View reviewed changes

ovitrif merged commit 1569477 into mainMar 16, 2026
2 checks passed

ovitrif deleted the fix/channel-monitor-migration branch March 18, 2026 10:21

ovitrif mentioned this pull request Mar 18, 2026
fix: bump ldk-node to v0.7.0-rc.33 synonymdev/bitkit-ios#498
Closed

ovitrifand others added 4 commits March 9, 2026 22:22

chore: bump version to 0.7.0-rc.33
530f449
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

chore: regenerate bindings for v0.7.0-rc.33
fdab2be
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

This comment was marked as resolved.
Sign in to view

ovitrif requested review from ben-kaufman and coreyphillips and removed request for ben-kaufman and coreyphillips March 10, 2026 18:31

ovitrifand others added 2 commits March 10, 2026 21:03

chore: regenerate bindings for v0.7.0-rc.33
e998973
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

ovitrif marked this pull request as draft March 10, 2026 22:01

ovitrif changed the title ~~fix: prevent channel monitor migration from overwriting newer state~~fix: prevent channel monitor overwrite in migrationsMar 10, 2026

ovitrif marked this pull request as ready for review March 11, 2026 00:46

ovitrif requested review from ben-kaufman and coreyphillips March 11, 2026 03:03

coreyphillips approved these changes Mar 11, 2026
View reviewed changes

fix: skip redundant monitor write when update_id is equal
cde6037
Use >= instead of > when comparing existing vs migrated update_id. When they are equal, the write is redundant — the store already has equivalent state. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

claudeBot reviewed Mar 11, 2026
View reviewed changes

Comment threadsrc/builder.rs

ovitrif marked this pull request as draft March 11, 2026 23:27

ovitrifand others added 2 commits March 12, 2026 00:29

ovitrif marked this pull request as ready for review March 12, 2026 01:17

chore: reword changelog and regenerate bindings for v0.7.0-rc.33
1531af7
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

ovitrif requested a review from coreyphillips March 12, 2026 09:04

ovitrif changed the title ~~fix: prevent channel monitor overwrite in migrations~~fix: prevent channel data overwrite in migrationsMar 13, 2026

ben-kaufman reviewed Mar 16, 2026
View reviewed changes

Comment threadsrc/builder.rs Outdated

ben-kaufman approved these changes Mar 16, 2026
View reviewed changes

ovitrif changed the title ~~fix: prevent channel data overwrite in migrations~~fix: prevent channel data overwrite in FS→KV migrationMar 16, 2026

fix: simplify channel manager migration comment
ffc0ec5
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

coreyphillips approved these changes Mar 16, 2026
View reviewed changes

ovitrif merged commit 1569477 into mainMar 16, 2026
2 checks passed

ovitrif deleted the fix/channel-monitor-migration branch March 18, 2026 10:21

ovitrif mentioned this pull request Mar 18, 2026
fix: bump ldk-node to v0.7.0-rc.33 synonymdev/bitkit-ios#498
Closed

ovitrifand others added 4 commits March 9, 2026 22:22

chore: bump version to 0.7.0-rc.33
530f449
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

chore: regenerate bindings for v0.7.0-rc.33
fdab2be
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

This comment was marked as resolved.
Sign in to view

ovitrif requested review from ben-kaufman and coreyphillips and removed request for ben-kaufman and coreyphillips March 10, 2026 18:31

ovitrifand others added 2 commits March 10, 2026 21:03

chore: regenerate bindings for v0.7.0-rc.33
e998973
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

ovitrif marked this pull request as draft March 10, 2026 22:01

ovitrif changed the title ~~fix: prevent channel monitor migration from overwriting newer state~~fix: prevent channel monitor overwrite in migrationsMar 10, 2026

ovitrif marked this pull request as ready for review March 11, 2026 00:46

ovitrif requested review from ben-kaufman and coreyphillips March 11, 2026 03:03

coreyphillips approved these changes Mar 11, 2026
View reviewed changes

fix: skip redundant monitor write when update_id is equal
cde6037
Use >= instead of > when comparing existing vs migrated update_id. When they are equal, the write is redundant — the store already has equivalent state. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

