The file lib/dbtexmf/core/sgmlent.txt, which contains mappings from SGML characters to Unicode, says that its contents are drawn from various sources and that the file is in Public Domain. The mere fact that the file has various sources makes me doubt that some of the data isn't copyrighted in some way.
Could you please review if it is indeed all right?
Blocking FE-LEGAL on this.
Even if it didn't, this file falls under the "Phone Book" exception to copyright, as it simply contains a list of SGML character entity names and their corresponding sets and char codes. So, it is safe to treat this as Public Domain. Closing NOTABUG.