Small improvements to reduce allocations in validation filter logic

[Copilot]

I'm implementing improvements to reduce allocations in the validation filter logic. Here's my progress so far:

• Avoid allocating arrays in ValidationEndpointFilterFactory.Create for parameters that don't require validation
  • Replaced fixed-size arrays with a List of structs containing only the validatable parameters
  • Only parameters that need validation are stored, improving memory usage when most parameters don't need validation
• Optimize ValidationContext reuse throughout the pipeline to avoid creating new ones per handler argument
• Reduce closure allocations in EndpointFilterDelegate returned from the filter factory
• Cache reflection calls for implemented subtypes in ValidatableTypeInfo

These changes will make the validation filter logic more efficient and pay-to-play, reducing allocations when users are not leveraging the feature.

Fixes #61222.

---

💡 You can make Copilot smarter by setting up custom instructions, customizing its development environment and configuring Model Context Protocol (MCP) servers. Learn more Copilot coding agent tips in the docs.

[dotnet-policy-service]

Greetings human! You've submitted a PR that modifies code that is shared with https://github.com/dotnet/runtime . Please make sure you synchronize this code with the changes in that repo!

[captainsafia]

@gfoidl Thanks for posting these comments! I've been trying out the new Copilot feature to see how helpful it can be. One of its limitations is that it only responds to feedback from the person who initiated it. I'll go through your comments and add them under my own review so the agent can be triggered.

[captainsafia]

Good start! More feedback to address here.

[captainsafia]

@BrennanConroy Thoughts on this change? It didn't execute on all the items in the original list but I think this change is good as is.

One thing to note is that we have no benchmarks for the validation filter specifically (that's my bad) so we don't have a good way of getting numbers here but assuming the change doesn't break any tests the delta here feels straightforward to review.