Hello mark,

On Sunday December 02 2007 13:09, you wrote to me:

 ml>> this happens, most of the time, when an echomail bundle has more
 ml>> PKTs added to it and there's already a partial download of it on
 ml>> your system...

 MvdV>> If that is the case, it implies a configuration error on the
 MvdV>> sender's side doesn't it? The tosser shouldn't touch mail
 MvdV>> packets that are being send.

 ml> it is possible... it wouls depend on the tosser and mailer software
 ml> being able to talk to each other, though...

Of course. There should be some intertask communication or steps should be
taken to insure the tatsks never run simultaneously.

 ml> in my hybrid environment, i have to choose between FD and Binkley
 ml> style semaphore languages because my tosser (in fact, none that i'm
 ml> aware of) doesn't speak both at the same time... add to that that i
 ml> have to use another tool to move mail between the two for sending and
 ml> things get more complicated...

Yes, a hybrid FD/BSO environment would comlicate things. Reason why I avoid it.

 ml> it could be as simple as the connection failing on the one file and
 ml> then another PKT being added before a reconnect to complete the
 ml> original transfer... in this case, you have part of the file, the
 ml> connection breaks, a PKT is added, you reconnect, you attempt to
 ml> resume the download of the partial file and you get this because the
 ml> sending file is now no longer the same as the piece you have due to
 ml> the addition while the connection was down...

Of course when the synchronisation between sender and reciever is lost.
Many things can happen.

 ml> make sense?

Yes it does and it seems to be what has happened. My Irex received a file
and something happened before the transfer was completed. So they were out
of sync and the protocol was not intelligent enough to restore the
condition on its own.

The problem was finding the corrupted file. I never expected it to reside
in a subdirectory of the hold directory...


Cheers, Michiel

--- GoldED+/W32-MINGW 1.1.5-b20070503