[dependabot]

Updated AndreGoepel.Core from 1.0.4 to 1.0.6.

Release notes

Sourced from AndreGoepel.Core's releases.

1.0.6

Publication complete: Core 1.0.6 and its symbols are published and independently verified. The package payload and NuGet repository signature match the original build; the symbol archive and DLL-bound PDB match their authenticated original hashes. The controlled symbol-only recovery succeeded on its first attempt, without re-uploading the main package or changing the original tag. Acceptance evidence is recorded in Core #​29.

Fixes

• Result.Fail(string error) and Result.Fail<T>(string error) now reject null errors with ArgumentNullException (ParamName: error), preventing an invalid failure result without an error message.
• Two regression tests cover this behavior. Non-null calls, public signatures, the .NET 10 target and the zero-dependency package remain unchanged.

Compatibility: callers that pass null despite the non-nullable parameter contract now receive an exception. Supply a meaningful error string instead.

Publication

This release uses the protected Sonar-gated preparation/publication path, with official NuGet/login and native dotnet nuget push. The legacy CI publisher has been retired. Production publication and remote package/symbol verification are tracked in Core #​29.

Source commit: 06e816a983d27e4d215042acc47e63f1ddba2075.

What's Changed

• feat: prepare opt-in shared Sonar analysis for core by @​andregoepel in feat: prepare opt-in shared Sonar analysis for core core#25
• chore: configure Core central Sonar policy and merged runner pin by @​andregoepel in chore: configure Core central Sonar policy and merged runner pin core#26
• fix: forward protected Sonar secrets to shared workflow by @​andregoepel in fix: forward protected Sonar secrets to shared workflow core#27
• chore: record the approved Core Sonar baseline by @​andregoepel in chore: record the approved Core Sonar baseline core#28
• Prepare Sonar-gated publication and new-code coverage evidence by @​andregoepel in Prepare Sonar-gated publication and new-code coverage evidence core#31
• Register disabled publication workflows and record new-code acceptance by @​andregoepel in Register disabled publication workflows and record new-code acceptance core#32
• chore: register disabled NuGet workflow identity by @​andregoepel in chore: register disabled NuGet workflow identity core#33
• test: add no-upload publication rehearsal by @​andregoepel in test: add no-upload publication rehearsal core#34
• feat: prepare tagged NuGet acceptance and retire the legacy publisher by @​andregoepel in feat: prepare tagged NuGet acceptance and retire the legacy publisher core#35
• feat: prepare production NuGet tag trigger by @​andregoepel in feat: prepare production NuGet tag trigger core#36
• feat: activate gated NuGet preparation and publishing by @​andregoepel in feat: activate gated NuGet preparation and publishing core#37

Full Changelog: andregoepel/core@v1.0.5...v1.0.6

1.0.5

What's Changed

• chore: update coverlet.collector to 10.1.0 by @​andregoepel in chore: update coverlet.collector to 10.1.0 core#20
• chore: update dependencies by @​andregoepel in chore: update dependencies core#23

Full Changelog: andregoepel/core@v1.0.4...v1.0.5

Commits viewable in compare view.

Updated AndreGoepel.Design.Blazor from 1.7.1 to 1.7.2.

Release notes

Sourced from AndreGoepel.Design.Blazor's releases.

1.7.2

What's Changed

• chore: update dependencies by @​andregoepel in chore: update dependencies design-blazor#132

Full Changelog: andregoepel/design-blazor@v1.7.1...v1.7.2

Commits viewable in compare view.

Updated AndreGoepel.Marten.Configuration from 1.2.6 to 1.2.7.

Release notes

Sourced from AndreGoepel.Marten.Configuration's releases.

1.2.7

What's Changed

• chore: propagate shared container image updates by @​andregoepel in chore: propagate shared container image updates marten-configuration#55
• chore: update dependencies by @​andregoepel in chore: update dependencies marten-configuration#61

Full Changelog: andregoepel/marten-configuration@v1.2.6...v1.2.7

Commits viewable in compare view.

Updated AndreGoepel.Marten.Identity.Blazor from 2.2.0 to 2.2.1.

Release notes

Sourced from AndreGoepel.Marten.Identity.Blazor's releases.

2.2.1

What's Changed

• chore: pin current database images in the identity sample by @​andregoepel in chore: pin current database images in the identity sample marten-identity#217
• chore: propagate shared container image updates by @​andregoepel in chore: propagate shared container image updates marten-identity#219
• chore: update dependencies by @​andregoepel in chore: update dependencies marten-identity#225

Full Changelog: andregoepel/marten-identity@v2.2.0...v2.2.1

Commits viewable in compare view.

Updated AndreGoepel.Marten.Testing from 1.2.6 to 1.2.7.

Release notes

Sourced from AndreGoepel.Marten.Testing's releases.

1.2.7

What's Changed

• chore: update dependencies by @​andregoepel in chore: update dependencies marten-testing#37

Full Changelog: andregoepel/marten-testing@v1.2.6...v1.2.7

Commits viewable in compare view.

Updated AndreGoepel.Testing.Bunit from 1.3.4 to 1.3.5.

Release notes

Sourced from AndreGoepel.Testing.Bunit's releases.

1.3.5

What's Changed

• chore: update dependencies by @​andregoepel in chore: update dependencies testing#32

Full Changelog: andregoepel/testing@v1.3.4...v1.3.5

Commits viewable in compare view.

Updated AndreGoepel.Testing.E2E from 1.3.4 to 1.3.5.

Release notes

Sourced from AndreGoepel.Testing.E2E's releases.

1.3.5

What's Changed

• chore: update dependencies by @​andregoepel in chore: update dependencies testing#32

Full Changelog: andregoepel/testing@v1.3.4...v1.3.5

Commits viewable in compare view.

Updated Aspire.Hosting from 13.6.0 to 13.6.4.

Release notes

Sourced from Aspire.Hosting's releases.

13.6.4

What's New in Aspire 13.6.4

Patch release for Aspire 13.6 that improves resource startup reliability, fixes Azure publishing with parameterized enum properties in TypeScript AppHosts, and tightens CLI diagnostic log permissions on Unix.

