your mail

• Messages sorted by:[ date ][ thread ][ subject ][ author ]

Hi,

John Wehle wrote:
> Gert Doering wrote:
> 
> > How is the istrip flag set when ttymon forks off "login"? (please
> > do a "stty -a" from a different window)
> 
> Here is the entire output from stty -a from when I run * cu * out
> one modem logged into the other modem.
[..]
> Call out modem (being used by * cu *):
> 
>  speed 38400 baud; 
[..]
>  rprnt = ^r; flush = ^o; werase = ^w; lnext = ^v;
>  parenb -parodd cs7 -cstopb hupcl cread -clocal -loblk crtscts crtsxoff -parext 

 ^^^^^^^^^^^^^^^^^^^
Solaris' man page is wrong.

> Call in modem (after having logged in):
[..]
>  -parenb -parodd cs8 -cstopb hupcl cread -clocal -loblk crtscts crtsxoff -parext 
>  -ignbrk brkint ignpar -parmrk -inpck istrip -inlcr -igncr icrnl -iuclc 
        ^^^^^^^

Which will take care of the 7E1 call...

> > What program do you use to call out? "tip"? For tip on FreeBSD, I know
> > for sure that it defaults to 7E2, but I have no idea about Solaris.
> 
> I guess I should of emphasized * cu * better when I mentioned it in
> item 2 of my earlier message. :-)

Hmmm. Consider that I receive about 10-30 mgetty mails every day, and
bear with my waning brain...

> > This may expose a "problem" on the calling side, masked by ttymon
> > stripping the parity bit.
> 
> Sure from a purely technical standpoint the calling side should exactly
> match the answering side. 

Definitely.

> From a practical standpoint it works with ttymon and not with mgetty. 

*From a practical standpoint, it works on each system in the world
except Solaris 2.5.1.

That gives?

> Also it is easier to have ISTRIP default
> to on rather then having to educate * each * dial in user on the
> proper technical setup, especially when that setup isn't the default
> for the software that they are using.

Well, on the other hand, just setting ISTRIP may cause other, hard-to-find
problems later. Like "if I log in with mgetty, I am not able to type
any special characters (german umlauts come to my mind) anymore".

> BTW, it's interesting to note that the following appears around line
> 394 of logname.c:
> 
> ch = ch & 0x7f;      /* strip to 7 bit */
> 
> Which effectively emulates ISTRIP.

Yep. I'm not really sure why I put it there, but most likely just for
that purpose.

IIRC, SCO's login program strips 8th bit as well, but then properly resets
istrip so the shell has an 8-bit clean data path.

[..]
> > (On all systems I have access to - SCO, Linux, AIX, FreeBSD - I can
> > login fine without "istrip")
> Interesting.

Well... -- actually I should phrase this differently: all programs that
I happened to use default to 8N1, except "tip", which I know to default to
7E1 and which I avoid for that reason.

Which gives a completely different picture, I admit.

> > Which might mimic ttymon's behaviour.
> A very good possibility.

Yup.

Concluding: I am not really willing to add this change, because I expect
it to cause more problems at other places, and because I think that
software that uses 7E1 should be extinguished as fast as possible, instead
of being cared for.

But since everybody has the source code, this doesn't really matter, as
everybody really needing this can easily change it.

I will add a note to the manual, however, explaining that problem.

gert

-- 
USENET is *not* the non-clickable part of WWW!
           //www.muc.de/~gert/
Gert Doering - Munich, Germany      gert@greenie.muc.de
fax: +49-89-3545980     gert.doering@physik.tu-muenchen.de
.