claudeBot reviewed Mar 11, 2026
View reviewed changes

Comment threadsrc/builder.rs

ovitrif marked this pull request as draft March 11, 2026 23:27

ovitrifand others added 2 commits March 12, 2026 00:29

ovitrif marked this pull request as ready for review March 12, 2026 01:17

chore: reword changelog and regenerate bindings for v0.7.0-rc.33
1531af7
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

ovitrif requested a review from coreyphillips March 12, 2026 09:04

ovitrif changed the title ~~fix: prevent channel monitor overwrite in migrations~~fix: prevent channel data overwrite in migrationsMar 13, 2026

ben-kaufman reviewed Mar 16, 2026
View reviewed changes

Comment threadsrc/builder.rs Outdated

ben-kaufman approved these changes Mar 16, 2026
View reviewed changes

ovitrif changed the title ~~fix: prevent channel data overwrite in migrations~~fix: prevent channel data overwrite in FS→KV migrationMar 16, 2026

fix: simplify channel manager migration comment
ffc0ec5
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

coreyphillips approved these changes Mar 16, 2026
View reviewed changes

ovitrif merged commit 1569477 into mainMar 16, 2026
2 checks passed

ovitrif deleted the fix/channel-monitor-migration branch March 18, 2026 10:21

ovitrif mentioned this pull request Mar 18, 2026
fix: bump ldk-node to v0.7.0-rc.33 synonymdev/bitkit-ios#498
Closed

ovitrifand others added 4 commits March 9, 2026 22:22

chore: bump version to 0.7.0-rc.33
530f449
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

chore: regenerate bindings for v0.7.0-rc.33
fdab2be
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

This comment was marked as resolved.
Sign in to view

ovitrif requested review from ben-kaufman and coreyphillips and removed request for ben-kaufman and coreyphillips March 10, 2026 18:31

ovitrifand others added 2 commits March 10, 2026 21:03

chore: regenerate bindings for v0.7.0-rc.33
e998973
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

ovitrif marked this pull request as draft March 10, 2026 22:01

ovitrif changed the title ~~fix: prevent channel monitor migration from overwriting newer state~~fix: prevent channel monitor overwrite in migrationsMar 10, 2026

ovitrif marked this pull request as ready for review March 11, 2026 00:46

ovitrif requested review from ben-kaufman and coreyphillips March 11, 2026 03:03

coreyphillips approved these changes Mar 11, 2026
View reviewed changes

fix: skip redundant monitor write when update_id is equal
cde6037
Use >= instead of > when comparing existing vs migrated update_id. When they are equal, the write is redundant — the store already has equivalent state. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

claudeBot reviewed Mar 11, 2026
View reviewed changes

Comment threadsrc/builder.rs

ovitrif marked this pull request as draft March 11, 2026 23:27

ovitrifand others added 2 commits March 12, 2026 00:29

ovitrif marked this pull request as ready for review March 12, 2026 01:17

chore: reword changelog and regenerate bindings for v0.7.0-rc.33
1531af7
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

ovitrif requested a review from coreyphillips March 12, 2026 09:04

ovitrif changed the title ~~fix: prevent channel monitor overwrite in migrations~~fix: prevent channel data overwrite in migrationsMar 13, 2026

ben-kaufman reviewed Mar 16, 2026
View reviewed changes

Comment threadsrc/builder.rs Outdated

ben-kaufman approved these changes Mar 16, 2026
View reviewed changes

ovitrif changed the title ~~fix: prevent channel data overwrite in migrations~~fix: prevent channel data overwrite in FS→KV migrationMar 16, 2026

fix: simplify channel manager migration comment
ffc0ec5
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

coreyphillips approved these changes Mar 16, 2026
View reviewed changes

ovitrif merged commit 1569477 into mainMar 16, 2026
2 checks passed

ovitrif deleted the fix/channel-monitor-migration branch March 18, 2026 10:21

ovitrif mentioned this pull request Mar 18, 2026
fix: bump ldk-node to v0.7.0-rc.33 synonymdev/bitkit-ios#498
Closed

ovitrifand others added 4 commits March 9, 2026 22:22

chore: bump version to 0.7.0-rc.33
530f449
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

chore: regenerate bindings for v0.7.0-rc.33
fdab2be
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

