dohertj2/lmxopcua
1
0
Fork 0
Code Issues 26 Pull Requests 1 Actions 2 Packages Projects Releases Wiki Activity

Upgrade OPC Foundation SDK 1.5.378 → 2.0.x (public SecureChannelContext.PeerAddress replaces #751 reflection) #782

New Issue
Open
opened 2026-11-09 20:23:45 -05:00 by dohertj2 · 7 comments
dohertj2 commented 2026-11-09 20:23:45 -05:00
Owner
Copy Link
Copy Source

OtOpcUa pins OPCFoundation.NetStandard.Opc.Ua.* 1.5.378.106. The 2.0 line is out (2.0.1 is in the local NuGet cache) and is a major-version upgrade.

Concrete driver for this issue: 2.0.1 adds a public Opc.Ua.SecureChannelContext.PeerAddress ("the observed network address of the peer, when the transport exposes one"). That is the API lmxopcua#770 item 4 has been waiting for. Today src/Server/ZB.MOM.WW.OtOpcUa.OpcUaServer/Security/SecureChannelRemoteEndpoint.cs (#751) reads TcpTransportListener.m_channels → channel Socket → IMessageSocket.RemoteEndpoint by reflection. That path is fail-soft and pinned by RemoteEndpointResolutionTests, so a 2.0 upgrade that drops those private members fails a test rather than silently degrading.

Also visible in the 2.0.1 surface: the transport layer was reworked (IUaSCByteTransport, TcpByteTransport, UaScConnectionAdmission). Expect the reflection path to break on upgrade, and expect other breaking changes across Server/Client/Configuration.

Scope:

  • Spike: bump to 2.0.x on a branch and catalogue the compile breaks across Server, Client.*, the OpcUaClient driver and the tests. Is it the 1.5.378 async-API port again, or bigger?
  • Replace SecureChannelRemoteEndpoint with SecureChannelContext.PeerAddress, keeping null = "address unknown" semantics.
  • Re-run OpcUaServer.IntegrationTests and the OpcUaClient wire tests, then live-check ScadaBridge's OPC UA adapter against the upgraded server (wire contract).
  • Coordinate with ScadaBridge, which also uses the OPC Foundation stack in its DCL OPC UA adapter.
OtOpcUa pins `OPCFoundation.NetStandard.Opc.Ua.*` **1.5.378.106**. The 2.0 line is out (2.0.1 is in the local NuGet cache) and is a major-version upgrade. Concrete driver for this issue: 2.0.1 adds a public **`Opc.Ua.SecureChannelContext.PeerAddress`** ("the observed network address of the peer, when the transport exposes one"). That is the API lmxopcua#770 item 4 has been waiting for. Today `src/Server/ZB.MOM.WW.OtOpcUa.OpcUaServer/Security/SecureChannelRemoteEndpoint.cs` (#751) reads `TcpTransportListener.m_channels` → channel `Socket` → `IMessageSocket.RemoteEndpoint` by reflection. That path is fail-soft and pinned by `RemoteEndpointResolutionTests`, so a 2.0 upgrade that drops those private members fails a test rather than silently degrading. Also visible in the 2.0.1 surface: the transport layer was reworked (`IUaSCByteTransport`, `TcpByteTransport`, `UaScConnectionAdmission`). Expect the reflection path to break on upgrade, and expect other breaking changes across Server/Client/Configuration. Scope: - [ ] Spike: bump to 2.0.x on a branch and catalogue the compile breaks across Server, Client.*, the OpcUaClient driver and the tests. Is it the 1.5.378 async-API port again, or bigger? - [ ] Replace `SecureChannelRemoteEndpoint` with `SecureChannelContext.PeerAddress`, keeping null = "address unknown" semantics. - [ ] Re-run OpcUaServer.IntegrationTests and the OpcUaClient wire tests, then live-check ScadaBridge's OPC UA adapter against the upgraded server (wire contract). - [ ] Coordinate with ScadaBridge, which also uses the OPC Foundation stack in its DCL OPC UA adapter.
dohertj2 added the type:debtpriority:p3cross-repo labels 2026-11-09 20:23:45 -05:00
dohertj2 commented 2026-11-09 20:32:52 -05:00
Author
Owner
Copy Link
Copy Source

Spike recorded on master: docs/plans/2026-11-09-opcua-sdk-2.0-upgrade-spike.md (doc only; package bump reverted).