🐛 Fixes

• 🖥️ Resources could fail to start after a DCP create request timed out — AppHost now checks whether a resource was created before retrying a timed-out or HTTP 504 request, avoiding a premature FailedToStart state when DCP completed the operation. API operations default to a 40-second timeout with an additional 20-second allowance before the first successful operation, and resource creation has a two-minute recovery budget. These limits can be configured through DcpPublisher:KubernetesApiTimeout, DcpPublisher:KubernetesInitializationAdditionalTimeout, and DcpPublisher:KubernetesCreateRecoveryTimeout. Established watch and log streams are not limited by these operation timeouts. Addresses startup timeouts reported in 13.6. Fixes #​20691. (#​20868, backport of #​20752, @​danegsta)

• ☁️ TypeScript AppHosts could fail to publish parameterized Azure enum properties — Assigning a bicep().parameter(...) expression to an enum-valued provisioning property, such as an Azure Storage SKU name, now works instead of failing with a type-assignment error. Type checking for literal values and incompatible expression types is preserved. Fixes a scenario introduced with the Azure provisioning proxy feature in 13.6. (#​20767, backport of #​20557, @​sebastienros)

• 🔒 CLI diagnostic logs now use owner-only permissions on Unix — Log files are restricted to owner read/write access (0600), including reused files, and the dedicated Aspire logs directory uses owner-only access (0700), repairing existing permissions when needed. Newly created custom log directories also use 0700; existing custom parent directories and Windows permission behavior are unchanged. Diagnostic logs can contain sensitive data, including dashboard login URLs, so review them before sharing. (#​21282, backport of #​21276, @​mitchdenny)

---

Full Changelog: v13.6.3...v13.6.4

Full commit: c89352703f78ae46de90fbf751ed8b52b77b89bf

13.6.3

What's New in Aspire 13.6.3

Patch release for Aspire 13.6 that improves resource orchestration reliability by preventing unread resource watches from stalling resource updates and operations.

🐛 Fixes

• 🖥️ Resource operations could stall when a watch stopped being read — AppHost now closes abandoned watch connections when setup is canceled and keeps reading resource events even when event handlers are slow. Resource event ordering, lifecycle callback dispatch, and the existing periodic reconnection safeguard are preserved. (#​21002, backport of #​20972, @​danegsta)

• 🔄 Slow watchers could block other resource updates or watcher shutdown — The bundled Developer Control Plane (DCP) is updated to 0.26.6, which keeps internal lifecycle watches draining and uses bounded, asynchronous watch delivery to prevent unread watches from blocking resource-change publication and unrelated watcher shutdown. (#​21215, @​danegsta; includes microsoft/dcp#​307)

---

Full Changelog: v13.6.2...v13.6.3

Full commit: 38294b1b31552dae42a516cffa0cbc737ec08976

13.6.2

Full commit: af07e33f8d1ef733645138b3f8e36067d073148d

What's Changed

• [release/13.6] Bump version to 13.6.2 by @​danegsta with @​Copilot in [release/13.6] Bump version to 13.6.2 microsoft/aspire#20877
• [release/13.6] Pin MessagePack to 2.5.307 for security fixes by @​eerhardt in [release/13.6] Pin MessagePack to 2.5.307 for security fixes microsoft/aspire#21156

Full Changelog: microsoft/aspire@v13.6.1...v13.6.2

13.6.1

What's New in Aspire 13.6.1

Patch release for Aspire 13.6 that reduces dashboard CPU usage during sustained metric ingestion, restores scrolling in the Metrics tree, improves dashboard navigation, fixes container exec watch failures, and ensures session-scoped containers are cleaned up when stopping .NET AppHosts on Windows.

🐛 Fixes

• 📊 Dashboard CPU usage could increase during long-running metric ingestion — Metric retention repeatedly scanned retained history on every insertion, causing steadily increasing CPU usage even with a fixed workload. Cleanup now skips dimensions below capacity and removes only the oldest surplus points when needed. Existing retention limits and exemplar behavior are unchanged, and no configuration changes are required. Regression introduced in 13.6. Fixes #​20725. (#​20747, backport of #​20736, @​JamesNK)

• 📈 The Metrics tree could not be scrolled — The meter and instrument tree now scrolls when it exceeds the viewport, allowing you to reach and select instruments below the fold. Regression introduced in 13.6. Fixes #​20728. (#​20762, backport of #​20758, @​heyysiri)

• 🧭 Dashboard navigation could fail after a browser disconnect — Pending page-state updates now persist session state before navigating and skip navigation when the browser has disconnected, avoiding an unhandled navigation failure. The navigation rail's overflow button also has consistent sizing, icon alignment, and hover styling in expanded and collapsed layouts and both light and dark themes. (#​20545, backport of #​20527, @​JamesNK)

• 🖥️ Container exec watches could terminate shortly after AppHost startup — A stalled DCP watch connection could produce a critical Watch task over Kubernetes ContainerExec resources terminated unexpectedly error and stop subsequent terminal state and log updates from being observed. Watch connection attempts now retry after timeouts, and periodic restarts correctly recreate connections without swallowing unrelated cancellation. Regression introduced in 13.6.0. (#​20664, backport of #​20466, @​radical)

• 🪟 Session-scoped containers could remain running after stopping a .NET AppHost on Windows — AppHosts launched through dotnet run or dotnet watch now allow DCP to finish resource cleanup during shutdown, including existing Ctrl+C and aspire stop flows. Persistent resources remain unaffected, while direct .NET launches and TypeScript/polyglot AppHosts retain their existing process protection. Fixes #​20495. (#​20514, backport of #​20494, @​danegsta)

---

Full Changelog: v13.6.0...v13.6.1

Full commit: 3751e615ad660621f2a7aa4c390ef48e08a0eed8

Commits viewable in compare view.

Updated Aspire.Hosting.AppHost from 13.6.0 to 13.6.4.

Release notes

Sourced from Aspire.Hosting.AppHost's releases.

