term bug report

[Bill C. Riemers]

'Francois-Rene Rideau'

>    Here's the result of linecheck 0 17 19 on my modem connection.
> (btw, linecheck is documented to output all this information to
> stderr, whereas it actually uses linecheck.log ...)

I can't find the place where it claims that the output goes to standard
error.  The only mention I could find is in linecheck.1:

      Linecheck must be run on both systems the same way term is run.  That
      is, the stdin and stdout should be directed to the serial port while
      the stderr goes to the log file linecheck.log.  Remotely you can type
      ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

This manual page is very poorly written.  Sometime, I should write a new
one.


> I gzip-ed and uuencoded those, as they are binary and may not pass
> well mail gates (I couldn't tupload the linecheck.remote).
>    My line is sevenbit, and I escape 0 17 19 on both sides, and 43
> on local side (to not confuse modem; but the modem waits 1 second
> to interpret this like a special sequence; can it be that term
> stays silent more than 1 second after sending three pluses +++ ?
> I don't think so, but you may tell me).
>    You see that the local linecheck has troubles with the high bit
> from the even parity, while the other end has no problem. Kermit
> leaves the line 7 bit -parodd when leaving, so this is a linecheck bug
> (it does some stty's it shouldn't).

Does look bothersome.  I just added the option --seven_in to linecheck,
so if you grab term-2.2o, you should be able to use this to force linecheck
to ignore the parity and give you a deceint report.

>    Term *almost* works while escaping 0 17 19, but tuploading binary
> afiles show that some high bits are moved from one byte to the next
> and sometimes other bits (a packet of errors every 10K or so).

OK.  I spent the whole day playing around with this.  I found that the problem
depended on the size of the packet.  So my previous tests, under ideal 
conditions always showed that the seven bit conversion was working...  So,
I've corrected the problem with the high bits.  I'm hoping that the other 
errors, are either indirectly due to the same problem, or something you 
will now be able to find with linecheck.

                                      Bill


-- 
<A HREF="     http://physics.purdue.edu/~bcr/homepage.html                  ">
<EM><ADDRESS> Dr. Bill C. Riemers,  bcr@physics.purdue.edu </ADDRESS></EM></A>
<A HREF="            http://www.physics.purdue.edu/                         ">
<EM>            Department of Physics, Purdue University             </EM></A>