[exim-dev] [Bug 1926] New: DKIM TXT record parsing too lenie…

https://bugs.exim.org/show_bug.cgi?id=1926

            Bug ID: 1926
           Summary: DKIM TXT record parsing too lenient ("\" escaping)
           Product: Exim
           Version: 4.87
          Hardware: x86-64
                OS: Linux
            Status: NEW
          Severity: bug
          Priority: medium
         Component: DKIM
          Assignee: tom@???
          Reporter: chibi@???
                CC: exim-dev@???

We had a wrongly configured TXT record for one domain that looked like this:
"v=DKIM1\\\; k=rsa\\\; p=..."

When verifying correctly signed mails for this domain Exim (in the receiving,
MX role) vetted them with "[verification succeeded]".

Whereas google.com or other DKIM verification systems took offence to the
triple backslash as opposed to the correct single one.

My (uneducated, non-source checking) guess would be that the parsing inside
Exim just ignores unexpected garbage in the "v=" and "k=" fields or recursively
goes through the escaping sequence until it arrives at a "correct" result.

Christian

--
You are receiving this mail because:
You are on the CC list for the bug.