13.6.4

What's New in Aspire 13.6.4

Patch release for Aspire 13.6 that improves resource startup reliability, fixes Azure publishing with parameterized enum properties in TypeScript AppHosts, and tightens CLI diagnostic log permissions on Unix.

🐛 Fixes

• 🖥️ Resources could fail to start after a DCP create request timed out — AppHost now checks whether a resource was created before retrying a timed-out or HTTP 504 request, avoiding a premature FailedToStart state when DCP completed the operation. API operations default to a 40-second timeout with an additional 20-second allowance before the first successful operation, and resource creation has a two-minute recovery budget. These limits can be configured through DcpPublisher:KubernetesApiTimeout, DcpPublisher:KubernetesInitializationAdditionalTimeout, and DcpPublisher:KubernetesCreateRecoveryTimeout. Established watch and log streams are not limited by these operation timeouts. Addresses startup timeouts reported in 13.6. Fixes #​20691. (#​20868, backport of #​20752, @​danegsta)

• ☁️ TypeScript AppHosts could fail to publish parameterized Azure enum properties — Assigning a bicep().parameter(...) expression to an enum-valued provisioning property, such as an Azure Storage SKU name, now works instead of failing with a type-assignment error. Type checking for literal values and incompatible expression types is preserved. Fixes a scenario introduced with the Azure provisioning proxy feature in 13.6. (#​20767, backport of #​20557, @​sebastienros)

• 🔒 CLI diagnostic logs now use owner-only permissions on Unix — Log files are restricted to owner read/write access (0600), including reused files, and the dedicated Aspire logs directory uses owner-only access (0700), repairing existing permissions when needed. Newly created custom log directories also use 0700; existing custom parent directories and Windows permission behavior are unchanged. Diagnostic logs can contain sensitive data, including dashboard login URLs, so review them before sharing. (#​21282, backport of #​21276, @​mitchdenny)

---

Full Changelog: v13.6.3...v13.6.4

Full commit: c89352703f78ae46de90fbf751ed8b52b77b89bf

13.6.3

What's New in Aspire 13.6.3

Patch release for Aspire 13.6 that improves resource orchestration reliability by preventing unread resource watches from stalling resource updates and operations.

🐛 Fixes

• 🖥️ Resource operations could stall when a watch stopped being read — AppHost now closes abandoned watch connections when setup is canceled and keeps reading resource events even when event handlers are slow. Resource event ordering, lifecycle callback dispatch, and the existing periodic reconnection safeguard are preserved. (#​21002, backport of #​20972, @​danegsta)

• 🔄 Slow watchers could block other resource updates or watcher shutdown — The bundled Developer Control Plane (DCP) is updated to 0.26.6, which keeps internal lifecycle watches draining and uses bounded, asynchronous watch delivery to prevent unread watches from blocking resource-change publication and unrelated watcher shutdown. (#​21215, @​danegsta; includes microsoft/dcp#​307)

---

Full Changelog: v13.6.2...v13.6.3

Full commit: 38294b1b31552dae42a516cffa0cbc737ec08976

13.6.2

Full commit: af07e33f8d1ef733645138b3f8e36067d073148d

What's Changed

• [release/13.6] Bump version to 13.6.2 by @​danegsta with @​Copilot in [release/13.6] Bump version to 13.6.2 microsoft/aspire#20877
• [release/13.6] Pin MessagePack to 2.5.307 for security fixes by @​eerhardt in [release/13.6] Pin MessagePack to 2.5.307 for security fixes microsoft/aspire#21156

Full Changelog: microsoft/aspire@v13.6.1...v13.6.2

13.6.1

What's New in Aspire 13.6.1

Patch release for Aspire 13.6 that reduces dashboard CPU usage during sustained metric ingestion, restores scrolling in the Metrics tree, improves dashboard navigation, fixes container exec watch failures, and ensures session-scoped containers are cleaned up when stopping .NET AppHosts on Windows.

🐛 Fixes

• 📊 Dashboard CPU usage could increase during long-running metric ingestion — Metric retention repeatedly scanned retained history on every insertion, causing steadily increasing CPU usage even with a fixed workload. Cleanup now skips dimensions below capacity and removes only the oldest surplus points when needed. Existing retention limits and exemplar behavior are unchanged, and no configuration changes are required. Regression introduced in 13.6. Fixes #​20725. (#​20747, backport of #​20736, @​JamesNK)

• 📈 The Metrics tree could not be scrolled — The meter and instrument tree now scrolls when it exceeds the viewport, allowing you to reach and select instruments below the fold. Regression introduced in 13.6. Fixes #​20728. (#​20762, backport of #​20758, @​heyysiri)

• 🧭 Dashboard navigation could fail after a browser disconnect — Pending page-state updates now persist session state before navigating and skip navigation when the browser has disconnected, avoiding an unhandled navigation failure. The navigation rail's overflow button also has consistent sizing, icon alignment, and hover styling in expanded and collapsed layouts and both light and dark themes. (#​20545, backport of #​20527, @​JamesNK)

• 🖥️ Container exec watches could terminate shortly after AppHost startup — A stalled DCP watch connection could produce a critical Watch task over Kubernetes ContainerExec resources terminated unexpectedly error and stop subsequent terminal state and log updates from being observed. Watch connection attempts now retry after timeouts, and periodic restarts correctly recreate connections without swallowing unrelated cancellation. Regression introduced in 13.6.0. (#​20664, backport of #​20466, @​radical)

• 🪟 Session-scoped containers could remain running after stopping a .NET AppHost on Windows — AppHosts launched through dotnet run or dotnet watch now allow DCP to finish resource cleanup during shutdown, including existing Ctrl+C and aspire stop flows. Persistent resources remain unaffected, while direct .NET launches and TypeScript/polyglot AppHosts retain their existing process protection. Fixes #​20495. (#​20514, backport of #​20494, @​danegsta)

---

Full Changelog: v13.6.0...v13.6.1

Full commit: 3751e615ad660621f2a7aa4c390ef48e08a0eed8

Commits viewable in compare view.

