Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP Path: utzoo!linus!decvax!cca!ima!ism780!isc.net From: isc.net@ism780.UUCP Newsgroups: net.micro.pc Subject: PC/IX - (nf) Message-ID: <392@ism780.UUCP> Date: Wed, 29-Aug-84 00:04:33 EDT Article-I.D.: ism780.392 Posted: Wed Aug 29 00:04:33 1984 Date-Received: Thu, 30-Aug-84 12:11:16 EDT Lines: 143 #N:ism780:12800007:000:5906 ism780!isc.net Aug 27 14:53:00 1984 The following is a response from your good friends at Interactive to some recent gripes/problems voiced on the net about PC/IX. The original notes are prefixed with ">"; our comments are indented. Note that the "next release of PC/IX" should be available quite soon. >There definitely seem to be some port related problems with PC/IX. Not >only the run-PC/IX-after-MSDOS problem Lee Merrill spoke of, but other >somewhat mysterious conditions where tty0 or tty1 will get into a wedged >state and refuse further commands. An attempt to shutdown at that point >will cause shutdown to sit quietly for close to a minute, print the >message "init: something won't die%s", and then MAYBE finish the shutdown. This behavior is the result of two bugs, both of which have been fixed for the next release of PC/IX. The first is a hardware bug in many of the async cards used in PCs; I believe this has been mentioned before on the net. We have altered the kernel to circumvent the problem. The second is a bug in the kernel shutdown procedure which we have fixed. >One problem that is still plaguing me is a behavior problem with certain >utility programs *only* when used from tty0 or tty1. They're fine from >the local console. Seems that any extra characters entered after >the final c/r of the password prompt and before the system types to you >when using login, su or passwd will kill the port. You can still send >to it and the machine will respond (you'd only know that if it were >right next to you), but no output of any kind comes from the port. Once >you have disconnected, the port is in its famous wedged state and nothing >at all seems to clear it. Only solution at that point is a one minute >shutdown (described above). IBM says its because PC/IX isn't running on >an IBM! What help. > >For example: login: yourid password: yourpass >will kill it every time. Anything entered at the point of the second c/r >after your password, but before you receive more output from PC/IX on >any of the three utilities I mentioned will do it. > >I'm using the MicroLine BabyBlue II board for both ports. It works >flawlessly in MS-DOS. > >Any suggestions? (please mail me directly, and I'll summarize if necessary) > >--Bill Blue {sdcsvax, sdchema, ihnp4}!bang!crash!bblue We cannot duplicate this behavior. It is quite possible that it is also due to the bad async card. Sorry, but that's the best we can tell you! ------------------------------------------------ >I have been "experimenting" with PC/IX since May 1, 1984. My problems >are/were: > >1. I was only able to install PC/IX after IBM replaced EVERY board >in my PC XT. restore -x didn't like my computer! Your hardware was a lemon. Sorry! >2. I could only get connect to work after spending several days >with local "gurus". When I connect to a DEC VAX 11/780, I >can't use vi, but with SIMTERM, etc. I can. Termcap problems? The configuration files associated with connect (/etc/sites and /usr/lib/INnet/connect.con) are really not that hard to figure out, although I suspect somewhat better documentation would be helpful. There was a bug in the connect "talker" program which prevented programs such as vi and e from operating properly when connected. It is fixed for the next release of PC/IX. >3. I had to rm /dev/lp and ln /dev/lp1 /dev/lp to get the printer >to work. We distribute the system with /dev/lp linked to /dev/lp0, which seems the most logical choice. You must have put your printer card in the second slot instead of the first, which obviously means reconfiguring the devices (or /etc/qconfig) accordingly. >4. I can't run more than one or two processed without crashing the >system. > mm filename > temp & > ps -ef >( the above will kill my system, ungracefully) There was a highly unfortunate bug in the shell which went out with PC/IX. It will manifest itself in relatively random ways whenever the shell is running on a heavily loaded system. I suspect that if just the two processes above killed your system, that you must be configured with a very small amount of memory (perhaps the minimal 256K?). The bug is fixed for the next release of PC/IX. However, you can (and probably should) circumvent the problem, at the cost of some efficiency, by executing the following commands: mv /bin/sh /bin/sh.old cp /bin/sh.old /bin/sh chmem +60000 /bin/sh >5. I can't print on standard output the graphics character set while >in connect mode ( e.g., echo "\0333" won't generate graphics ) Same bug as item 2., I believe. >6. usr/games/bj doesn't work! Shocking! Actually, this is pretty surprising, as the level of testing that went into the PC/IX release should have caught this easily. We will look into this bug. >7. You cant type in the password after the login UNTILL the password >prompt appears This is standard UNIX System III. An ioctl with a TCSETAF command is issued just prior to printing the password prompt, in order to discourage people from typing in their password while echoing is still enabled. >8. INed ( e editor ) is the SLOWEST editor I have ever used. It can't >keep up with my typing speed which is 10 words/minute on a good day. Again, you must be configured with VERY little memory. On our 640K systems, e keeps up with anything we can give it; indeed, it performs about as well as it does on a lightly-loaded VAX. Also, keep in mind that while some of the function keys may seem slow, they often perform complex tasks with a single keystroke, so the illusion of sluggishness can sometimes manifest itself. >The bottom line is that I am going to shelve the product and try >Santa Cruze's XENIX. No comment. Have fun. >David S. Green Bell Labs mhuxi!dsg phone 201-564-4468 --isc.net (ima!ism780!isc.net) INTERACTIVE Systems Corp.