  • Target 2.0.1 (2026-10-12), same package ids. Restore needs Microsoft.Extensions.Logging(.Abstractions) 10.0.12 (we pin 10.0.7).
  • Compile errors: 42 in the first wave, 103 after the trivial fixes. The whole solution is extrapolated to roughly 600–1,200, mostly in tests.
  • Root causes: built-in types become immutable structs (ArrayOf<T>, ByteString, obsolete Variant.Value/DataValue.Value, StatusCode no longer converts to uint), node-manager overrides change signature and go async, the identity/certificate API is reworked (async UserIdentity.CreateAsync, CertificateManager), and SDK-wide nullable annotations trip warnings-as-errors.
  • SecureChannelContext.PeerAddress is an IPAddress (no port), available at ActivateSessionAsync/OperationContext.ChannelContext. It can replace the #751 reflection, but whether the method is virtual and whether the TCP transport populates it are unverified.
  • Estimate: about 2 person-weeks (6–10 days to compile clean with unit tests green, plus 3–5 days for integration, ScadaBridge and live checks). Try the SDK's MigrationAnalyzer fixers first, and consider an interim hop to 1.5.378.182.
  • Risks: stricter session security (needs a live check against ScadaBridge's client), aggregate expected values may move, PKI path changes on deployed boxes, and the Microsoft.Extensions floor across family libs.
    Keeping open as the upgrade tracker; scheduling it is an owner call (p3, nothing forces it).
Spike recorded on master: `docs/plans/2026-11-09-opcua-sdk-2.0-upgrade-spike.md` (doc only; package bump reverted). - Target 2.0.1 (2026-10-12), same package ids. Restore needs Microsoft.Extensions.Logging(.Abstractions) 10.0.12 (we pin 10.0.7). - Compile errors: 42 in the first wave, 103 after the trivial fixes. The whole solution is extrapolated to roughly 600–1,200, mostly in tests. - Root causes: built-in types become immutable structs (`ArrayOf<T>`, `ByteString`, obsolete `Variant.Value`/`DataValue.Value`, `StatusCode` no longer converts to uint), node-manager overrides change signature and go async, the identity/certificate API is reworked (async `UserIdentity.CreateAsync`, `CertificateManager`), and SDK-wide nullable annotations trip warnings-as-errors. - `SecureChannelContext.PeerAddress` is an `IPAddress` (no port), available at `ActivateSessionAsync`/`OperationContext.ChannelContext`. It can replace the #751 reflection, but whether the method is virtual and whether the TCP transport populates it are unverified. - Estimate: about 2 person-weeks (6–10 days to compile clean with unit tests green, plus 3–5 days for integration, ScadaBridge and live checks). Try the SDK's MigrationAnalyzer fixers first, and consider an interim hop to 1.5.378.182. - Risks: stricter session security (needs a live check against ScadaBridge's client), aggregate expected values may move, PKI path changes on deployed boxes, and the Microsoft.Extensions floor across family libs. Keeping open as the upgrade tracker; scheduling it is an owner call (p3, nothing forces it).
dohertj2 commented 2026-11-09 22:13:24 -05:00
Author
Owner
Copy Link
Copy Source

Owner decision 2026-11-09: park the 2.0 upgrade (keep this issue as the tracker; schedule when 1.5 reaches end of support or a feature needs it). Do the patch-level hop 1.5.378.106 → 1.5.378.182 now.

**Owner decision 2026-11-09:** park the 2.0 upgrade (keep this issue as the tracker; schedule when 1.5 reaches end of support or a feature needs it). Do the patch-level hop **1.5.378.106 → 1.5.378.182** now.
dohertj2 commented 2026-11-09 22:48:55 -05:00
Author
Owner
Copy Link
Copy Source

Interim hop done: OPC Foundation SDK 1.5.378.106 -> 1.5.378.182 (latest 1.5.378.x), master 2da58d03. 2.0 upgrade stays parked; this issue stays open as the 2.0 tracker.