Updated Aspire.Hosting.PostgreSQL from 13.6.0 to 13.6.4.

Release notes

Sourced from Aspire.Hosting.PostgreSQL's releases.

13.6.4

What's New in Aspire 13.6.4

Patch release for Aspire 13.6 that improves resource startup reliability, fixes Azure publishing with parameterized enum properties in TypeScript AppHosts, and tightens CLI diagnostic log permissions on Unix.

🐛 Fixes

• 🖥️ Resources could fail to start after a DCP create request timed out — AppHost now checks whether a resource was created before retrying a timed-out or HTTP 504 request, avoiding a premature FailedToStart state when DCP completed the operation. API operations default to a 40-second timeout with an additional 20-second allowance before the first successful operation, and resource creation has a two-minute recovery budget. These limits can be configured through DcpPublisher:KubernetesApiTimeout, DcpPublisher:KubernetesInitializationAdditionalTimeout, and DcpPublisher:KubernetesCreateRecoveryTimeout. Established watch and log streams are not limited by these operation timeouts. Addresses startup timeouts reported in 13.6. Fixes #​20691. (#​20868, backport of #​20752, @​danegsta)

• ☁️ TypeScript AppHosts could fail to publish parameterized Azure enum properties — Assigning a bicep().parameter(...) expression to an enum-valued provisioning property, such as an Azure Storage SKU name, now works instead of failing with a type-assignment error. Type checking for literal values and incompatible expression types is preserved. Fixes a scenario introduced with the Azure provisioning proxy feature in 13.6. (#​20767, backport of #​20557, @​sebastienros)

• 🔒 CLI diagnostic logs now use owner-only permissions on Unix — Log files are restricted to owner read/write access (0600), including reused files, and the dedicated Aspire logs directory uses owner-only access (0700), repairing existing permissions when needed. Newly created custom log directories also use 0700; existing custom parent directories and Windows permission behavior are unchanged. Diagnostic logs can contain sensitive data, including dashboard login URLs, so review them before sharing. (#​21282, backport of #​21276, @​mitchdenny)

---

Full Changelog: v13.6.3...v13.6.4

Full commit: c89352703f78ae46de90fbf751ed8b52b77b89bf

13.6.3

What's New in Aspire 13.6.3

Patch release for Aspire 13.6 that improves resource orchestration reliability by preventing unread resource watches from stalling resource updates and operations.

🐛 Fixes

• 🖥️ Resource operations could stall when a watch stopped being read — AppHost now closes abandoned watch connections when setup is canceled and keeps reading resource events even when event handlers are slow. Resource event ordering, lifecycle callback dispatch, and the existing periodic reconnection safeguard are preserved. (#​21002, backport of #​20972, @​danegsta)

• 🔄 Slow watchers could block other resource updates or watcher shutdown — The bundled Developer Control Plane (DCP) is updated to 0.26.6, which keeps internal lifecycle watches draining and uses bounded, asynchronous watch delivery to prevent unread watches from blocking resource-change publication and unrelated watcher shutdown. (#​21215, @​danegsta; includes microsoft/dcp#​307)

---

Full Changelog: v13.6.2...v13.6.3

Full commit: 38294b1b31552dae42a516cffa0cbc737ec08976

13.6.2

Full commit: af07e33f8d1ef733645138b3f8e36067d073148d

What's Changed

• [release/13.6] Bump version to 13.6.2 by @​danegsta with @​Copilot in [release/13.6] Bump version to 13.6.2 microsoft/aspire#20877
• [release/13.6] Pin MessagePack to 2.5.307 for security fixes by @​eerhardt in [release/13.6] Pin MessagePack to 2.5.307 for security fixes microsoft/aspire#21156

Full Changelog: microsoft/aspire@v13.6.1...v13.6.2

13.6.1

What's New in Aspire 13.6.1

Patch release for Aspire 13.6 that reduces dashboard CPU usage during sustained metric ingestion, restores scrolling in the Metrics tree, improves dashboard navigation, fixes container exec watch failures, and ensures session-scoped containers are cleaned up when stopping .NET AppHosts on Windows.

🐛 Fixes

• 📊 Dashboard CPU usage could increase during long-running metric ingestion — Metric retention repeatedly scanned retained history on every insertion, causing steadily increasing CPU usage even with a fixed workload. Cleanup now skips dimensions below capacity and removes only the oldest surplus points when needed. Existing retention limits and exemplar behavior are unchanged, and no configuration changes are required. Regression introduced in 13.6. Fixes #​20725. (#​20747, backport of #​20736, @​JamesNK)

• 📈 The Metrics tree could not be scrolled — The meter and instrument tree now scrolls when it exceeds the viewport, allowing you to reach and select instruments below the fold. Regression introduced in 13.6. Fixes #​20728. (#​20762, backport of #​20758, @​heyysiri)

• 🧭 Dashboard navigation could fail after a browser disconnect — Pending page-state updates now persist session state before navigating and skip navigation when the browser has disconnected, avoiding an unhandled navigation failure. The navigation rail's overflow button also has consistent sizing, icon alignment, and hover styling in expanded and collapsed layouts and both light and dark themes. (#​20545, backport of #​20527, @​JamesNK)

• 🖥️ Container exec watches could terminate shortly after AppHost startup — A stalled DCP watch connection could produce a critical Watch task over Kubernetes ContainerExec resources terminated unexpectedly error and stop subsequent terminal state and log updates from being observed. Watch connection attempts now retry after timeouts, and periodic restarts correctly recreate connections without swallowing unrelated cancellation. Regression introduced in 13.6.0. (#​20664, backport of #​20466, @​radical)

• 🪟 Session-scoped containers could remain running after stopping a .NET AppHost on Windows — AppHosts launched through dotnet run or dotnet watch now allow DCP to finish resource cleanup during shutdown, including existing Ctrl+C and aspire stop flows. Persistent resources remain unaffected, while direct .NET launches and TypeScript/polyglot AppHosts retain their existing process protection. Fixes #​20495. (#​20514, backport of #​20494, @​danegsta)

---

Full Changelog: v13.6.0...v13.6.1

Full commit: 3751e615ad660621f2a7aa4c390ef48e08a0eed8

Commits viewable in compare view.