This comment was marked as resolved.
Sign in to view

ovitrif requested review from ben-kaufman and coreyphillips and removed request for ben-kaufman and coreyphillips March 10, 2026 18:31

ovitrifand others added 2 commits March 10, 2026 21:03

chore: regenerate bindings for v0.7.0-rc.33
e998973
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

ovitrif marked this pull request as draft March 10, 2026 22:01

ovitrif changed the title ~~fix: prevent channel monitor migration from overwriting newer state~~fix: prevent channel monitor overwrite in migrationsMar 10, 2026

ovitrif marked this pull request as ready for review March 11, 2026 00:46

ovitrif requested review from ben-kaufman and coreyphillips March 11, 2026 03:03

coreyphillips approved these changes Mar 11, 2026
View reviewed changes

fix: skip redundant monitor write when update_id is equal
cde6037
Use >= instead of > when comparing existing vs migrated update_id. When they are equal, the write is redundant — the store already has equivalent state. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

claudeBot reviewed Mar 11, 2026
View reviewed changes

Comment threadsrc/builder.rs

ovitrif marked this pull request as draft March 11, 2026 23:27

ovitrifand others added 2 commits March 12, 2026 00:29

ovitrif marked this pull request as ready for review March 12, 2026 01:17

chore: reword changelog and regenerate bindings for v0.7.0-rc.33
1531af7
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

ovitrif requested a review from coreyphillips March 12, 2026 09:04

ovitrif changed the title ~~fix: prevent channel monitor overwrite in migrations~~fix: prevent channel data overwrite in migrationsMar 13, 2026

ben-kaufman reviewed Mar 16, 2026
View reviewed changes

Comment threadsrc/builder.rs Outdated

ben-kaufman approved these changes Mar 16, 2026
View reviewed changes

ovitrif changed the title ~~fix: prevent channel data overwrite in migrations~~fix: prevent channel data overwrite in FS→KV migrationMar 16, 2026

fix: simplify channel manager migration comment
ffc0ec5
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

coreyphillips approved these changes Mar 16, 2026
View reviewed changes

ovitrif merged commit 1569477 into mainMar 16, 2026
2 checks passed

ovitrif deleted the fix/channel-monitor-migration branch March 18, 2026 10:21

ovitrif mentioned this pull request Mar 18, 2026
fix: bump ldk-node to v0.7.0-rc.33 synonymdev/bitkit-ios#498
Closed

ovitrifand others added 4 commits March 9, 2026 22:22

chore: bump version to 0.7.0-rc.33
530f449
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

chore: regenerate bindings for v0.7.0-rc.33
fdab2be
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

This comment was marked as resolved.
Sign in to view

ovitrif requested review from ben-kaufman and coreyphillips and removed request for ben-kaufman and coreyphillips March 10, 2026 18:31

ovitrifand others added 2 commits March 10, 2026 21:03

chore: regenerate bindings for v0.7.0-rc.33
e998973
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

ovitrif marked this pull request as draft March 10, 2026 22:01

ovitrif changed the title ~~fix: prevent channel monitor migration from overwriting newer state~~fix: prevent channel monitor overwrite in migrationsMar 10, 2026

ovitrif marked this pull request as ready for review March 11, 2026 00:46

ovitrif requested review from ben-kaufman and coreyphillips March 11, 2026 03:03

coreyphillips approved these changes Mar 11, 2026
View reviewed changes

fix: skip redundant monitor write when update_id is equal
cde6037
Use >= instead of > when comparing existing vs migrated update_id. When they are equal, the write is redundant — the store already has equivalent state. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

claudeBot reviewed Mar 11, 2026
View reviewed changes

Comment threadsrc/builder.rs

ovitrif marked this pull request as draft March 11, 2026 23:27

ovitrifand others added 2 commits March 12, 2026 00:29

ovitrif marked this pull request as ready for review March 12, 2026 01:17

chore: reword changelog and regenerate bindings for v0.7.0-rc.33
1531af7
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

ovitrif requested a review from coreyphillips March 12, 2026 09:04

ovitrif changed the title ~~fix: prevent channel monitor overwrite in migrations~~fix: prevent channel data overwrite in migrationsMar 13, 2026

