[comp.unix.xenix] Getty timeout problem

smith@compound.se (Bj|rn Smith) (02/15/90)

How do I specify the -t (timeout) option for getty in a Xenix system.
In a System V Unix it is specified in /etc/inittab but in /etc/ttys
I cant see where it fits in !?

--
Bj|rn Smith	Compound Systems, phone +46 8 7923689
UUCP:	 	{uunet,mcsun}!sunic!enea!comsys!smith
Internet:	smith@compound.se

perry@bluemtn.uucp (Perry Minyard (3MTA3)) (02/20/90)

In article <221@comsys.UUCP> smith@compound.se (Bj|rn Smith) writes:
>How do I specify the -t (timeout) option for getty in a Xenix system.
>In a System V Unix it is specified in /etc/inittab but in /etc/ttys
>I cant see where it fits in !?
>
>--
>Bj|rn Smith	Compound Systems, phone +46 8 7923689
>UUCP:	 	{uunet,mcsun}!sunic!enea!comsys!smith
>Internet:	smith@compound.se

Right out of the manual page for init(M):  ``getty/login" procedure:
Once init has opened a line, it executes the getty program, passing
the line mode as an arguement.   

The "line mode" is the second 'argument' in the /etc/ttys file.
An example /etc/ttys entry would be "1mtty01", (tty01 enabled at 9600baud)
"m" is the line mode, and is a pointer to the label identifier in the
in the file /etc/gettydefs file.   Below is an example /etc/gettydefs entry:


m # B9600 HUPCL # B9600 CS8 SANE HUPCL TAB3 ECHOE IXANY #\r\nn@!login: # m

Each entry in the /etc/gettydefs file has the following format:

label# initial-flags # final-flags # login prompt # next label

The man page for inittab, under initial-flags states that it:  
   Sets the initial ioctl(S) settings if a terminal type is not specified
   to getty.  The flags that getty understands are the same as the ones
   listed in tty(M).  Normally only the speed flag is required in the 
   initial-flags.  Getty automatically sets the terminal to raw input
   mode and takes care of most of the other flags.
                          ^^^^
       hmmmmm? ------------|

tty(M) makes mention of all the termio(M) flags, and nowhere in termio(M)
does it mention being able to use a flag like "-t timeout" so it looks like it
can't be done!...  Unless you could modify init, which only comes compiled, 
or if Xenix used inittab instead of ttys..  
   
You might want to play around with telinit, but I dont think that will help
you either.

If you do figure it it, I'd appreciate somebody letting me know.

-Perry Minyard, blue mountain software.


    ...!uunet!emory!bluemtn!perry

goer@sophist.uucp (Richard Goerwitz) (02/21/90)

perry@bluemtn.UUCP (uunet!bluemtn!perry)(Perry Minyard) writes:
>
>In article <221@comsys.UUCP> smith@compound.se (Bj|rn Smith) writes:
>>How do I specify the -t (timeout) option for getty in a Xenix system.
>>In a System V Unix it is specified in /etc/inittab but in /etc/ttys
>>I cant see where it fits in !?

>tty(M) makes mention of all the termio(M) flags, and nowhere in termio(M)
>does it mention being able to use a flag like "-t timeout" so it looks like it
>can't be done!...  Unless you could modify init, which only comes compiled, 
>or if Xenix used inittab instead of ttys..  

>You might want to play around with telinit, but I dont think that will help
>you either.

Telinit under Xenix has a rather limited usefulness with its getty com-
mand.  Its format is rigid, and is pretty much a translation of what you
get in /etc/ttys.  Have you tried installing another getty command?  I
have not actually done so myself, but there are some replacements float-
ing around the source groups.  You have, for instance, agetty2 in comp.
sources.misc vol. 10 and newgetty in comp.sources.unix vol. 15.  At one
point I *tried* installing them.  Newgetty has to be mucked around with a
bit so that duplicate case expressions are removed (Xenix handles cer-
tain input characters differently than most Sys V variations, and de-
fines them as such in one another header file, resulting in duplicate
case constants in a one or another of the switch statements).  I never
got very far, because the Xenix init program does not seem to like a
real System V-looking inittab file.  It might be possible to hack one
or another of these getty replacements to take its command options,
not from init, but from a secondary file somewhere.  Just an idle idea.
Like Perry Minyard said, please let us know if you develop some sort
of workaround.  SCO promises a more full integration of telinit and
enable/disable approaches in a future release of Xenix, so this is not
meant to be a criticism of them.  For now, hack we must.

   -Richard L. Goerwitz              goer%sophist@uchicago.bitnet
   goer@sophist.uchicago.edu         rutgers!oddjob!gide!sophist!goer

perry@bluemtn.uucp (Perry Minyard (3MTA3)) (02/21/90)

In article <7724@tank.uchicago.edu> goer@sophist.UUCP (Richard Goerwitz) writes:
>perry@bluemtn.UUCP (uunet!bluemtn!perry)(Perry Minyard) writes:
>>
>>You might want to play around with telinit, but I dont think that will help
>>you either.
>
>Telinit under Xenix has a rather limited usefulness with its getty com-
>mand.  Its format is rigid, and is pretty much a translation of what you

[...] 

>                SCO promises a more full integration of telinit and
>enable/disable approaches in a future release of Xenix, so this is not
>meant to be a criticism of them.  For now, hack we must.
>

Yea.. right in the manual, and inittab file, it says "It is intended that
a full integration of these two approaches will be provided in future 
versions of Xenix."

I've heard they aren't going to have any further releases of Xenix.  
I have version 2.3.3.  Hey SCO... are you out there listening?  Am I
wrong or am I right?

-Perry Minyard