Just after my last message to this list asking for assistance calculating body 
hashes, I worked out what the problem was. This is after staring at the code 
for nearly a week until my eyes hurt.

Why does this always seem to happen? It's actually quite embarrassing to 
have bothered a lot of high-powered people with a problem that ended up 
being quite trivial, and which I found within minutes of sending the message.

Turns out it was a bug in the routine I had written to strip and update line 
terminations. If the line happened to end with LWS + CRLF, the update 
routine would remove the LWS as part of its operation. Because the lines 
were changing only to the tune of a single space, it was actually quite hard to 
notice that it was happening.

Body hashes are now working correctly on all the test messages I have, for 
both simple and relaxed canonicalization.

Anyway, sorry for having troubled you - I think I have this one under control 
now. I'll go away and bury my head in the sand for a while.

Thanks again,

-- David Harris --

------------------ David Harris -+- Pegasus Mail ----------------------
Box 5451, Dunedin, New Zealand | e-mail: [email protected]
              Phone: Number provided on request only.

Real newspaper headlines from US Papers:
   "Experts sure the Dow will either rise or decline".


_______________________________________________
Ietf-dkim mailing list -- [email protected]
To unsubscribe send an email to [email protected]

Reply via email to