Confusion about PipeReader

[GF-Huang]

According to https://learn.microsoft.com/en-us/dotnet/standard/io/pipelines#pipereader:

PipeReader manages memory on the caller's behalf. Always call PipeReader.AdvanceTo after calling PipeReader.ReadAsync. This lets the PipeReader know when the caller is done with the memory so that it can be tracked. The ReadOnlySequence<byte> returned from PipeReader.ReadAsync is only valid until the call the PipeReader.AdvanceTo. It's illegal to use ReadOnlySequence<byte> after calling PipeReader.AdvanceTo.

If I want to reference a ReadOnlyMemory<byte> (which inside ReadOnlySequence<byte>) in memory for a long time, should I ReadOnlySequence<byte>.CopyTo/SequenceReader<byte>.TryCopyTo? If I use SequenceMarshal.TryGetReadOnlyMemory, are there any potential problems?

[AlgorithmsAreCool]

Not sure if you have gotten your answer elsewhere, but if you want to keep the data after you call AdvanceTo you will need to copy it. If you use TryGetReadOnlyMemory it will give you a reference to the underlying memory that the Pipe owns. After AdvanceTo is called then the Pipe may reuse or clobber that memory.

The alternative of course is to delay the call to AdvanceTo until you are done with the memory, but that might not be possible as you are hinting.