On TIKA-2118, Seva Alekseyev shared a document that causes an UnsupportedCodePage exception. The file is an old xls (BIFF5) that is encrypted. After looking through https://www.openoffice.org/sc/excelfileformat.pdf and experimenting with some files that cause similar exceptions in Tika's regression corpus, it appears that all records after a file pass record are encrypted, even the contents of the code page record. Let's throw an EncryptedDocumentException (Encryption not supported for old excel files). I did find one file that doesn't appear to be encrypted (attached to TIKA-2118). I can open it, but it is write protected...so even though I can open it and copy and paste contents out of it, the inner contents are encrypted.
r1765829