Avoid failing scans outright when encountering extraction failures

[egibs]

We currently fail scans when encountering extraction errors which will prevent any report generation. This is not great when scanning many files wherein only one or a handful of files cannot be processed.

This PR introduces an opt-out flag (--exit-extraction) to ignore these failures. By default, the behavior remains the same but it's easy enough to default this to false in the future. I added two new test cases to cover the new configuration option and it behaves as expected.

I also addressed an inconsistency between mimetype.Detect and Path. The latter would not return a valid MIME type which would cause issues if we were expecting ft.MIME to be populated. We also need to keep a list of valid archive extensions and at least one MIME type in the supportedKind map. In cases like .zip or .gz extractions, we can do another lookup to cover additional types in cases where mimetype.Detect is used over Path.

FWIW, we do mitigate this in other places by doing something like:

if err != nil {
	if strings.Contains(err.Error(), "failed to extract") {
		log.Error(err.Error())
		return nil
	}
	return err
}

[eslerm]

Appreciate that a minified test is included 🙏