[michaelnebel]

Instead of searching through all files (including source code files) we attempt to read the files as XML instead for detecting framework and package dependencies (and other project related settings). This will

• Implicitly enable "fail-fast" (as an exception will be thrown if a file is not XML).
• Make it easier to extend the logic for detecting certain properties.

Note, that this change impacts the semantics of the heuristic.

• Only files parseable as XML are taken into consideration.
• The matching is made case-sensitive.

We expect this to be an implicit improvement as well

• Accuracy improvement: XML snippets in text files (or source code files) are no longer taken into consideration.
• Performance improvement: We do less parsing of text as the XmlReader gives up and throws an exception, when a file is not parseable as XML (before we read all text from all files and did regular expression matching).

DCA looks good

• Drastic reduction in the number of calls with missing call target for dotnet/efcore.
• Performance appears unaffected.