ben-kaufman reviewed Mar 16, 2026
View reviewed changes

Comment threadsrc/builder.rs Outdated

ben-kaufman approved these changes Mar 16, 2026
View reviewed changes

ovitrif changed the title ~~fix: prevent channel data overwrite in migrations~~fix: prevent channel data overwrite in FS→KV migrationMar 16, 2026

fix: simplify channel manager migration comment
ffc0ec5
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

coreyphillips approved these changes Mar 16, 2026
View reviewed changes

ovitrif merged commit 1569477 into mainMar 16, 2026
2 checks passed

ovitrif deleted the fix/channel-monitor-migration branch March 18, 2026 10:21

ovitrif mentioned this pull request Mar 18, 2026
fix: bump ldk-node to v0.7.0-rc.33 synonymdev/bitkit-ios#498
Closed

ovitrifand others added 4 commits March 9, 2026 22:22

chore: bump version to 0.7.0-rc.33
530f449
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

chore: regenerate bindings for v0.7.0-rc.33
fdab2be
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

This comment was marked as resolved.
Sign in to view

ovitrif requested review from ben-kaufman and coreyphillips and removed request for ben-kaufman and coreyphillips March 10, 2026 18:31

ovitrifand others added 2 commits March 10, 2026 21:03

chore: regenerate bindings for v0.7.0-rc.33
e998973
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

ovitrif marked this pull request as draft March 10, 2026 22:01

ovitrif changed the title ~~fix: prevent channel monitor migration from overwriting newer state~~fix: prevent channel monitor overwrite in migrationsMar 10, 2026

ovitrif marked this pull request as ready for review March 11, 2026 00:46

ovitrif requested review from ben-kaufman and coreyphillips March 11, 2026 03:03

coreyphillips approved these changes Mar 11, 2026
View reviewed changes

fix: skip redundant monitor write when update_id is equal
cde6037
Use >= instead of > when comparing existing vs migrated update_id. When they are equal, the write is redundant — the store already has equivalent state. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

claudeBot reviewed Mar 11, 2026
View reviewed changes

Comment threadsrc/builder.rs

ovitrif marked this pull request as draft March 11, 2026 23:27

ovitrifand others added 2 commits March 12, 2026 00:29

ovitrif marked this pull request as ready for review March 12, 2026 01:17

chore: reword changelog and regenerate bindings for v0.7.0-rc.33
1531af7
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

ovitrif requested a review from coreyphillips March 12, 2026 09:04

ovitrif changed the title ~~fix: prevent channel monitor overwrite in migrations~~fix: prevent channel data overwrite in migrationsMar 13, 2026

ben-kaufman reviewed Mar 16, 2026
View reviewed changes

Comment threadsrc/builder.rs Outdated

ben-kaufman approved these changes Mar 16, 2026
View reviewed changes

ovitrif changed the title ~~fix: prevent channel data overwrite in migrations~~fix: prevent channel data overwrite in FS→KV migrationMar 16, 2026

fix: simplify channel manager migration comment
ffc0ec5
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

coreyphillips approved these changes Mar 16, 2026
View reviewed changes

ovitrif merged commit 1569477 into mainMar 16, 2026
2 checks passed

ovitrif deleted the fix/channel-monitor-migration branch March 18, 2026 10:21

ovitrif mentioned this pull request Mar 18, 2026
fix: bump ldk-node to v0.7.0-rc.33 synonymdev/bitkit-ios#498
Closed

ovitrifand others added 4 commits March 9, 2026 22:22

chore: bump version to 0.7.0-rc.33
530f449
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

chore: regenerate bindings for v0.7.0-rc.33
fdab2be
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