Updated Aspire.Hosting.Testing from 13.6.0 to 13.6.1.

Release notes

Sourced from Aspire.Hosting.Testing's releases.

13.6.1

What's New in Aspire 13.6.1

Patch release for Aspire 13.6 that reduces dashboard CPU usage during sustained metric ingestion, restores scrolling in the Metrics tree, improves dashboard navigation, fixes container exec watch failures, and ensures session-scoped containers are cleaned up when stopping .NET AppHosts on Windows.

🐛 Fixes

• 📊 Dashboard CPU usage could increase during long-running metric ingestion — Metric retention repeatedly scanned retained history on every insertion, causing steadily increasing CPU usage even with a fixed workload. Cleanup now skips dimensions below capacity and removes only the oldest surplus points when needed. Existing retention limits and exemplar behavior are unchanged, and no configuration changes are required. Regression introduced in 13.6. Fixes #​20725. (#​20747, backport of #​20736, @​JamesNK)

• 📈 The Metrics tree could not be scrolled — The meter and instrument tree now scrolls when it exceeds the viewport, allowing you to reach and select instruments below the fold. Regression introduced in 13.6. Fixes #​20728. (#​20762, backport of #​20758, @​heyysiri)

• 🧭 Dashboard navigation could fail after a browser disconnect — Pending page-state updates now persist session state before navigating and skip navigation when the browser has disconnected, avoiding an unhandled navigation failure. The navigation rail's overflow button also has consistent sizing, icon alignment, and hover styling in expanded and collapsed layouts and both light and dark themes. (#​20545, backport of #​20527, @​JamesNK)

• 🖥️ Container exec watches could terminate shortly after AppHost startup — A stalled DCP watch connection could produce a critical Watch task over Kubernetes ContainerExec resources terminated unexpectedly error and stop subsequent terminal state and log updates from being observed. Watch connection attempts now retry after timeouts, and periodic restarts correctly recreate connections without swallowing unrelated cancellation. Regression introduced in 13.6.0. (#​20664, backport of #​20466, @​radical)

• 🪟 Session-scoped containers could remain running after stopping a .NET AppHost on Windows — AppHosts launched through dotnet run or dotnet watch now allow DCP to finish resource cleanup during shutdown, including existing Ctrl+C and aspire stop flows. Persistent resources remain unaffected, while direct .NET launches and TypeScript/polyglot AppHosts retain their existing process protection. Fixes #​20495. (#​20514, backport of #​20494, @​danegsta)

---

Full Changelog: v13.6.0...v13.6.1

Full commit: 3751e615ad660621f2a7aa4c390ef48e08a0eed8

Commits viewable in compare view.

Updated Aspire.Hosting.Testing from 13.6.0 to 13.6.4.

Release notes

Sourced from Aspire.Hosting.Testing's releases.

13.6.4

What's New in Aspire 13.6.4

Patch release for Aspire 13.6 that improves resource startup reliability, fixes Azure publishing with parameterized enum properties in TypeScript AppHosts, and tightens CLI diagnostic log permissions on Unix.

🐛 Fixes

• 🖥️ Resources could fail to start after a DCP create request timed out — AppHost now checks whether a resource was created before retrying a timed-out or HTTP 504 request, avoiding a premature FailedToStart state when DCP completed the operation. API operations default to a 40-second timeout with an additional 20-second allowance before the first successful operation, and resource creation has a two-minute recovery budget. These limits can be configured through DcpPublisher:KubernetesApiTimeout, DcpPublisher:KubernetesInitializationAdditionalTimeout, and DcpPublisher:KubernetesCreateRecoveryTimeout. Established watch and log streams are not limited by these operation timeouts. Addresses startup timeouts reported in 13.6. Fixes #​20691. (#​20868, backport of #​20752, @​danegsta)

• ☁️ TypeScript AppHosts could fail to publish parameterized Azure enum properties — Assigning a bicep().parameter(...) expression to an enum-valued provisioning property, such as an Azure Storage SKU name, now works instead of failing with a type-assignment error. Type checking for literal values and incompatible expression types is preserved. Fixes a scenario introduced with the Azure provisioning proxy feature in 13.6. (#​20767, backport of #​20557, @​sebastienros)

• 🔒 CLI diagnostic logs now use owner-only permissions on Unix — Log files are restricted to owner read/write access (0600), including reused files, and the dedicated Aspire logs directory uses owner-only access (0700), repairing existing permissions when needed. Newly created custom log directories also use 0700; existing custom parent directories and Windows permission behavior are unchanged. Diagnostic logs can contain sensitive data, including dashboard login URLs, so review them before sharing. (#​21282, backport of #​21276, @​mitchdenny)

---

Full Changelog: v13.6.3...v13.6.4

Full commit: c89352703f78ae46de90fbf751ed8b52b77b89bf

13.6.3

What's New in Aspire 13.6.3

Patch release for Aspire 13.6 that improves resource orchestration reliability by preventing unread resource watches from stalling resource updates and operations.

🐛 Fixes

• 🖥️ Resource operations could stall when a watch stopped being read — AppHost now closes abandoned watch connections when setup is canceled and keeps reading resource events even when event handlers are slow. Resource event ordering, lifecycle callback dispatch, and the existing periodic reconnection safeguard are preserved. (#​21002, backport of #​20972, @​danegsta)

• 🔄 Slow watchers could block other resource updates or watcher shutdown — The bundled Developer Control Plane (DCP) is updated to 0.26.6, which keeps internal lifecycle watches draining and uses bounded, asynchronous watch delivery to prevent unread watches from blocking resource-change publication and unrelated watcher shutdown. (#​21215, @​danegsta; includes microsoft/dcp#​307)

---

Full Changelog: v13.6.2...v13.6.3

Full commit: 38294b1b31552dae42a516cffa0cbc737ec08976

13.6.2

Full commit: af07e33f8d1ef733645138b3f8e36067d073148d

What's Changed

• [release/13.6] Bump version to 13.6.2 by @​danegsta with @​Copilot in [release/13.6] Bump version to 13.6.2 microsoft/aspire#20877
• [release/13.6] Pin MessagePack to 2.5.307 for security fixes by @​eerhardt in [release/13.6] Pin MessagePack to 2.5.307 for security fixes microsoft/aspire#21156

Full Changelog: microsoft/aspire@v13.6.1...v13.6.2

13.6.1

What's New in Aspire 13.6.1

Patch release for Aspire 13.6 that reduces dashboard CPU usage during sustained metric ingestion, restores scrolling in the Metrics tree, improves dashboard navigation, fixes container exec watch failures, and ensures session-scoped containers are cleaned up when stopping .NET AppHosts on Windows.

