🚀 go-pugleaf

In article <D0CLHB.775@zoo.toronto.edu>,
Henry Spencer <henry@zoo.toronto.edu> wrote:
>In article <D0BGKF.2qr@telly.on.ca> evan@telly.on.ca (Evan Leibovitch) writes:
>>What I've found is that a significant amount of incoming news batches are
>>garbled. Brought over compressed, they won't uncompress cleanly; when it's
>>forced, the first half of each batch looks fine, but as one goes further and
>>further in the file it becomes more and more garbled until it's unreadable.

>I'd suggest making sure by UUCPing some files back and forth a few times,
>using the same paths and protocols employed for news, and seeing if they
>make it intact.  Some experiments with more suspicious protocols, and
>possibly encapsulations with checksums (such as the bencode/bdecode stuff
>that ships with C News), would also seem in order.

I believe that there is some sort of problem with Taylor's "i" protocol at
higher window and packet sizes.  If I remember rightly, Evan should try
forcing the window size below 16 (it can go to 31).  I think there may be a
comment to that fact in the Taylor distribution somewhere.
-- 
Chris Lewis: _Una confibula non sat est_
Phone: Canada 613 832-0541
Latest psroff: FTP://ftp.uunet.ca/distrib/chris_lewis/psroff3.0pl17/*
Latest hp2pbm: FTP://ftp.uunet.ca/distrib/chris_lewis/hp2pbm/*