API Proposal PipeReader ReadAsync could take an expected size

[Drawaes]

Because the flow is inverted with pipelines there is less understanding at the lower levels of the expected data. Often the application protocols have an idea (or if no the application protocol at least the intermediate ones).

For instance in Tls if I only get a partial frame, I know the size of the rest of the frame. Rather than having to loop over and over on ReadAsync() if I haven't got enough yet it would be good to be able to call

var result = await ReadAsync(expectedSize : 40)

and it not return until the connection closes/dies/or there are enough bytes to fill that amount.

[AlgorithmsAreCool]

I think this can be done in an extension method right?. It would just be a loop that that calls Advance(Buffer.Start, Buffer.End) until Buffer.Length exceeds expected size or until ReadResult.IsCompleted returns as true.

It would probably be pretty useful.

[Drawaes]

It could be, and that might be the first iteration but you might never call the ReadAsync continuation of there isn't enough data as well, saving a whole lot of code from running.

[ahsonkhan]

@maryamariyan - the bot didn't label this issue.

[davidfowl]

Fixed in #51979. Dupe of #25063