🐛 Fixes

• 📊 Dashboard CPU usage could increase during long-running metric ingestion — Metric retention repeatedly scanned retained history on every insertion, causing steadily increasing CPU usage even with a fixed workload. Cleanup now skips dimensions below capacity and removes only the oldest surplus points when needed. Existing retention limits and exemplar behavior are unchanged, and no configuration changes are required. Regression introduced in 13.6. Fixes #​20725. (#​20747, backport of #​20736, @​JamesNK)

• 📈 The Metrics tree could not be scrolled — The meter and instrument tree now scrolls when it exceeds the viewport, allowing you to reach and select instruments below the fold. Regression introduced in 13.6. Fixes #​20728. (#​20762, backport of #​20758, @​heyysiri)

• 🧭 Dashboard navigation could fail after a browser disconnect — Pending page-state updates now persist session state before navigating and skip navigation when the browser has disconnected, avoiding an unhandled navigation failure. The navigation rail's overflow button also has consistent sizing, icon alignment, and hover styling in expanded and collapsed layouts and both light and dark themes. (#​20545, backport of #​20527, @​JamesNK)

• 🖥️ Container exec watches could terminate shortly after AppHost startup — A stalled DCP watch connection could produce a critical Watch task over Kubernetes ContainerExec resources terminated unexpectedly error and stop subsequent terminal state and log updates from being observed. Watch connection attempts now retry after timeouts, and periodic restarts correctly recreate connections without swallowing unrelated cancellation. Regression introduced in 13.6.0. (#​20664, backport of #​20466, @​radical)

• 🪟 Session-scoped containers could remain running after stopping a .NET AppHost on Windows — AppHosts launched through dotnet run or dotnet watch now allow DCP to finish resource cleanup during shutdown, including existing Ctrl+C and aspire stop flows. Persistent resources remain unaffected, while direct .NET launches and TypeScript/polyglot AppHosts retain their existing process protection. Fixes #​20495. (#​20514, backport of #​20494, @​danegsta)

---

Full Changelog: v13.6.0...v13.6.1

Full commit: 3751e615ad660621f2a7aa4c390ef48e08a0eed8

Commits viewable in compare view.

Updated bunit from 2.11.3 to 2.13.3.

Release notes

Sourced from bunit's releases.

2.13.3

Fixed

• Find threw TypeLoadException with AngleSharp 1.8.6 because HtmlSelectElementWrapper did not implement the new SelectedIndex setter. Reported by @​poxet in #​1957. Fixed by @​linkdotnet.
• A WaitFor* call whose first check passed left its timeout timer running, which kept the helper, the renderer and everything the renderer holds alive until the timeout elapsed. With a long DefaultWaitTimeout, a test suite could hold every rendered component in memory for its whole run. Fixed by @​xaviergxf.

2.12.5

Fixed

• WrapperBase.WrappedElement could return null and throw NullReferenceException when an element was accessed concurrently with component re-renders. Reported by @​psu-prog in #​1937. Fixed by @​Nacl-HCl.
• DisposeComponentsAsync() could dispose nothing when the renderer's dispatcher was busy, e.g. with a timer calling InvokeAsync. Reported by @​xfunc-tim-schneider in #​1939. Fixed by @​linkdotnet.
• Markup generation and FindComponents could recurse until the test host crashed with a stack overflow when a render threw while instantiating a child component. Reported in #​1935. Fixed by @​linkdotnet.
• Find() could keep returning DOM nodes from the previous render when a render, e.g. from a timer, happened while the nodes were parsed, causing UnknownEventHandlerIdException. Reported by @​xfunc-tim-schneider in #​1938. Fixed by @​linkdotnet.

Commits viewable in compare view.

Updated MailKit from 4.18.1 to 4.18.2.

Release notes

Sourced from MailKit's releases.

No release notes found for this version range.

Commits viewable in compare view.

Updated Marten from 9.45.0 to 9.50.3.

Release notes

Sourced from Marten's releases.

9.50.3

The Native AOT and hard-delete release. One new feature on the event store, two schema-naming fixes that could not be worked around, and three behaviour changes — read those before upgrading.

Note on these notes. They were rewritten after publication. 9.50.3 was cut as the JasperFx 2.90.1 dependency bump, but five pull requests had merged in the minutes before the tag, so the tag carried all of them too. The original notes described only the dependency bump and said no code change was needed. Everything below is what 9.50.3 actually contains; nothing about the published packages changed.

New

• TombstoneStream — a hard stream delete (#​5673, #​5674). session.Events.TombstoneStream(id) queues a single-statement delete of a stream's tag rows, its events and its stream row — plus the mt_streams_identity row under strict identity enforcement — scoped to the session tenant under conjoined tenancy. The event sequence is never touched, and projected documents are deliberately left alone. Queued until SaveChangesAsync, like every other write.

  It is on the shared IEventStoreOperations contract as a default member that throws NotSupportedException, so a store without a hard delete stays source-compatible, and the shared StreamTombstoningCompliance suite now runs against Marten.

