[TriLUG] OT: mail servers corrupting our attachments

[jonc]

We battled this a lot with Mailman. If recall correctly part of the
problem *was* the file extension. The receiving servers treated them as
text when they should have treated them as binary... multiple problems
ensued.

The problem was made worse by multipart mime encoding/decoding, which
some servers did one way, while *others* went their own way totally
ignoring any kind of standards.

The first step is to pick a nice extension that Mail servers will
respect and leave alone. Try that out and see if it helps!

Good Luck -
Jon Carnes

On Thu, 2005-08-25 at 11:41, Christopher L Merrill wrote:
> I guess this is OT, since OUR servers (BSD and Linux) do NOT have this
> problem.  We send out license keys to our customers via email.  The keys
> are encrypted files (i.e. binary).  We've always received reports of the
> license keys arriving corrupted.  But the reports are becoming more and
> more frequent.  When we zip the key and resend it (manually), the keys
> always get through.
> 
> Firstly, I don't understand why the servers are corrupting any attachments.
> I could understand if they were simply stripped...by malware filters, for
> instance.
> 
> I surmise that if the servers were actually inspecting the contents of
> the files, both look like binary files, so both should get munged.  But
> since only the non-.zip files get munged, the servers are looking at
> either the first few bytes to identify Zip files (PK...) or they are
> looking at the file extensions.  Since I've heard people recommend
> renaming files to *.zip to get past mail filters, I suspect that some
> or all of these servers are looking at file extensions.
> 
> 
> So finally...my questions:
> 
> Is it common for mail servers or malware scanners to use the filename
> extension to determine the attachment type?
> 
> If we simply renamed our licence keys to *.zip, would we cut down the
> frequency of our corrupted key compaints?
> 
> Any other suggestions?
> 
> 
> TIA,
> C
> 
> 
> 
> -- 
> -------------------------------------------------------------------------
> Chris Merrill                  |  http://www.webperformanceinc.com
> Web Performance Inc.
> 
> Website Load Testing and Stress Testing Software
> -------------------------------------------------------------------------