Behaviour-neutral adaptations

  • Aggregates: SDK .182 flipped the AggregateManager default TreatUncertainAsBad to true. We keep ours (false; Part 13 default as recalled, not re-read). OtOpcUaSdkServer.CreateAggregateManager now pushes ServerAggregateConfiguration into the SDK manager so live AggregateFilter subscriptions and history reads agree. The #677 parity test is now an explicit deliberate-divergence tripwire (pins ours=false, asserts the SDK default is true so a future change is noticed) plus a booted-server test (mutation-checked).
  • FindPredefinedNode(NodeId, Type) -> FindPredefinedNode<T>; ApplicationConfiguration.TraceConfiguration is obsolete (line removed).
  • Transitive floor: Microsoft.Extensions.Logging.Abstractions 10.0.7 -> 10.0.8 (Opc.Ua.Types .182). No other floor moved.
  • OpcUaClientDriverStallWireTests: every behavioural assertion still holds (same session, one Bad, in-place recovery, no regression). Only the positive control 'SDK late backlog flush arrives' no longer fires; .176 notes 'Preserve publish queue timestamp ordering' and 'Fix publish queue lost wake-up', so the hazard the control guards appears gone SDK-side. The wait is now best-effort (15 s); ordering check kept.

Release notes .106 -> .182 (security/cert/session): .134 UserNameIdentityToken returns BadIdentityTokenInvalid on failed verification, RSA min key size honored, reconnect user-token-policy reuse; .145 certificate-validation StatusCodes filtered; .156 full cert chain sent when CA is in issuer store, issuer/CA KeyUsage (keyCertSign, cRLSign) validated (stricter: a CA cert lacking them would now be rejected), ActivateSession NRE fix; .176 unordered endpoint validation fix, Cancel scoped to current session, ValidateAsync defaults TransportQuotas, MaxNotificationsPerPublish=0 treated as unlimited; .182 bounded array/XML depth/PubSub metadata. Nothing else touching aggregates.

Tests (Release, AKKA_TEST_TIMEFACTOR=3, all 0 failed): OpcUaServer.IntegrationTests 132/132; OpcUaServer.Tests 1027 pass + 1 skip; Client.CLI 168; Client.UI 130; Client.Shared 240; Runtime 988 (+3 skip); Host.Tests 247; Host.IntegrationTests 287 (+2 skip); Driver.OpcUaClient 295 + 5 integration; Browser 19 + 5. Release build 0 warnings. Not deployed.

Interim hop done: OPC Foundation SDK 1.5.378.106 -> 1.5.378.182 (latest 1.5.378.x), master `2da58d03`. 2.0 upgrade stays parked; this issue stays open as the 2.0 tracker. **Behaviour-neutral adaptations** - Aggregates: SDK .182 flipped the AggregateManager default `TreatUncertainAsBad` to true. We keep ours (false; Part 13 default as recalled, not re-read). `OtOpcUaSdkServer.CreateAggregateManager` now pushes `ServerAggregateConfiguration` into the SDK manager so live AggregateFilter subscriptions and history reads agree. The #677 parity test is now an explicit deliberate-divergence tripwire (pins ours=false, asserts the SDK default is true so a future change is noticed) plus a booted-server test (mutation-checked). - `FindPredefinedNode(NodeId, Type)` -> `FindPredefinedNode<T>`; `ApplicationConfiguration.TraceConfiguration` is obsolete (line removed). - Transitive floor: `Microsoft.Extensions.Logging.Abstractions` 10.0.7 -> 10.0.8 (Opc.Ua.Types .182). No other floor moved. - `OpcUaClientDriverStallWireTests`: every behavioural assertion still holds (same session, one Bad, in-place recovery, no regression). Only the positive control 'SDK late backlog flush arrives' no longer fires; .176 notes 'Preserve publish queue timestamp ordering' and 'Fix publish queue lost wake-up', so the hazard the control guards appears gone SDK-side. The wait is now best-effort (15 s); ordering check kept. **Release notes .106 -> .182 (security/cert/session)**: .134 UserNameIdentityToken returns BadIdentityTokenInvalid on failed verification, RSA min key size honored, reconnect user-token-policy reuse; .145 certificate-validation StatusCodes filtered; .156 full cert chain sent when CA is in issuer store, issuer/CA KeyUsage (keyCertSign, cRLSign) validated (stricter: a CA cert lacking them would now be rejected), ActivateSession NRE fix; .176 unordered endpoint validation fix, Cancel scoped to current session, ValidateAsync defaults TransportQuotas, MaxNotificationsPerPublish=0 treated as unlimited; .182 bounded array/XML depth/PubSub metadata. Nothing else touching aggregates. **Tests** (Release, AKKA_TEST_TIMEFACTOR=3, all 0 failed): OpcUaServer.IntegrationTests 132/132; OpcUaServer.Tests 1027 pass + 1 skip; Client.CLI 168; Client.UI 130; Client.Shared 240; Runtime 988 (+3 skip); Host.Tests 247; Host.IntegrationTests 287 (+2 skip); Driver.OpcUaClient 295 + 5 integration; Browser 19 + 5. Release build 0 warnings. Not deployed.
dohertj2 added the blocked label 2026-11-10 08:59:05 -05:00
dohertj2 commented 2026-11-10 08:59:05 -05:00
Author
Owner
Copy Link
Copy Source

Labelled blocked. The owner decided to park the 2.0 upgrade; the hop to 1.5.378.182 is done in 2da58d03.

