Rachel Willmer wrote:
I think the spec does define this. It defines all of those items that
you mention in the grammar in section 11.3. Any OpenPGP-compliant
application should be able to write such a message in such a way that
any other OpenPGP-complaint application can read it, or one or both of
the implementations aren't OpenPGP compliant.
What the spec doesn't define is whether an application must process an
"OpenPGP Message, OpenPGP Message" -- two messages concatenated
together. There is an assumption in the spec that a single stream
contains a single message,
If that is intended, it should be explicitly stated, IMHO.
Just in an effort to get closure on this issue, I see it as
"undefined."
That is, it is strictly undefined whether messages can be
concatonated.
This makes sense to me, as how the app packages these
messages up and ships them is ... undefined, including in
the trivial case of files containing concatonations.
And if that is the case, then ok, I can get back to writing test
cases, and forget about writing RFCs.
If it is stated somewhere and I've missed it, can you point me at the
reference please?
Now, maybe the spec should say that the issue is undefined,
I'm not sure.
I'm also not sure whether we can even add a change like that
at this late stage?
iang