From CBFatMIT-MC Thu Feb 26 00:00:00 1981 From: CBFatMIT-MC (Charles Frankston) Date: 26 February 1981, 00:00 Subject: H19's on ITS Message-ID: Date: 25 Feb 1981 2031-EST From: EBM at MIT-XX :TCTYP H19 would actually be useful if ITS would pad the ins/del line sequences. Our experience indicates that 17ms. per line inserted/deleted is required at 9600 baud, and no padding is needed at 1200 and lower. We just pad proportionately in between. (I.e., 4800 baud => 8.5ms pad, etc.) I don't think this will screw anybody very much, and would certainly please H19 users, though having ins/del char would be nice, too. Getting that would require adding a terminal type, though (sigh). Anyway, is anybody out there willing the hack TS3TTY to put in the padding? You message makes no sense, but I implemented it anyway. I don't think the H19 is 17ms PER line inserted or deleted, and I don't see why it only takes 8.5ms if the UART is running at 4800 baud. I implemented a padding of .625 ms per line moved at speeds higher than 1200 baud, and no padidng at 1200 baud. This is only in the source of course, some poor sucker who brings up the next ITS will have to debug it. From EBM Wed Feb 25 00:00:00 1981 From: EBM (EBM) Date: 25 Feb 1981, 00:00 Subject: H19's on ITS Message-ID: :TCTYP H19 would actually be useful if ITS would pad the ins/del line sequences. Our experience indicates that 17ms. per line inserted/deleted is required at 9600 baud, and no padding is needed at 1200 and lower. We just pad proportionately in between. (I.e., 4800 baud => 8.5ms pad, etc.) I don't think this will screw anybody very much, and would certainly please H19 users, though having ins/del char would be nice, too. Getting that would require adding a terminal type, though (sigh). Anyway, is anybody out there willing the hack TS3TTY to put in the padding? ------- From JLK at MIT-MC Mon Feb 23 00:00:00 1981 From: JLK at MIT-MC (JLK at MIT-MC) Date: 23 Feb 1981 00:00 Subject: MUX, ITS, and Tektronix wierdness Message-ID: After painstaking investigations, I have determined that ITS does some very wierd things on Chaos Supdup connections with a TCTYP of %TNTEK. There seems to be a bug where it ends up throwing away 2 packets completely. The only thing I change in my icp packet is the tctyp field (to say, teleray or somesuch) and it works fine. This is repeatable. Nothing in my software depends on the tctyp. The observed wedged state is that 2 packets are missing and ITS refuses to resend them, even though ITS agrees that they have not been acknowledged. Its hard for me to see how this dependency could exist (output reset wierdness maybe?) From SLH Mon Feb 23 00:00:00 1981 From: SLH (Stephen L. Hain) Date: 23 FEB 1981, 00:00 Subject: No subject Message-ID: Apparently a disk error has trashed my mail file, on ty's directory Laurel From RWK at MIT-MC Thu Feb 19 00:00:00 1981 From: RWK at MIT-MC (RWK at MIT-MC) Date: 19 Feb 1981 00:00 Subject: No subject Message-ID: The IMP seems to be having troubles. From CSTACY Thu Feb 19 00:00:00 1981 From: CSTACY (Christopher C. Stacy) Date: 19 FEB 1981, 00:00 Subject: No subject Message-ID: This is really weird; I must be overlooking something simple. When I connect to any ITS from DCECTIP or PENTIP (which are the only ones I tried), funny things happen to my connection. These funny things are consistant, and do not happen when I use this same terminal/modem (300bd,btw) on non-net systems, or on a Unix I tried which is on the net. These funny things also do not happen on an AI Dialup. The funny things: the system pases in the middle of the SSTATUS and hangs for a few secs, sometimes I have to control-G out to pword toplevel. Various characters are received and echoed wrong. When I first saw this I thought that my terminal or modem was dropping bits, but it is now clear that the problem only occurs connecting to an ITS over a TIP. To wit, s := ^S n := ^N but N:=N a := ^A space := nothing? (sounds like its xmitting it, but no echo) atsign := same as space I cant imagine whats wrong. Help? Thanks, Chris From EB Mon Feb 16 00:00:00 1981 From: EB (Edward Barton) Date: 16 FEB 1981, 00:00 Subject: ai paper tape punch Message-ID: MSG: DEAD PUNCH the ai papertape punch, hasn't worked in years and has miniscule to disappearing chance of ever