info-vax@ucbvax.ARPA (06/23/85)
From: John M Sellens <jmsellens%waterloo.csnet@csnet-relay.ARPA>
Under VMS 4.1, I want to write a C program that can obtain single character
input from the keyboard, without any editing, and without waiting for the
user to press return. The character should not be echoed on the screen,
my program will do all the necessary echoing.
I am using the SMG$ routines for screen output. I tried using
SMG$ENABLE_UNSOLICITED_INPUT to enable a routine that will be activated
each time a key is pressed. I tried to make it reasonably intelligent,
but I seemed to have very little luck. One problem that I ran into was
that I couldn't convince it to accept an escape character without a
following escape sequence (our program uses escape to indicate "undo"
or "stop what I'm doing").
I would also like to be able to inquire whather or not a character or
characters is waiting (or alternatively, do something like a non-blocking
getchar()).
There is only a small amount of VMS expertise here a Waterloo, so asking
around hasn't been too successful. I will greatly appreciate any help
that you can offer me.
Thanks!
John
UUCP: {decvax|utzoo|ihnp4|allegra|clyde}!watmath!jmsellens
CSNET: jmsellens%watmath@waterloo.csnet
ARPA: jmsellens%watmath%waterloo.csnet@csnet-relay.arpainfo-vax@ucbvax.ARPA (06/24/85)
From: stew%lhasa@harvard.ARPA I have long wished that dec would make crmode() and nocrmode() work by setting a flag in the C runtimes which would make them do single character qio's. Is this not easy? Would this not work? To get escapes through, you have to turn off the TT$M_ESCAPE bit in your terminal's basic characteristics word, typically with the IO$_SENSEMODE and IO$_SETMODE functions of SYS$QIO. A good example of this can be found in the support for vms that I wrote for the new version of C-KERMIT, which is available by anonymous ftp from CU20A. This implements, entirely in C, the necessary calls to system services to do single character IO and other useful things like typeahead sensing. It is in KER:CKVTIO.C. You may be able to use this directly, or at least it will show you how to do such things from C. If I remember correctly, it does not turn off TT$M_ESCAPE, it uses pasthru mode instead, but perhaps this is what you want anyway. Stew
info-vax@ucbvax.ARPA (06/25/85)
From: stew%lhasa@harvard.ARPA Whoops! I said that kermit is available from CU20A via ftp. I meant to say CU20B (ecc chip blew in my brain). It is also available from CUVMA via bitnet, from the columbia computer center on a tape, and from okstate via uucp. If you smother him with thankyous and goodjobs, you might even get INFO-KERMIT-REQUEST@CU20B to mail you you the distribution info or even the CKVTIO.C file itself (last resort!). Stew
info-vax@ucbvax.ARPA (07/16/85)
From: ecsvax!jerryr%mcnc.csnet@csnet-relay.ARPA I am wondering if the SMG$READ_STRING routine wouldn't be sufficient for this. I don't think it will give you character waiting information, but from what I've read (unfortunately not enough time to test it yet) it appears that this routine uses qio's since you can give it max lengths, io modifiers and custom terminator blocks. You should be able to set up a terminator mask which indicates that all ASCII characters are valid terminators, and receive which terminator in the separate TERMINATOR-CODE parameter. There are some rather cryptic limitations for its use listed under the DISPLAY-ID parameter in the RUN-TIME LIBRARY REFERENCE which need more investigation on my part. But I'd be happy to have someone experiment and report the results. In fact, I'd be interested in any comments regarding the absence of usable raw(), etc. routines in the VMS Curses package that comes with VMS "C" V.2. I'd think (and was going to test this out) that one could use the above routine for standard input, with the use of a few others to set the terminator masks and the read modifiers in pre-defined fields which then would be used by SMG$READ_STRING. raw() would set the terminator mask suggested above as well as set the IO$M_NOFILTR read bit, noraw() back to defaults, echo() reset the IO$M_NOECHO read bit, noecho() set it, etc. I'm suspecting that it's not as simple as it seems, but then again *knowing DEC...* Jerry Robinson UNC School of Medicine UUCP: ...mcnc!ecsvax!jerryr