Labelled `blocked`. The owner decided to park the 2.0 upgrade; the hop to 1.5.378.182 is done in `2da58d03`.
dohertj2 commented 2026-11-10 11:46:44 -05:00
Author
Owner
Copy Link
Copy Source

Execution resumed under the owner’s explicit instruction. Review branch codex/opcua-sdk-2 is pushed (2541ef39); no deployment or release approval.

Implemented exact OPC Foundation SDK 2.0.1, coherent required Microsoft.Extensions uplift, server/client/driver API migration, retained identity/key/ACL behavior, owned certificate lifecycle, single-use session/node-bound history continuation points, filtered alarm/model/audit delivery, and write-conversion/revert compatibility. Aggregate results retain the original 1.5.378.182 algorithms through a private versioned source adapter: 6,144 independently generated baseline vectors pass both directly and through the booted runtime manager, across all 37 factories. No second SDK is loaded in the application.

Latest verification: server unit 1,109 passed / 0 failed / 1 existing skip; full real server wire suite 150/150 passed; trusted mutual-DER secure-channel tests 5/5 passed (Sign and SignAndEncrypt, narrow-grant positive/revocation, invalid credentials and untrusted client); target certificate-manager/endpoint identity tests 7/7 passed. Astra reviewed the scoped application audit permission seam and verified its encrypted wire proof: one application audit reaches the Read audience, zero leak to denied readers, and ordinary SDK automatic audit remains denied. Full solution convergence and other G3/G4 gates remain in progress.

Baseline historian PercentGood parity debt was independently reproduced: original and candidate have the same ten failing tests/messages out of 275; application results and independent expected values are retained. This will be linked as explicit baseline debt, not hidden by changing aggregate policy.

Still pending: complete solution tests, cross-version separate-process matrix, real ScadaBridge/backends, performance and two-hour soak, publish/runtime smoke and final Astra release review. Existing Wonder-box remote-execution hold remains binding. Issue stays open; parked label describes an earlier owner decision superseded by the current execution request.

Execution resumed under the owner’s explicit instruction. Review branch `codex/opcua-sdk-2` is pushed (2541ef39); no deployment or release approval. Implemented exact OPC Foundation SDK 2.0.1, coherent required Microsoft.Extensions uplift, server/client/driver API migration, retained identity/key/ACL behavior, owned certificate lifecycle, single-use session/node-bound history continuation points, filtered alarm/model/audit delivery, and write-conversion/revert compatibility. Aggregate results retain the original 1.5.378.182 algorithms through a private versioned source adapter: 6,144 independently generated baseline vectors pass both directly and through the booted runtime manager, across all 37 factories. No second SDK is loaded in the application. Latest verification: server unit 1,109 passed / 0 failed / 1 existing skip; full real server wire suite 150/150 passed; trusted mutual-DER secure-channel tests 5/5 passed (Sign and SignAndEncrypt, narrow-grant positive/revocation, invalid credentials and untrusted client); target certificate-manager/endpoint identity tests 7/7 passed. Astra reviewed the scoped application audit permission seam and verified its encrypted wire proof: one application audit reaches the Read audience, zero leak to denied readers, and ordinary SDK automatic audit remains denied. Full solution convergence and other G3/G4 gates remain in progress. Baseline historian PercentGood parity debt was independently reproduced: original and candidate have the same ten failing tests/messages out of 275; application results and independent expected values are retained. This will be linked as explicit baseline debt, not hidden by changing aggregate policy. Still pending: complete solution tests, cross-version separate-process matrix, real ScadaBridge/backends, performance and two-hour soak, publish/runtime smoke and final Astra release review. Existing Wonder-box remote-execution hold remains binding. Issue stays open; parked label describes an earlier owner decision superseded by the current execution request.
dohertj2 commented 2026-11-10 13:31:57 -05:00
Author
Owner
Copy Link
Copy Source

Implementation and evidence are pushed on codex/opcua-sdk-2; draft PR #802 is open for review (do not merge). Latest pushed evidence commit b218321b includes the completed CI2 counters and the failed supported-Linux RSS run. New test-only pre-assert/native/JIT/profile diagnostics are being verified before the next push.

