In .NET 6, @Maoni0 introduced the GCGlobalHeapHistory_V4 in https://github.com/dotnet/runtime/pull/55888 that expose some timing information for each phase of the GC and I did the parsing of it in https://github.com/microsoft/perfview/pull/1758. It turns out just parsing it is not enough, we need to expose it as property on TraceGC in order to be consumed, so this PR expose that array.
Unfortunately, the compact time can often be wrong because a bug I fixed in https://github.com/dotnet/runtime/pull/100599, practically that means for some traces, the compact time will be wrong (often seen as 0)
cshung
commented
2 months ago
In .NET 6, @Maoni0 introduced the
GCGlobalHeapHistory_V4in https://github.com/dotnet/runtime/pull/55888 that expose some timing information for each phase of the GC and I did the parsing of it in https://github.com/microsoft/perfview/pull/1758. It turns out just parsing it is not enough, we need to expose it as property onTraceGCin order to be consumed, so this PR expose that array.Unfortunately, the compact time can often be wrong because a bug I fixed in https://github.com/dotnet/runtime/pull/100599, practically that means for some traces, the compact time will be wrong (often seen as 0)