This comment was marked as resolved.
Sign in to view

ovitrif requested review from ben-kaufman and coreyphillips and removed request for ben-kaufman and coreyphillips March 10, 2026 18:31

ovitrifand others added 2 commits March 10, 2026 21:03

chore: regenerate bindings for v0.7.0-rc.33
e998973
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

ovitrif marked this pull request as draft March 10, 2026 22:01

ovitrif changed the title ~~fix: prevent channel monitor migration from overwriting newer state~~fix: prevent channel monitor overwrite in migrationsMar 10, 2026

ovitrif marked this pull request as ready for review March 11, 2026 00:46

ovitrif requested review from ben-kaufman and coreyphillips March 11, 2026 03:03

coreyphillips approved these changes Mar 11, 2026
View reviewed changes

fix: skip redundant monitor write when update_id is equal
cde6037
Use >= instead of > when comparing existing vs migrated update_id. When they are equal, the write is redundant — the store already has equivalent state. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

claudeBot reviewed Mar 11, 2026
View reviewed changes

Comment threadsrc/builder.rs

ovitrif marked this pull request as draft March 11, 2026 23:27

ovitrifand others added 2 commits March 12, 2026 00:29

ovitrif marked this pull request as ready for review March 12, 2026 01:17

chore: reword changelog and regenerate bindings for v0.7.0-rc.33
1531af7
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

ovitrif requested a review from coreyphillips March 12, 2026 09:04

ovitrif changed the title ~~fix: prevent channel monitor overwrite in migrations~~fix: prevent channel data overwrite in migrationsMar 13, 2026

ben-kaufman reviewed Mar 16, 2026
View reviewed changes

Comment threadsrc/builder.rs Outdated

ben-kaufman approved these changes Mar 16, 2026
View reviewed changes

ovitrif changed the title ~~fix: prevent channel data overwrite in migrations~~fix: prevent channel data overwrite in FS→KV migrationMar 16, 2026

fix: simplify channel manager migration comment
ffc0ec5
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

coreyphillips approved these changes Mar 16, 2026
View reviewed changes

ovitrif merged commit 1569477 into mainMar 16, 2026
2 checks passed

ovitrif deleted the fix/channel-monitor-migration branch March 18, 2026 10:21

ovitrif mentioned this pull request Mar 18, 2026
fix: bump ldk-node to v0.7.0-rc.33 synonymdev/bitkit-ios#498
Closed

Uh oh!

Conversation

ovitrif commented Mar 9, 2026• edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Summary

Test plan

Release

Uh oh!

This comment was marked as resolved.

Uh oh!

ben-kaufman commented Mar 11, 2026

Uh oh!

coreyphillips commented Mar 11, 2026• edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

ovitrif commented Mar 11, 2026• edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

ovitrif commented Mar 11, 2026• edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Uh oh!

ben-kaufman commented Mar 12, 2026

Uh oh!

ovitrif commented Mar 12, 2026

Uh oh!

ovitrif commented Mar 13, 2026• edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

Uh oh!

Conversation

ovitrif commented Mar 9, 2026• edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Summary

Test plan

Release

Uh oh!

This comment was marked as resolved.

Uh oh!

ben-kaufman commented Mar 11, 2026

Uh oh!

coreyphillips commented Mar 11, 2026• edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

ovitrif commented Mar 11, 2026• edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

ovitrif commented Mar 11, 2026• edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Uh oh!

ben-kaufman commented Mar 12, 2026

Uh oh!

ovitrif commented Mar 12, 2026

Uh oh!

ovitrif commented Mar 13, 2026• edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

Uh oh!

Conversation

ovitrif commented Mar 9, 2026•
edited
Loading

coreyphillips commented Mar 11, 2026•
edited
Loading

ovitrif commented Mar 11, 2026•
edited
Loading

ovitrif commented Mar 11, 2026•
edited
Loading

ovitrif commented Mar 13, 2026•
edited
Loading

ovitrif commented Mar 9, 2026•
edited
Loading

coreyphillips commented Mar 11, 2026•
edited
Loading

ovitrif commented Mar 11, 2026•
edited
Loading

ovitrif commented Mar 11, 2026•
edited
Loading

ovitrif commented Mar 13, 2026•
edited
Loading

ovitrif commented Mar 9, 2026•
edited
Loading

coreyphillips commented Mar 11, 2026•
edited
Loading

ovitrif commented Mar 11, 2026•
edited
Loading

ovitrif commented Mar 11, 2026•
edited
Loading

ovitrif commented Mar 13, 2026•
edited
Loading

ovitrif commented Mar 9, 2026•
edited
Loading

coreyphillips commented Mar 11, 2026•
edited
Loading