Completed local validation: solution Release build0 warnings/errors; full CI2 18,475 passed /11 failed /635 existing skips (18,486 executed, guard passed). All11 failures independently reproduced on original baseline: #796 ten PercentGood oracle cases, #798 ScriptWorker250ms timing assertion. Original application aggregate behavior retained, 6,144 independent original-SDK vectors pass directly and through booted manager. Server wire187, unit1,115+1 existing skip, Shared324, final security17/17. Separate-process original/candidate Shared plus actual driver/CLI matrix62/62. Real SQLServer195 and QuestDB233+1 existing skip pass on each version. Four self-contained Host/CLI publishes; actual Linux Host two clean start/stop cycles preserve PKI and pass HMAC/strict-peer checks.

Actual unchanged ScadaBridge:18 None/API-key cases pass; secure setup fails equally on old/new because its own application certificate is unconfigured (ScadaBridge#121). The narrow opaque scripted alarm grant defect is baseline behavior tracked separately in #797.

Astra independently reviewed integrated source at aa9af0dc: no new must-fix migration source/security finding. It explicitly withheld release approval. #799 now records failed unchanged-cap combined client/server RSS workloads on BOTH macOS (~17min) and Linux (~15min); cause unresolved, no budget waiver or speculative production patch. Native attribution and matched Linux baseline profiling continue. The separate actual production Shared/driver/live-reinitialize/browser two-hour Linux run remains active; no pass is claimed yet.

Windows actual runtime/service gate #800 and approved live AVEVA/consumer/paired-cluster gate #801 remain pending. Wonder-box remote-execution hold is respected; no deployment or master merge performed. #782 remains open; external gate issues do not convert the failed software resource gate into a pass.

Implementation and evidence are pushed on `codex/opcua-sdk-2`; draft PR #802 is open for review (do not merge). Latest pushed evidence commit b218321b includes the completed CI2 counters and the failed supported-Linux RSS run. New test-only pre-assert/native/JIT/profile diagnostics are being verified before the next push. Completed local validation: solution Release build0 warnings/errors; full CI2 18,475 passed /11 failed /635 existing skips (18,486 executed, guard passed). All11 failures independently reproduced on original baseline: #796 ten PercentGood oracle cases, #798 ScriptWorker250ms timing assertion. Original application aggregate behavior retained, 6,144 independent original-SDK vectors pass directly and through booted manager. Server wire187, unit1,115+1 existing skip, Shared324, final security17/17. Separate-process original/candidate Shared plus actual driver/CLI matrix62/62. Real SQLServer195 and QuestDB233+1 existing skip pass on each version. Four self-contained Host/CLI publishes; actual Linux Host two clean start/stop cycles preserve PKI and pass HMAC/strict-peer checks. Actual unchanged ScadaBridge:18 None/API-key cases pass; secure setup fails equally on old/new because its own application certificate is unconfigured (ScadaBridge#121). The narrow opaque scripted alarm grant defect is baseline behavior tracked separately in #797. Astra independently reviewed integrated source at aa9af0dc: no new must-fix migration source/security finding. It explicitly withheld release approval. #799 now records failed unchanged-cap combined client/server RSS workloads on BOTH macOS (~17min) and Linux (~15min); cause unresolved, no budget waiver or speculative production patch. Native attribution and matched Linux baseline profiling continue. The separate actual production Shared/driver/live-reinitialize/browser two-hour Linux run remains active; no pass is claimed yet. Windows actual runtime/service gate #800 and approved live AVEVA/consumer/paired-cluster gate #801 remain pending. Wonder-box remote-execution hold is respected; no deployment or master merge performed. #782 remains open; external gate issues do not convert the failed software resource gate into a pass.
dohertj2 commented 2026-11-10 16:10:06 -05:00
Author
Owner
Copy Link
Copy Source

Execution update: candidate and complete evidence pushed through ed60970f, draft PR #802 remains HOLD. Final integrated build has zero warnings/errors. CI4 executes 18,489 tests: 18,478 passed / 11 independently reproduced baseline failures / 635 existing skips, test1/guard0. The extra CI3 fixture/test-ordering failures are resolved and pass in this same final run; original failed attempts are retained.

