[FEATURE] Fixing greatest CS problem of cache invalidation by giving people Roslyn code analyzers

[tvardero]

It might be a completely new project tho. It will be really large with lots of rules, quick fixes, etc. (like Roslynator).

I currently don't know how to determine at compile time where cache is not invalidated, this might require analysis of code flow, integration with EF change tracking (again at compile time change tracker does not even exist yet).

For now what can be done is configuration analysis and validation, like if FailSafe time is less than Duration.

I'm opening this issue as idea for some unknown future.

[tvardero]

Actually cache invalidation problem can already be analyzed at some extent by code QL / AI PR review tools.

[jodydonetti]

Hi @tvardero

interesting idea, even though I find it hard to envision something that actually works...

It might be a completely new project tho.

Yeah most definitely.

I currently don't know how to determine at compile time where cache is not invalidated, this might require analysis of code flow, integration with EF change tracking (again at compile time change tracker does not even exist yet).

For now what can be done is configuration analysis and validation, like if FailSafe time is less than Duration.

I'm opening this issue as idea for some unknown future.

Good idea, let's keep this as a playgrtound to explore potential ideas.

Actually cache invalidation problem can already be analyzed at some extent by code QL / AI PR review tools.

Do you have any example of that?

I can already imagine a lot of scenarios where that would be basically impossible... I'm necessarily comparing it to the halting problem, but still definitely in that realm 😅

Let me know, and thanks for sharing!