by rdigqd » Wed Sep 30, 2009 5:12 pm
[quote="josh"]
yes, please
If you can replicate it with a simple test text-only email
message, then the full envelop/text of that email message
would help a lot.[/quote]
...
Recent changes:
1) Over that couple of weeks, this has been happening
a LOT more
(or maybe I just didn't notice it before, until the incidents
beginning in late July that caught my eye cuz stuff I
was looking for in msg copies CC-ed or BCC-ed
to my SG id were truncated.)
2) it is happening on e-mails sent to me
(versus ones that I sent and CC/BCC-ed to myself)
After I noticed the earlier reported problem, I made
some changes in my e-mail sending procedure:
BCC-ing to _both_ the SG id indicated in the
return-addr of the email and to a non-SG e-mail id.
That way, when the msg gets back to me I can easily
see if the one which was processed thru SG is intact
or "chopped".
Now, a "new" behaviour arrives:
One of my other e-mail id's ("public id" provided by a
professional organization) is widely known and so I
auto-FWD it to an SG id, and then that SG id sends
anything that gets past the "trusted sender" scan
back to me at a non-public email id.
For unrelated reasons, some time ago I changed the
settings on that "public id" to auto-FWD to _both_
the SG id and to the non-public id.
In essence, if _two_ copies of a message are received
when I run my e-mail client and do Get-Mail, then the
msg is from a "trusted sender". If only _one_ copy arrives,
I can either ignore/delete it, or use that occurrence to
update my "trusted sender" list.
A key difference in the past weeks or so is that truncation
is happening to e-mails _sent_ to me ... that is, not e-mail
that _I_ sent which was then CC-ed/BCC-ed back to me.
I first noticed this "new" situation after 2009.09.22
To date, I have captured full and "chopped" versions
of 14 messages.
On 2009.09.10, I PM-ed you with a description and
a URL where TEXT-format of the body-and-all-headers
versions of the emails I knew about to that date could
be found.
If these more recent "hits" of this problem will help
you track down the 'gremlin' I can send them too.
==>> were the previous items I sent helpful?
==>> do you want these recent (14x2) items ??
==>> if so, thru what means and in what format??
Lemme know if you want these msgs, and how,
and I'll get them to you ASAP.
Delays from time-to-time in SG thru-put (eg: during spammer
DDOS parties) is just a cost-of-doing-business that loyal SG
users have come to acknowledge, but when the integrity and completeness of email msgs processed thru SG is unreliable
or unpredictable, that fast becomes a worry and potential
show-stopper.
.
.
A passing thought, FWIW
--------------------------------
I have no concrete rationale, only a niggling feeling,
.
.
but I wonder if a change I think was made to SG
probably back towards the start or early part of the
year might somehow be involved.
Back in that time-frame I had a bunch of emails from
"trusted senders" that were eaten or caused the counter
on several of my SG ids to increment (ie: they were not
recognized as "trusted senders").
It took me a while, but eventuallllllly, I managed to track
down and suss out that SG had, it seemed, been changed
to either _start_ using "envelope-from" info, or to use it
as superior to "From Addr" info (when both were present)
for doing the scan.
My investigation found that some folks had changed their
e-mail client, so this info was now presented in the
msgs they sent; for others, it was a change made
by their ISP.
Whatever, that "envelope-from" was now prsent when I
examined their msg headers ... and... the only way these
folks could again be given "trusted sender" treatment was
to add their "envelope-from" info to my list, ... EVEN if it
differered from their real, true, actual "return addr"
Like I said above: no solid rhyme or reason or rationale for
thinking the "truncation" gremlin might be somehow related,
but only a gut feel from past experience and given the timing
that it mebbe-is or mebbe-is-not a coincidence. HTH.
jim