steve@eleazar.dartmouth.edu (Steve Campbell) (10/11/88)
I have a foreign language application that needs to send 8-bit characters to the terminal. On our 4.3BSD system the only way I can do this is with RAW mode, and that does lots of other stuff I *dont't* want to do - like defeating input character functions and fiddling with CR/LF on output. Isn't there a way under BSD just to get 8-bit output? There should be. Steve Campbell
ed@mtxinu.UUCP (Ed Gould) (10/12/88)
>Isn't there a way under BSD just to get 8-bit output? There should be.
The bit to set for exactly this purpose is LLITOUT. It's in the
local modes word. See tty(4) for details.
--
Ed Gould mt Xinu, 2560 Ninth St., Berkeley, CA 94710 USA
{ucbvax,uunet}!mtxinu!ed +1 415 644 0146
"I'll fight them as a woman, not a lady. I'll fight them as an engineer."robert@pvab.UUCP (Robert Claeson) (10/13/88)
In article <666@mtxinu.UUCP>, ed@mtxinu.UUCP (Ed Gould) writes: > >Isn't there a way under BSD just to get 8-bit output? There should be. > > The bit to set for exactly this purpose is LLITOUT. It's in the > local modes word. See tty(4) for details. I think that disables output processing as well. It might or might not work for the particular application Mr. Gould was describing. Anyway, this does not enable 8-bit character input (I know he wasn't asking for this, but if the application uses 8-bit characters, he might want to input 8-bit characters as well). -- Robert Claeson, ERBE DATA AB, P.O. Box 77, S-175 22 Jarfalla, Sweden Tel: +46 758-202 50 Fax: +46 758-197 20 Email: robert@pvab.se (soon rclaeson@erbe.se)
guy@auspex.UUCP (Guy Harris) (10/19/88)
>> >Isn't there a way under BSD just to get 8-bit output? There should be. >> >> The bit to set for exactly this purpose is LLITOUT. It's in the >> local modes word. See tty(4) for details. > >I think that disables output processing as well. It does. >Anyway, this does not enable 8-bit character input (I know he wasn't asking >for this, but if the application uses 8-bit characters, he might want to input >8-bit characters as well). In 4.3BSD you can turn on PASS8 which will give you 8-bit character input. However, you still can't get 8-bit output without turning other output processing off. This will, presumably, be fixed in whatever future release of BSD has a POSIX-compatible tty driver. Some systems commonly thought of as BSD systems can already handle 8-bit output without completely turning output processing off, e.g. SunOS 4.0 (which has a POSIX-style - I'd say "POSIX-compatible", except that IEXTEN or whatever it's called wasn't in POSIX in time to get it into 4.0 - tty driver, with a BSD compatibility streams module atop it). Setting PASS8 in SunOS 4.0 gives you 8 bits in both directions.
loverso@xenna.Encore.COM (John Robert LoVerso) (10/25/88)
In article <17@auspex.UUCP> guy@auspex.UUCP (Guy Harris) writes: > In 4.3BSD you can turn on PASS8 which will give you 8-bit character > input. However, you still can't get 8-bit output without turning other > output processing off. > > ...Setting PASS8 in SunOS 4.0 gives you 8 bits in both directions. Unfortunately, 4.3BSD (and SunOS up to 3.5) has a bug in the implementation of PASS8. Cooked input liternal next processing `quotes' a character by setting the 8th bit. Thus, by turning on PASS8 and then inputing something innocuous like ^Vx, a read will really return 'x'|0200. This makes PASS8 somewhat useless for input of multi-lingual character sets, not to mention making the csh/sh barf (they expect 7 bits only). This isn't fixed in 4.3BSD-Tahoe. It is SunOS4.0. John Robert LoVerso, Encore Computer Corp encore!loverso, loverso@Encore.COM
guy@auspex.UUCP (Guy Harris) (10/26/88)
>Unfortunately, 4.3BSD (and SunOS up to 3.5) has a bug in the implementation >of PASS8. Cooked input liternal next processing `quotes' a character >by setting the 8th bit. Thus, by turning on PASS8 and then inputing >something innocuous like ^Vx, a read will really return 'x'|0200. This >makes PASS8 somewhat useless for input of multi-lingual character sets, Well, not completely useless; the work-around (admittedly grotty) is not to use ^V. However, what makes it much less useful is the fact that turning PASS8 on doesn't give you an 8-bit *output* path, so the system won't be able to print - or even echo - those nice 8-bit characters.... >not to mention making the csh/sh barf (they expect 7 bits only). ...since they also use the 8th bit internally for quoting characters. >This isn't fixed in 4.3BSD-Tahoe. Although it is presumably fixed in the future POSIX-compliant BSD tty driver (right, Marc?). I don't know whether they'll fix the Bourne or C shells then or not. > It is SunOS4.0. The tty driver problem is fixed (since it is completely different internally). The "sh" problem is also fixed (since it uses the S5R3.1 Bourne shell, which no longer uses the 8th bit for quoting). The "csh" problem is not fixed in SunOS 4.0 (although it may be fixed in a future release).
marc@monet.Berkeley.EDU (Marc Teitelbaum) (11/02/88)
It's true that PASS8 and literal next (or "\") don't live a happy co-existance. To have both function requires a bit more work than was probably affordable at the time. The next release of BSD fixes the problem. Marc ------------------------------- Marc Teitelbaum +1-415-643-6448 457 Evans Hall Computer Systems Research Group, CSRG / DEC University of California Berkeley, CA 94720