Bugs found in the IRIX operating system

I’m trying to compile more random X programs, and for xcpustate, this
is the error which is reported:

cfe: Error: /usr/include/sys/sysmp.h, line 161: prototype and non-prototype
declaration found for sysmp, ellipsis terminator not allowed
 extern int sysmp(int, …);
 ———————-^

   Looking at /usr/include/sys/sysmp.h, I saw that that does contain
such a line. This is 5.2, and we just installed it 2 weeks ago. Is
this the way things should be, and if so, why does the cc die?

   Thanks in advance.

Nathan Mates

–
* Nathan Mates nat…@cco.caltech.edu * MSC #850, Pasadena CA 91126 *
* Ftp humor archiver: ftp to cco.caltech.edu, look in pub/humor     *
* Largest collection of Clinton Jokes, other canonical lists        *
* Support Twilight II, the best screen saver for the Apple IIGS!    *

I’ve just inherited an Indy and immediately upgraded it to 5.2.  I am trying
to use it as a router using PPP through the serial port.

The two serial ports, /dev/ttyd1 and /dev/ttyd2, have a serious baud rate problem,
and will not work at all.  Using a ‘scope I figured out that the baud rates
are not as advertised, i.e. 9600 baud comes out around 6000 baud, etc.

By poking the table (zduart.baud_table in the kernel) i figured out that the
correct entry for 9600 baud is ’7′, not the ’10′ as computed by the kernel at
compile time.

I assume that my particular serial port clock was not taken into account in this
release.  But there are further problems I haven’t figured out:  setting the
table entry fixes character output, but the machine doesn’t recognize incoming
characters.  I will ponder further in an older listing of the driver.

Bill Cheswick
c…@research.att.com

Bell Labs

Last week, I upgraded an Indigo^2 XZ from IRIX 5.2 beta 2 to IRIX 5.2.
I started the installation by using mkfs to clear out the / and /usr
partitions. The installation went smoothly, but after the system was
rebooted the soft link from /bin to usr/bin disappeared!

I was able to telnet in and re-create the link, but it would disappear
again each time we rebooted.

The problem seems to have cured itself just as mysteriously as it appeared.
When the system was rebooted this past weekend, the link remained.
This seems to indicate the problem goes away if /bin is ‘old’ ??

The same problem happened before, when IRIX 5.2 beta 2 was first installed
on the system, but at that time I thought it must have been a quirk in the
beta release.

So, I have seen this twice on the same Indigo^2 XZ, but never on the
10 Indy systems I have installed. All of them have separate / and /usr.

Has anyone else seen this problem?
—
Steve Kotsopoulos  P.Eng.                         st…@ecf.toronto.edu
Systems Analyst,  Engineering Computing Facility, University of Toronto

        I can not contact to any ftp sites.  Please look into this.

                                                Thank you,

                                                James.

our lpd is dying every time is run, with the following error message.
What does it mean?

May 16 15:21:59 suerte lpd[236]: setsockopt (SO_DEBUG): Invalid argument

If anyone can tell us what this means, and how to fix it, we’d be very
appreciative.

Your,

Chris Lee

     We have an Indigo2 R4000 100MHz Extreme running IRIX 4.0.5H.  I got
mosaic from the SGI ftp server and put it on our machine.  Most things
seem to work ok, however, it I sellect ANY audio hypertext, the machine
plays the entire audio and then propomptly DIES and while it is dumping
a system core file, a LOUD tone is played until the reset button is pressed.
Has anyone had this happen to them?

  Brent L. Bates                                Phone:(804) 864-2854
  M.S. 361                                        FAX:(804) 864-8469
  NASA Langley Research Center
  Hampton, Virginia  23681-0001
  E-mail: blba…@aero36.larc.nasa.gov or B.L.BA…@larc.nasa.gov

pressing "Save" in the Mail Aliases window
rewrites ~/.zmailrc, stripping any user-added
comments.

this deletes valuable information from the file.

seth

–

 princeton CS dept, 35 olden st, princeton NJ 08544-2087 USA
 tel 609 258 3946 _________________________ fax 609 258 1771
 web site:  http://www.cs.princeton.edu/staff/seth/home.html

in the latest zmail, pressing Include in the Compose
pane always inserts the last-read message at the end
of the composed message, regardless of cursor position.

this is annoying.

seth.

–

 princeton CS dept, 35 olden st, princeton NJ 08544-2087 USA
 tel 609 258 3946 _________________________ fax 609 258 1771
 web site:  http://www.cs.princeton.edu/staff/seth/home.html

I have encountered a problem with the use of the logical function
"EQV" which appears not to work when used with variables declared
as logical*1.
The following program when compiled in the ways suggested will
illustrate the problem

I would welcome comments

Peter Bladon

(cba…@vaxa.strath.ac.uk)
Phone/fax +44 (0)41-776-1718

—————————-cut here——————————
      Program testeqv
C  To test the action of the logical ‘eqv’ operator
C  which does not work with logical*1 variables!!!

C  There are two statements where the logical variables
C  a, b, c, and d are declared in two alternative ways.
C  One and only one of these statements must be "active",
C  the other one must be commented-out.

C  When this small program is compiled with the
C  variables a,b,c,d declared as logical (*4) in the
C  normal way, then the correct values of c and d
C  are computed.  When the logicals are declared as
C  logical*1, then the operator EQV does not compute
C  correctly.  The alternative (longwinded) statement
C  using AND and OR operators does give the correct
C  results with both the logical*4 and logical*1
C  alternatives.

C  Identical behaviour was found using the 3.5.1 and 4.0.1
C  versions of the fortran compiler, under IRIX 4.0.5
C  and 5.2 respectively.

      logical    a,b,c,d
C      logical*1  a,b,c,d

10    continue
      write (6,6000)
      write (6,6001)
      read  (5,5001) a,b

      write (6,6006)
      c = (a.eqv.b)
      write (6,6002)  c
      if (c) then
         write (6,6003)
      else
         write (6,6004)
      endif

      write (6,6005)
      d = ((a.and.b).or.((.not.a).and.(.not.b)))
      write (6,6002) d
      if (d) then
         write (6,6003)
      else
         write (6,6004)
      endif

      go to 10

5001  format (2l1)
6000  format (‘——————————————————’)
6001  format (‘Enter as "tt","ff","tf", or "ft" the values of A and B’)
6002  format (l1)
6003  format (‘A and B have the same logical state’)
6004  format (‘A and B do not have the same logical state’)
6005  format (//’Use of AND and OR operators’)
6006  format (//’Use of EQV operator’)
      end

—————————-cut here——————————