The Android Soong Build system can generate mostly correct SPDX docs.
It does include the external LICENSE text for all components in though hasExtractedLicensingInfos / extractedText.
Some extractedText values include the formfeed JSON control character: \f.
The JSON formfeed control character will deserialize to \x0c in python (ASCII formfeed control char).
The problem is that when converted to XML these control characters will corrupt the RDF+XML document as
most control characters in XML are not supported in general (both XML 1.0 and 1.1).
The proposed fix (being implemented) is to strip the \b and \f JSON control characters as those have no semantical value.