That specifies what string
will separate each subsequent attachment,
but I meant the header part of the PDF attachment itself,
like this one which I just sent to myself:
Content-Type: application/pdf; name="SevenColors.pdf"
Content-Disposition: attachment; filename="SevenColors.pdf"
The first few characters of the PDF file should also be
a standard PDF header; for this file it happens to be:
If you can't obtain the actual first bytes from the stored file,
the first line of the "base64" encoding (following the header)
would suffice, as we can decode that, such as on-line at:
Any comparison of PDF attachments (headers and first bytes)
which are okay, vs. those not okay, might also be helpful.