• DCB tag types can be registered for Native AOT (#​5649, #​5674, #​5676). RegisterTagType<TTag>() discovers the tag's inner type by reflecting over its members and closes a generic registration at runtime, neither of which a native image can do — so a store with a DCB tag type could not even be configured under AOT.

  opts.Events.RegisterTagType<StudentId, Guid>(x => x.Value, "student")
      .ForAggregate<StudentCourseEnrollment>();

  Naming the inner type closes the generic at compile time and the unwrapper replaces the reflection. The suffix is optional and defaults to the tag type's own name, so for a tag the inferring overload could register, this one registers the same inner type and the same table. It works identically under a JIT.

  The inferring overloads are now annotated [RequiresUnreferencedCode]/[RequiresDynamicCode], and — because an analyzer warning only reaches an application that is itself AOT-analyzed, while a store configured from inside a library gets nothing — a failure at runtime now throws naming the exact line to write, keeping the original exception as its inner exception. Automatic tag registration from projection identities logs a warning naming the same line, including the projection's ForAggregate<T>().

  With explicit registrations, tagging, tag rules, EventTagQuery, EventsExistAsync, AggregateByTagsAsync and FetchForWritingByTags into a [BoundaryAggregate] all run from a native binary. A CI job publishes and runs them.

Fixed

• A custom foreign key name is no longer discarded under conjoined tenancy (#​5672, #​5678). Moving tenant_id to the front of a document foreign key regenerated the key's name unconditionally, so a name set through the foreignKeyConfiguration callback was thrown away. That is what made the identifier-length failure unfixable: the exception tells you to override the database object name, and overriding it had no effect. A generated name is still regenerated; a chosen one survives, as does the name of a ForeignKey added by hand.

• Archiving a tagged stream no longer fails with a foreign key violation (#​5675, #​5679). With UseArchivedStreamPartitioning on and any DCB tag type registered, ArchiveStream came back as 23503. Each mt_event_tag_* table is partitioned by is_archived and carries a composite foreign key to mt_events; mt_archive_stream moved the events between partitions without moving the tag rows that referenced them. It now moves them, in the order that keeps both sides of a non-deferrable key consistent at every statement.

• AllObjects() no longer applies the Hi-Lo DDL (#​5665, #​5670). Describing the schema provisioned part of it. The describing seam is now separate from the applying one, and is proven offline against an unreachable host.

• A failed Rebuild-mode run can be restarted and finished from its committed progress (#​5668, #​5669).

Performance

• A commit that only starts streams no longer queries for a snapshot that cannot exist (#​5671, thanks @​lahma). An inline single-stream projection asks storage only for the ids of streams being appended to, so a start-only commit passed an empty id array — and still paid a round trip on id = ANY($1) that could only return no rows, plus a second connection with the default UseIdentityMapForAggregates = false. The same guard removes a wasted round trip from an async daemon batch whose slices all hit the aggregate write cache.

Behaviour changes

1. An empty DCB tag table suffix is refused. RegisterTagType<T>("") used to be accepted and named mt_event_tag_ — a real table that every tag type registered that way would share. Pass a suffix, or null for the tag type's own name.

2. A tag type that cannot be read now throws InvalidOperationException, with the original InvalidValueTypeException as its inner exception, and a message naming the explicit registration. This happens whether or not dynamic code is available, because the same exception means "not a single-value wrapper" under a JIT and "the members were trimmed away" in a trimmed image, and nothing at runtime tells those apart.

3. Document foreign key and index names longer than 63 characters are shortened with a deterministic hash suffix, the same treatment the document table's primary key already got. If you raised PostgreSQL's NAMEDATALEN and have a foreign key or index whose generated name is between 64 characters and your own limit, that object will be renamed and you will see it in a migration diff.

Known, not fixed

DCB tag queries return archived events, where every other event query excludes them by default — #​5677. It matters most for EventsExistAsync, which is what a consistency boundary assertion is built on: an archived stream still answers yes. Independent of storage mode and of archived-stream partitioning.

Dependencies

... (truncated)

9.50.2

A maintenance release. Nothing in a shipped Marten package changed, and nothing here requires a configuration or code change.

Dependencies

• The NuGet vulnerability audit is clean (#​5666). Marten's shipped packages were never affected by GHSA-846f-f764-76mx (CVE-2026-84008, System.Security.Cryptography.Pkcs). Npgsql doesn't bring that package in. Every finding was in the Nuke build tool or the samples:
  • The build tool now uses patched System.Security.Cryptography.Pkcs, System.Security.Cryptography.Xml and NuGet.Packaging.
  • The OpenTelemetry packages used by the EventPublisher sample and the Aspire sample moved from 1.8.x to 1.19.x.

The same advisory does reach Polecat and Weasel's SQL Server and Oracle packages, through their database drivers. Those are fixed in Polecat 5.40.2 and Weasel 9.45.1.

9.50.1

A patch release with one addition, for Wolverine's agent assignment. Nothing here requires a configuration or code change.

Event store

• IEventStore.TryDescribeDatabasesAsync() is answered straight from the tenancy (#​5664). JasperFx 2.87.1 adds this narrower call, the database half of TryCreateUsage(). It returns each database's server, name and tenant ids without reading the highest event sequence or describing every projection, subscription, event type and tag type. Wolverine 6.50.0 asks for it on every leader heartbeat, for every store, where it used to build the full usage (JasperFx/wolverine#​5002). It follows on from #​5652, so thanks again to @​uniquelau for a report that came with the heartbeat maths already done.

Dependencies

• JasperFx 2.87.1.

9.50.0

A bug-fix release. Nothing here requires a configuration or code change, but three things change what an application observes — read the box below before upgrading.

Observable changes

• TotalResults changes value for Stats() + Include() + Select() + Take(). It was reporting the size of the page; it now reports the number of matches (#​5654). If you compute page counts from it, those counts change — to the correct ones.
• A new warning may appear at startup if tag-type auto-discovery cannot register a projection's identity type. That condition is not new; it was previously silent (#​5649).
• db-patch drop files no longer DROP ... CASCADE every schema — they roll back only what the patch touched (Weasel 9.45.0, weasel#​706). This is the one to look at if you apply drop files in an environment you care about.

