ByteRange, incremental update, and CR + LF = EOL

[a20god]

Suppose a PDF/A-2 document ends with CR, like this one:
B.1-unknown-3.pdf
The ByteRange entry covers the entire document, the document conforms to PDF/A-2. We add an incremental update. The incremental update begins with a LF as the object number must be preceded by an EOL marker (ISO 19005-2:2001, 6.1.9). Now, veraPDF reports that the ByteRange no longer covers the entire document (as signed). This is probably due to the LF being considered to by part of the document (as signed) as together with the CR it forms an EOL marker:
B.1-unknown-4.pdf
That is, the document (as signed) became longer by one byte. Is this really the intended behavior?

(For a PDF writer it's probably safer to start an incremental update with a CR rather than a LF in order to avoid this problem. Doesn't help for existing signed documents, though.)

[a20god]

veraPDF 1.11.11 still reports a ByteRange not covering the entire file for B.1-unknown-4.pdf above.

[a20god]

The logic might still be wrong. Consider this document:
B.1-01-fail-4.pdf
The document ends with CR+LF and ByteRange does not cover the final LF. veraPDF 1.11.19 does not report a violation of clause B.1. The logic allowing the document to be a bit longer (by a CR or CR+LF) should come into play only if an incremental update was added after the signature.