being repaired. If the punch is really permanently losing, OPENs on PTP: should be made to fail with some appropriate error code. From RWK Sat Feb 14 00:00:00 1981 From: RWK (Robert W. Kerns) Date: 14 FEB 1981, 00:00 Subject: File clobberage Message-ID: Page 0 of AI:.MAIL.;^Q ^Q LIST ^Q QUEUE was bashed with random mail (what looked like a random mail file). I renamed the file .MAIL.; LIST CLOBRD, and copied the old set of (MSGS REMIND MASTER) to .SAVE *, and did a MAKQFI. It seems to be working OK now. From ED at MIT-ML Thu Feb 12 00:00:00 1981 From: ED at MIT-ML (ED at MIT-ML) Date: 12 Feb 1981 00:00 Subject: Index to History Message-ID: Those of you who want to experience the old days might wish to run ML:SYS;TS MONIT, which for some reason does not have symbols nor the correct start address (1300). It is pretty badly broken, but great fun to play with nevertheless. Too bad we don't have any working microtapes to really hack it with. From MOON5 Thu Feb 12 00:00:00 1981 From: MOON5 (David A. Moon) Date: 12 FEB 1981, 00:00 Subject: No subject Message-ID: I would kind of like to know why single pages of the inquir data base have been clobbered to what looks like pages out of other files three times in the last two days, on AI. I don't think anyone maps it in read-write, so it's hard to see how to blame it on Howard's memory. Anyone have any ideas? Anyone observe any other file clobbering problems? From MOON Sun Feb 8 00:00:00 1981 From: MOON (David A. Moon) Date: 8 FEB 1981, 00:00 Subject: No subject Message-ID: I tried to take AI down using the DOWN/KILL command in LOCK so that I could bring back a fixed disk. The command fails to do anything, it just gives an _ prompt. From RWKatMIT-MC Sun Feb 8 00:00:00 1981 From: RWKatMIT-MC (Robert W. Kerns) Date: 8 February 1981, 00:00 Subject: No subject Message-ID: KMP at MIT-MC 02/07/81 05:20:27 To: (BUG ITS) at MIT-MC MC has been randomly resetting its arpa connections this evening. I don't know why. I didn't notice anyone running LOCK. I guess it does it on its own sometimes. Usually the net goes off, on, off then there's a bit of a gap (as much as a minute or two) and then back on again... There was a system message a while back informing everyone that the TIP and IMP were being moved. Perhaps this is what was doing it? From MP Sun Feb 8 00:00:00 1981 From: MP (Mark A. Plotnick) Date: 8 FEB 1981, 00:00 Subject: long leaders Message-ID: I just tried to get onto ML from a TIP. It wouldn't connect me, saying that the Host didn't support extended leaders. Is this something new? I thought extended leaders were required since Jan. 1. From JONL Sat Feb 7 00:00:00 1981 From: JONL (Jon L White) Date: 7 FEB 1981, 00:00 Subject: Crunch on AI:SYS;, part II Message-ID: ITSDFS 4 ==> SYSENG; SAIDFS 17 ==> SYSENG' Both are "old" files. current versions still on SYS; From JONL Sat Feb 7 00:00:00 1981 From: JONL (Jon L White) Date: 7 FEB 1981, 00:00 Subject: Directory space crunch on AI:SYS; Message-ID: I moved ATSIGN ODEVIC and ATSIGN OHACTR to SYSBIN;, since one hadn't been referenced in over 2 years, and the other in almost a year. From KMP at MIT-MC Sat Feb 7 00:00:00 1981 From: KMP at MIT-MC (KMP at MIT-MC) Date: 07 Feb 1981 00:00 Subject: No subject Message-ID: MC has been randomly resetting its arpa connections this evening. I don't know why. I didn't notice anyone running LOCK. I guess it does it on its own sometimes. Usually the net goes off, on, off then there's a bit of a gap (as much as a minute or two) and then back on again... From CBFatMIT-MC Sat Feb 7 00:00:00 1981 From: CBFatMIT-MC (Charles Frankston) Date: 7 February 1981, 00:00 Subject: confused about char ins/del Message-ID: Date: 7 February 1981 00:52-EST From: Earl A. Killian The C100 would probably work fairly well on ITS without using CRTSTY, just as it does on XX. This is to say that while it is completely trivial to construct legitimate sequences of requests to the virtual terminal support that will lose, those sequences don't tend to come up while editing in EMACS. With CRTSTY, however, it is extremely common for CRTSTY to produce such sequences as the result of other unrelated optimizations that it is doing. As a result CRTSTY definitely cannot use insert/delete character unless your terminal supports the set blank