Egor Bogatov

> @EgorBo does it mean that in the final .NET7 release this will be fixed? If so I can close the issue No, Main is .NET8, we need to check...

Relaxed checks made NativeAOT unhappy, let me see why

``` Failed test: nativeaot_SmokeTests::_DynamicGenerics_DynamicGenerics_DynamicGenerics_cmd (nativeaot\\SmokeTests\\DynamicGenerics\\DynamicGenerics\\DynamicGenerics.cmd) Unhandled Exception: EETypeRva:0x00496880(System.Diagnostics.DebugProvider+DebugAssertException): IsInstanceOfClass called with parameterized MethodTable at System.Diagnostics.DebugProvider.Fail(String, String) + 0x36 at System.Diagnostics.Debug.Fail(String, String) + 0x21 at System.Runtime.TypeCast.IsInstanceOfClass(MethodTable*, Object) + 0x48 at System.Runtime.TypeCast.CheckCastClass(MethodTable*,...

> The optimization is valid for NativeAOT. > > The problem seems to be that the codegen is violating the JIT helper contract. NativeAOT asserts. CoreCLR has a test hole...

@jkotas thanks for suggestion, the code is cleaner now! @dotnet/jit-contrib PTAL

Hm something is off on CI - with `Git fetch`

I assume it makes `GCHeap::UpdateFrozenSegment` that I added in https://github.com/dotnet/runtime/pull/49576 useless, right? (it reports to gc when vm commits more memory for a frozen segment)

> is it okay to throw OutOfMemory at the spot you call RegisterFrozenSegment or UpdateFrozenSegment? I think so, we already throw OOM in some cases from those call sites

Seems to be passing CI tests (SPMI failures are due to JIT-EE change), @jkotas when you have time could you review the vm side?

CI failures are weird, e.g. https://helixre107v0xdeko0k025g8.blob.core.windows.net/dotnet-runtime-refs-pull-75573-merge-3616944558404cd193/System.Linq.Queryable.Tests/1/console.13d74af2.log?helixlogtype=result had no luck trying to reproduce any, going to kick off a single gc stress job - maybe it will help to find a...