MTP v2 Breaking Changes

[Youssef1313]

Planned

• [Breaking][MTP v2]: Remove ITestApplicationLifecycleCallbacks in favor of ITestHostApplicationLifetime #5733
• [MTP v2] Consider breaking old dotnet test experience #5722
• [MTP] Error when VSTest arguments are passed via old dotnet test when they will be ignored #5697 (NOTE: Not needed in v2 if we remove the whole thing in v2 ;))
• Drop support for < net8.0 in MSTest and MTP #5614 (NOTE: This is common for MTP and MSTest)
• [MTP v2][Breaking] Revise MTP properties and where they do belong (test node vs update message) #5524
• [MTP v2][Breaking] Avoid the use of C# records in public API #5517
• [MTP][Breaking] Remove TrxFullyQualifiedTypeNameProperty once we get rid of vstestProvider capability #5474
• [MTP v2] Add CancellationToken parameter to more async methods. #5342
• Rename diagnostic options to use kebab case #3630
• Make the TestNodeUpdateMessage and in general every payload structure immutable. #3473
• Review all exposed extensibility points and substitute the couple SessionId and CancellationToken with Microsoft.Testing.Platform.Services.ITestSessionContext #3356
• Microsoft.Testing.Platform.Extensions.TestHost.IDataConsumer should be defined inside the Microsoft.Testing.Platform.Extensions #3354
• TrxReport should use TestNodeUid as-is if it's a Guid already.

To discuss

• [ ]

Cancelled

• [MTP][Breaking for v2] AddSelfRegisteredExtensions shouldn't need to have string[] args parameter #5453
• ITrxReportCapability interface design has a flaw where "enabled" state is unknown to callers #5532
• use ValueTask instead of Task if possible #3474

[Youssef1313]

This will mostly affect non-VSTestBridge users. As for VSTestBridge, we will be addressing these ourselves in the VSTestBridge package.

cc @thomhurst @bradwilson

[bradwilson]

What's the proposed release timeframe? I assume it's aligned with .NET 10...?

[Youssef1313]

Not yet known: #5614 (comment)
It's unlikely to be before .NET 10. But whether it's aligned with .NET 10 or some time after is something we don't know yet.