xfer problems sco-sun

[Bill Vermillion]

On Thu, Apr 01, 2004 at 11:23:19AM -0800, Bill Campbell thus spoke:
> On Thu, Apr 01, 2004, Bill Vermillion wrote:
> >On Thu, Apr 01, 2004 at 12:49:34PM -0500, Jay R. Ashworth thus spoke:
> >> On Thu, Apr 01, 2004 at 12:59:53PM -0500, Tim Fischer wrote:
> >> > I'm looking for that swapcpu program, but I can't find it anywhere.  I
> >> > assume it would be in the fp directory along with *clerk, etc?
> >
> >> I dunno; the time I had to go 3b2 to SCO, I had to write one myself.

> >Hm.  The fpxfer worked just fine for the time I went from a 3B2 to
> >and SCO system.   It was really really slow, so I moved some files
> >out of the directory tree and moved them individually, the ones
> >that didn't have the byte order problem.   The headers in the key
> >files swapped just fine with xfer.   

> If I remember correctly, fpxfer ran someplace between 1200 and
> 2400 baud, even on directly connected 9600 baud serial lines.

That's about right - with one glaring exception.  When I moved
data from about 5 Xenix systems into FP on a Netware operation
the best error-free speed that could be managed was 300 bps.
I set one transfer up on Friday afternoon, and when I arrived at
the site on Monday afternoon it was still not done.

So I wrote somce screipts that swapped the file names from
prc.<name> to <name>.prc.   The one area where FP is inconsistant
between the platforms is that in both areas it it is screen.<N>
or screen.<XXX>/screen.<XXXXX>.    

> I normally did transfers using compressed tar files of the FilePro
> directories, then wrote a C program to do the byte swapping in the key
> file's headers.

Thankfully the above was pure intel to intel.  The biggest
complaint was how slow FP was on the 200MHz Pentiums compared
to the 16MHz '386 Xenix.

Bill


-- 
Bill Vermillion - bv @ wjv . com