[dkim-dev] Should granularity (g= tag) be case sensitive?

[Jason Long]

On Tue, 2008-04-08 at 16:29 -0400, Jason Long wrote:
> Well, I checked this point on various responders by constructing a message
> that tests this specific issue... Here's what I found:

Ok, there was a problem with my test message, which led to some false
results. (The From header didn't match the i= tag, and it turns out that
some of the responders checked the granularity against the From header
instead of the i= tag. A different problem :]..)

Here are the updated results:

the following responders ARE case-sensitive:
    blackops.org - responded with (granularity mismatch)
    sendmail.net - signature confirmed BAD
    elandsys.com - responded with (granularity mismatch)
    altn.com - responded with (:DKIM_SELECTOR_GRANULARITY_MISMATCH)
    verifier.port25.com - responded with (key granularity "jl*g" doesn't match local part "JLong")


So I think we're all in agreement that g= is case-sensitive. :). I'll
update my implementation.

Jason