Completed two-hour actual production client/driver/browser lifecycle: 5,420 cycles, 43,360 verified manager disposals, zero weak survivors/PKI descriptors. Fresh HG1.0 four publishes and actual Linux two-start/stop smoke passed. Historiangw#131 is closed on consumer master230cef4e; baseline test repair #805 is closed on master52d1373e; both merged into candidate. No SDK2 master merge/deployment.

Astra source/security review closed with no new must-fix finding. Release remains on HOLD: #799 cold RSS failed, despite the bounded canonical factory-saturation diagnostic meeting its selected criterion; the original adapter honestly records partial proof and unsupported ownership/process-memory parity. No warmed acceptance baseline, cap waiver or production workaround was approved. #800/#801/ScadaBridge#121 remain external gates; #796/#798 baseline failures and #797/#806 baseline functionality gaps remain linked.

Execution evidence, operator runbook. The scadaproj main index/history was updated and pushed with actual consumer versions and gate limits. #782 remains OPEN; software/release completion is not claimed.

Execution update: candidate and complete evidence pushed through **ed60970f**, draft PR #802 remains HOLD. Final integrated build has zero warnings/errors. CI4 executes 18,489 tests: **18,478 passed / 11 independently reproduced baseline failures / 635 existing skips**, test1/guard0. The extra CI3 fixture/test-ordering failures are resolved and pass in this same final run; original failed attempts are retained. Completed two-hour actual production client/driver/browser lifecycle: 5,420 cycles, 43,360 verified manager disposals, zero weak survivors/PKI descriptors. Fresh HG1.0 four publishes and actual Linux two-start/stop smoke passed. Historiangw#131 is closed on consumer master230cef4e; baseline test repair #805 is closed on master52d1373e; both merged into candidate. No SDK2 master merge/deployment. Astra source/security review closed with no new must-fix finding. **Release remains on HOLD:** #799 cold RSS failed, despite the bounded canonical factory-saturation diagnostic meeting its selected criterion; the original adapter honestly records partial proof and unsupported ownership/process-memory parity. No warmed acceptance baseline, cap waiver or production workaround was approved. #800/#801/ScadaBridge#121 remain external gates; #796/#798 baseline failures and #797/#806 baseline functionality gaps remain linked. [Execution evidence](https://gitea.dohertylan.com/dohertj2/lmxopcua/src/commit/ed60970f/docs/plans/2026-11-10-opcua-sdk-2-evidence.md), [operator runbook](https://gitea.dohertylan.com/dohertj2/lmxopcua/src/commit/ed60970f/docs/plans/2026-11-10-opcua-sdk-2-rollout.md). The scadaproj main index/history was updated and pushed with actual consumer versions and gate limits. #782 remains OPEN; software/release completion is not claimed.
Sign in to join this conversation.
No Branch/Tag Specified
Branches Tags
Labels
Clear labels
auto-managed
queue management
 blocked
Blocked by dependency or external hold
 cross-driver
queue management
 cross-repo
Needs coordinated change in another family repo (wire contract etc.)
 driver/abcip
queue management
 driver/ablegacy
queue management
 driver/focas
queue management
 driver/mqtt
MQTT / Sparkplug B broker-subscribe driver
 driver/opcuaclient
queue management
 driver/s7
queue management
 driver/twincat
queue management
 migrated
Migrated from a markdown backlog (2026-11 standardization)
 phase/1
queue management
 phase/2
queue management
 phase/3
queue management
 phase/4
queue management
 phase/5
queue management
 phase/6
queue management
 priority:p0
Must-have / blocking
 priority:p1
High priority
 priority:p2
Normal priority
 priority:p3
Low priority / nice to have
 queue/blocked
queue management
 queue/done
queue management
 queue/failed
queue management
 queue/in-progress
queue management
 queue/queued
queue management
 type:bug
Defect in shipped behaviour
 type:debt
Technical debt / hygiene
 type:deploy-gate
Rollout/deploy/migration step not yet done on an environment
 type:docs
Documentation work
 type:feature
Feature implementation
 type:followup
Follow-up deferred from a plan, PR or gate
 type:infra
Infrastructure and scaffolding
 type:review-finding
Unfixed finding from an architecture/code review
 type:test
Test work
No Label blocked cross-repo priority:p3 type:debt
Milestone
No items
No Milestone
Projects
Clear projects
No project
Assignees
Clear assignees
dohertj2 (dohertj2)
No Assignees
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: dohertj2/lmxopcua#782
Delete Branch "%!s()"

Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?