Question about how to use Recyclable memory stream manager about microsoft.io.recyclablememorystream HOT 6 CLOSED

The call to memoryStream.ToArray() is undoing any possible optimizations of using a pooled stream. The memory is probably going up in production because you have so much spare memory that no GC happens.

In your case, you would need to keep the stream alive (undisposed) until after you were done with the TransferFile it was being used in. If TransferFile can only accept a buffer, it should also accept a length so that it knows how much of the buffer is valid.

from microsoft.io.recyclablememorystream.

The call to memoryStream.ToArray() is undoing any possible optimizations of using a pooled stream. The memory is probably going up in production because you have so much spare memory that no GC happens.

In your case, you would need to keep the stream alive (undisposed) until after you were done with the TransferFile it was being used in. If TransferFile can only accept a buffer, it should also accept a length so that it knows how much of the buffer is valid.

I can't avoid using the ToArray() because it is calling a webservice that only accepts a byte array. Is it in this case better to use a normal MemoryStream instead of a Recyclable one?

from microsoft.io.recyclablememorystream.

from microsoft.io.recyclablememorystream.

This goes a little bit above my head how all this works, the reason I started with the RecyclableMemoryStream was because I move a lott of data back and forward. I used to have regular out of memory exceptions even on byte arrays smaller than 200MB. These problems were gone when starting to use this library.

Is it possible to use GetBuffer() instead of ToArray() to get the underlying byte array to avoid that a new copy is made in memory? Or is this asking for other problems?

from microsoft.io.recyclablememorystream.

GetBuffer is preferred to ToArray because it will return a reference to the underlying buffer currently in use, but since those buffers are pooled, they will be larger than the amount of data you have stored. If the API you pass the buffer to doesn't understand that, it might assume the whole buffer has valid data, which isn't the case. That's why it's important to also use the length of the stream. ToArray always creates a new array of exactly the right length. Useful for those APIs that need just a buffer, but it destroys the benefits of pooling.

from microsoft.io.recyclablememorystream.

Thanks for helping me out with these questions, best chrismas wishes and have a good new year.

from microsoft.io.recyclablememorystream.