LINQ — Include() joined on the queried document's own id

Three separate defects in the same corner, found by pulling on the first one. An Include() joined on the queried document's own id builds a temp table, and that temp table was missing the id it is joined on.

• #​5650 — failed with 42703: column d.id does not exist in a query session when the queried document had any duplicated field. Lightweight, identity-map and dirty-tracked sessions were unaffected. Thanks to @​Blackclaws for the report and the fix (#​5651).
• #​5653 — the same 42703, in a query session, when the query also called Stats(). Unlike #​5650 this did not depend on a duplicated field; Stats() wraps the select clause, which hid it from the check that puts the id back.
• #​5654 — Stats() with an Include() and a Select() projection reported TotalResults as the size of the page rather than the number of matches, whenever Take() was used. No exception, just a wrong number, on every session kind. The temp table already held the correct count — PostgreSQL evaluates a window function before LIMIT — but re-applying the projection emitted a second count(*) over() against the already-truncated temp table.

If you page a projected grid with Stats() and Include(), #​5654 is the one to read.

Event store

• #​5652 — a store with no event types and no projections never has its event tables migrated, so building a usage descriptor asked Postgres for a table that was not there. Marten swallowed the failure, but Postgres logged an ERROR for every attempt; under UseWolverineManagedEventSubscriptionDistribution the leader does this for every store on a ten-second heartbeat, which is tens of thousands of server-side ERROR lines a day for a store that can never hold an event. Marten now asks the catalog first, so nothing reaches the server log. Thanks to @​uniquelau for a report that came with the heartbeat maths already done.
• #​5658 — IEventStore.FindDatabaseForTenantAsync, so a caller waiting on or querying a single tenant touches only that tenant's database instead of all of them. Goes through ITenancy.TryFindDatabase, never FindOrCreateDatabase, so a lookup cannot bring an unknown tenant into existence.

Native AOT

• #​5649 (partial) — a DCB [BoundaryAggregate] fetch failed natively with MissingMethodException, because EventGraph.Build<TDoc>() closed SingleStreamProjection<,> at runtime even though both type arguments were known at compile time. And tag-type auto-discovery swallowed the AOT failure in the same breath as a benign one, leaving a store running with the tag type quietly absent; it now says so. Verified by publishing and running the native binary, not by a clean publish.

  Still open: RegisterTagType<TTag>() closes its own open generic and remains broken under Native AOT. That needs a reflection-free construction path in JasperFx.Events plus a new overload — see the discussion on #​5649. Thanks to @​skoging for an unusually thorough report, with native runs and three costed alternatives.

Dependencies

• Weasel 9.45.0 — db-patch drop files no longer drop every schema with CASCADE; they roll back only what the patch touched (weasel#​706). This changes what an existing drop file does. Primary keys can also now be added to and removed from existing tables (weasel#​707).
• JasperFx 2.87.0 — across all five packages.

Known issue

Stats() with Include() and a Select() projection is fixed here, but note that the Include() itself still follows the page, not the match set — it reads from the same truncated temp table. Only the statistics column outlives the Take(). That is existing, intended behaviour; it is called out because #​5654 makes the distinction visible for the first time.

9.49.0

The LINQ projection surface works under Native AOT

The theme of this release. #​5624 came in as "a projection into an anonymous type cannot be
materialized under Native AOT". Publishing Marten.AotRuntimeSmoke natively and running it showed
something much wider: every Select() projection into a new object failed from a native binary,
a named DTO exactly as hard as an anonymous one, and it failed before the serializer was ever
reached.

MissingMethodException: No parameterless constructor defined for type
'Marten.Linq.SqlGeneration.SelectDataSelectClause`1[YourDto]'

A projection is parsed out of an expression tree, so the parser only has selectExpression.Type and
closed its select clause over that Type with MakeGenericType + Activator.CreateInstance — the
defect class of #​5328, one step later in the pipeline. Only a bare string scalar survived, because
its clause happens not to be generic, which is what made the surface look healthy.

Shapes that threw from a native binary before this release and run now:

• Select(x => new Summary { ... }) — a named DTO
• Select(x => x.Lines) — a collection member, and the element type of a SelectMany(x => x.Lines)
• Select(x => x.Amount) for every framework scalar type, decimal? and the other nullable ones included
• Select(x => x.Status) and MaxAsync(x => x.Status) for an enum of your own — see the one line below
• GroupBy(...).Select(g => new Tally { ... })

The fix is SelectClauseResolvers, which does for projected types what #​5328's
DocumentStorageResolvers does for document types: a registry populated from
MartenLinqQueryable<T>'s constructor, which every projected type reaches because
Queryable.Select<TSource, TResult>() asks the provider for CreateQuery<TResult>(). The reflective
path remains the fallback wherever a JIT is available, so nothing changes for JIT-mode apps.

One line for an enum or struct of your own

Marten's scalar select clause is declared where T : struct and the registration point above
constrains its own T to notnull, so Marten can root the scalar types it renders itself —
decimal, Guid, DateOnly and the rest need nothing — but not one out of your assembly:

opts.Schema.For<Order>();
opts.RegisterScalarType<OrderStatus>();   // new in 9.49.0; no-op under a JIT

Only bare scalar projections need it. Projecting the same enum as a member of a named type, and
filtering on it, never did. When the registration is missing, the failure now names that exact call
instead of reporting a missing parameterless constructor on an internal generic.

An anonymous projection is refused with the remedy

With a source-generated JsonSerializerContext assigned to
... (truncated)

9.48.0

Document diagnostics: filter and order by document properties (jasperfx#​869, #​5631)

IDocumentStoreDiagnostics.QueryDocumentsAsync now applies DocumentQueryOptions.Where / OrderBy / Arguments instead of refusing them. The text goes onto Query<T>() through JasperFx 2.84.0's Dynamic LINQ helper, so member casing, enum storage, duplicated fields, conjoined tenancy, soft deletes and hierarchies are Marten's own; only the SELECT list is swapped for the diagnostics read's columns, so the JSON comes back byte-exact with its metadata, and ...

Description has been truncated

Dependabot will...

Description has been truncated