Improve ImageSharpMiddleware no-op case's performance

[lahma]

Prerequisites

• I have written a descriptive pull-request title
• I have verified that there are no overlapping pull-requests open
• I have verified that I am following matches the existing coding patterns and practice as demonstrated in the repository. These follow strict Stylecop rules 👮.
• I have provided test coverage for my change (where applicable)

Description

I profiled Orchard Core as @sebastienros gave the idea. I saw that ImageSharp was doing some unnecessary processing as part of the pipeline when the request ended to be no-op on ImageSharp's behalf. I tweaked following bits:

• reduce usage of LINQ, which is always more expensive than "code nearer to the metal"
• change error handling to be more easy to inline (micro optimization)
• use concrete types as fields, no extra penalty from interface access
• split the actual work to ProcessRequest when actually needed
• don't parse in OnParseCommands when result is discarded

[CLAassistant]

All committers have signed the CLA.

[JimBobSquarePants]

This is really good! Just a few minor changes required and I'll be very happy to merge this in. Thanks! 👍

[lahma]

Thank you for the review, good points! I hope I got the fixes right and PR should be good now.

[JimBobSquarePants]

Perfect! I'll merge this now, thanks for your help!

[lahma]

You're most welcome, thanks for accepting the PR!