these analyzers will create rules on the fly based on given options. this doesn't fit to current diagnostic analyzers which assumes those sets are known before-hand.
we need a way to support this in compiler diagnostic analyzer framework, otherwise, this should use other mechanism than diagnostic analyzer since it doesn't fit into the diagnostic analyzer architecture at all.
currently IDE has namingstyle analyzers. http://source.roslyn.io/#Microsoft.CodeAnalysis.Features/Diagnostics/Analyzers/NamingStyles/NamingStyleDiagnosticAnalyzerBase.cs,11
these analyzers will create rules on the fly based on given options. this doesn't fit to current diagnostic analyzers which assumes those sets are known before-hand.
we need a way to support this in compiler diagnostic analyzer framework, otherwise, this should use other mechanism than diagnostic analyzer since it doesn't fit into the diagnostic analyzer architecture at all.