Outlook is sending the PDFs using quoted printable rather than base64 encoding. Unfortunately, it’s not encoding the line breaks within the file, and when your mail client opens the file it correctly switches these to system line endings. In other words 0D 0A in the original gets replaced by 0A when you detach it. This is why the corrupted file is smaller than the original. Unfortunately, PDFs are not text files, and these substitutions will break the file.

I can’t explain why it works when it is resent. Presumably Outlook, for reasons best known to itself, switches to base64 encoding. You can check this by looking at the source of the emails and seeing which encoding Outlook has used in both cases.

My guess is that the PDF was sent not with the MIME (Multipurpose Internet Mail Extensions) Content Type of application/pdf and rather used the fall back MIME Content Type of “Content-Type: application/octetstream;”

The IETF’s RFC 2046 says:

The recommended action for an implementation that receives an
“application/octet-stream” entity is to simply offer to put the data
in a file, with any Content-Transfer-Encoding undone, or perhaps to
use it as input to a user-specified process.

The problem is that the developer of an email client can interpret the RFC as they see fit. As shown below every PDF attachment is done a little differently.

AOL

--part1_5602.21b11067.3d165f85_boundary
Content-Type: application/pdf; name="SSS.pdf"
Content-Transfer-Encoding: base64


Web Generated

-----=_NextPart_002_1DDF_0D61693F.F2222C51
Content-Type: application/octetstream;
name="BKD-7362630486.pdf"
Content-Disposition: attachment;
filename="BKD-7362630486.pdf"
Content-Transfer-Encoding: base64


Outlook

------=_NextPart_000_005D_01CCEFEE.E3B18A60
Content-Type: application/pdf;
name="QQQ.pdf"
Content-Transfer-Encoding: base64
Content-Disposition: attachment;
filename="QQQ.pdf"


UNKNOWN

Apple
--Apple-Mail-18--970524984
Content-Disposition: inline;
filename="AAA.PDF"
Content-Type: application/pdf;
name="AAA.PDF"
Content-Transfer-Encoding: base64


So that was my eventual solution: if an attachment is borked in Thunderbird, download it via the GMail web interface. Unsatisfying, but practical.

I know and have no reason to think that things have changed – that Google continues to scan the accounts and uses the information for purpose of advertising. Has anyone considered that Google may be damaging these files while scanning them?

I made “datanoia” up on the spot, though doubtless I wasn’t the first, so it really just meant what I had in my head, which is: “paranoia about the data one receives”, with a secondary definition of “paranoid about the data one shares”—sort of a mirror image of the first definition. That was all. Though I suppose it also makes for a useful way to ferret out blog-scrapers!

Drives me insane.

I wonder if there might be .js in there, perhaps doing something funky or even mean?

FWIW,
-s.