extract the union of features from multiple extractors

[williballenthin]

when multiple extractors, e.g. native PE and dotnet, can extract features from a binary, e.g. a mixed-mode dotnet program, should we extract the union of features from both extractors? or consider the features separately?

[williballenthin]

the union seems intuitive and is my intuitive preference, but needs further thought and/or discussion

[mr-tz]

introduced in #955 multiple file-level extractors can run on the same file

the full-feature extraction is not yet implemented that way and will require either

• harder:
  • more refactoring
  • ways to merge metadata, capabilities, etc. by each extractor
• easier:
  • modifying the results document to contain multiple results
  • display multiple results tables on the CLI

so far this would only apply to mixed-mode assemblies, which are rather rare