In article <4l108i$t0j@netnews.upenn.edu>,
Michael Hicks <mwh@gradin.cis.upenn.edu> wrote:
>Olin Shivers (shivers@ai.mit.edu) wrote:
>
>: I've got the following six lines of code running on lambda.ai.mit.edu:
>
>: (bind-listen-accept-loop protocol-family/internet
>: (lambda (sock addr)
>: (write-string "Eat flaming death, earthling.\n"
>: (socket:outport sock))
>: (close-socket sock))
>: 8006)
>: which you can test by saying
>: telnet lambda.ai.mit.edu 8006
>: until I reboot my workstation.
>
>Thanks for the effort. I tried out your server, and liked it so much
>that I tried it a couple more times ...
>
>Anyway, I tried to run the above code here and it failed with the same
>server error I mentioned before. Additionally, I installed scsh on
>another machine with a different OS and architecture and got the
>same error. The same "I think I connected" behavior occurred for clients
>on both machines.
>
>I also tried to run the smtp client and the web server from
>the scsh code distribution, and got the same stuff. <sigh>
>
>A note about the systems I've tried this on:
>
>1) SPARCserver 1000 with two CPU's running
> SunOS Release 5.4 Version Generic_101945-29 [UNIX(R) System V Release 4.0]
>
> which is interesting, because the configure script thought I was running
> solaris ...
SunOS 5.4 *is* Solaris (Solaris 2.4). From uname -a:
SunOS sunny.stat.iastate.edu 5.4 Generic_101945-10 sun4d sparc
(and I know it's Solaris 2.4...). The Solaris2/FAQ
states:
1.1) What's a Solaris anyway?
Solaris(tm) is Sun's name for their UNIX-based user environment,
including the UNIX(tm) operating system, window system (X11-based),
and other stuff too.
Solaris 1.x is a retroactive (marketing?) name for SunOS4.1.x (x>=1),
a version of UNIX that is BSD-like with some SVR4 features,
along with OpenWindows 3.0.
Solaris 2.x (which is what most everybody means by "Solaris")
includes SunOS5.x, which is an SVR4-derived UNIX,
along with OpenWindows 3.x, tooltalk, and other stuff.
(See 1.5 for a chart with more info)
(I'd rather not include the chart, as it's long..).
Now, I have scsh-0.4.0 on sunny, and it dies with:
sunny% scsh
Scsh 0.4
> ,load net.scsh
net.scsh
Error: 122
"Operation not supported on transport endpoint"
#{Procedure 9398 %listen}
3
5
1>
(where net.scsh is the appropriate 6 lines of flaming death..)
I'm a bit embarrassed about that..as I sent Olin a few notes on building
scsh on Solaris back last fall..obviously I goofed in my build at least. I'll
see what I can come up with that works..
FWIW, the flaming death loop doesn't die on Ultrix..
>
>2) SGI Challenge L with four IP19 processors
> running IRIX 5.3
On the poor little Indy I have running IRIX 5.3...
vincent% scsh
Scsh 0.4
> (bind-listen-accept-loop protocol-family/internet
(lambda (sock addr)
(write-string "Eat flaming death, earthling.\n"
(socket:outport sock))
(close-socket sock))
8006)
Error: 122
"Operation not supported on transport endpoint"
#{Procedure 9389 %listen}
3
5
1>
Now, one thing I know is that Solaris changed some magic numbers, and
I'm pretty sure scsh-0.4.0 doesn't reflect this, but doesn't scsh-0.4.2?
I'll try to find time soon to look at it, but I've got a fileserver
problem and 'senior faculty' breathing down my neck...
>
>If you are interested in any other information, let me know.
>
>I did notice some interesting things during the make on the SPARC:
>
>****
>
>[solaris-defs
>[defenum-package
>./scsh/enumconst.scm
>]
>./scsh/machine/fdflags.scm ./scsh/machine/errno.scm ./scsh/machine/signals.scm
>.
>/scsh/machine/netconst.scm ..........................
>................................................................................
>.....................................
>......................................
>................................................
>
>Warning: Structure has undefined exports
> #{Structure 120 solaris-defs}
> (ip/options ip/time-to-live socket/cant-signal)
>]
>
>*****
>and on the SGI:
>*****
>
> gcc -g -o scshvm main.o unix.o dynload.o prescheme.o extension.o
> scheme48v
>m.o cig/libcig.o cig/libcig1.o proc2.o scsh/dirstuff1.o scsh/fdports1.o
>scsh
>/flock.o scsh/flock1.o scsh/jcontrol2.o scsh/jcontrol1.o
>scsh/machine/stdio_de
>p.o scsh/machine/time_dep1.o scsh/network.o scsh/network1.o scsh/putenv.o
>sc
>sh/rdelim.o scsh/re.o scsh/re1.o scsh/regexp/regexp.o scsh/regexp/regsub.o
>s
>csh/select.o scsh/select1.o scsh/syscalls.o scsh/syscalls1.o scsh/time.o
>scsh/
>time1.o scsh/tty.o scsh/tty1.o scsh/userinfo1.o -ldl -lelf -lsocket -lnsl
>-lg
>en -lm
>/usr/lib/ld:
>The shared object /usr/lib/libdl.so did not resolve any symbols.
> You may want to remove it from your link line.
>The shared object /usr/lib/libsocket.so did not resolve any symbols.
> You may want to remove it from your link line.
>
>*****
>
>The make output was about the same on both machines except for these
>differences.
>
>Are there successful networking scsh's installed under these conditions?
>I also have access to an INDIGO2 uniprocessor using IRIX 5.3, but I haven't
>tried it there yet. Any suggestions to what "configure" should be
>telling me? How about the output of "make install?" Is there one
>I can compare to for differences?
>
>Thanks in advance for any help,
>Mike
|