characteristics and set insert mode commands. On ITS it's not so clear, but I think having the default be off is still the right thing; users with the proper microcode or bravery can always try :TCTYP +%TOCID (I hope this works). :TCTYP +%TOCID does not work, in the sense that ITS has no entries for char ins/del on a C100 whether on by default or not. I was going to put entries in, but unfortunately, the C100 uses Escape-Null to exit insert mode. I can't put the null in an ASCIZ string which is what the table entries consist of, and I don't know TS3TTY well enough to figure out a good kludge.. From CBF Sat Feb 7 00:00:00 1981 From: CBF (Charles Frankston) Date: 7 FEB 1981, 00:00 Subject: TS3TTY terminal padding Message-ID: Remember the feature to vary a terminal's line insert/delete padding in proportion to the number of lines left to be moved on the screen? Anyway, it was never exercised in any terminal defintions. EAK and I tried patching the entry for a C100 with PADCR=1 on ML and found it seems to work fine. However, when calculating the values for a 1200 baud C100 or for a Teleray 1061, we find that they would really want to be 1.7 and 2.5 milliseconds respectively. Rounding 2.5 ms up to 3 ms makes a difference of 12 ms (* 10 lines moved can get annoying). Therefore in the fine tradition of TS3TTY kludges, if the table entry is negative (which used to mean the value given is milliseconds per line to be moved), then it is the number of 1/8 milliseconds per line to be moved. This is MC:SYSTEM;TS3TTY 271, I have updated the terminal entries for the C100 and Teleray. Good luck to the next ITS debugger, why am I so verbose? From HIC at MIT-MC Wed Feb 4 00:00:00 1981 From: HIC at MIT-MC (HIC at MIT-MC) Date: 04 Feb 1981 00:00 Subject: No subject Message-ID: Oh well, I was confused...iot looks like some subtle off by a few error when the MMP almost gets full. From HIC at MIT-MC Wed Feb 4 00:00:00 1981 From: HIC at MIT-MC (HIC at MIT-MC) Date: 04 Feb 1981 00:00 Subject: No subject Message-ID: MC just tried to make an extra MMP page (one more than allowed). MEMFR was 7. Is someone forgetting to decrement this guy? From RWKatMIT-MC Wed Feb 4 00:00:00 1981 From: RWKatMIT-MC (Robert W. Kerns) Date: 4 February 1981, 00:00 Subject: No subject Message-ID: Date: 2 February 1981 21:59-EST From: Ed Schwalenberg To: GNU at MIT-AI cc: BUG-ITS at MIT-AI What is the effect of linking to a free STY (i.e. Finger shows it as "STY not in use")? I saw someone doing it earlier tonight and wondered. It depends on the state of the STY. If you just link to an unused sty, you get a ?, since the tty you want to link to does not exist. What must have happened is that the sty became unused after the link existed, as for instance if someone logs out from a sty and you link to him just before or after the logout (but before the net connection is closed. I believe that the system doesn't really do anything with stys whose only activity is a link, and they will vanish, breaking the link. Actually, this isn't right at all. You can link to a free sty, and slave it, etc. The fact that you have done so is recorded. It doesn't work very well, tho, since there isn't anybody on the other end of the STY to gobble up it's output, so it quicly fills up it's output buffer and can't print anything more. FINGER's 'Sty not in use' merely means that it is unowned. From CENT Tue Feb 3 00:00:00 1981 From: CENT (Pandora B. Berman) Date: 3 FEB 1981, 00:00 Subject: bitten again Message-ID: comsat decided to take another walk over my mail file. today i logged in and typed :rmail, and found myself looking at a message sent to alan and bee. neither of whom i am. this was somewhat disconcerting, so i quit out of the rmail and poked around a bit. by now the rmail had obligingly copied the contents of my mail file into my rmail file and had deleted the mail file, so there was no way of finding out information about it. i have copied the entire bunch of new mail (everything before the first message that i know i have seen in rmail before; i use r-option append and gmsgs so the break was pretty clearly after the newest system msg) to the file ai:vsdb;foo2 mail. i also looked at this stuff: i am sure that the first seven msgs are not for me, and that the eighth (which starts with gubbish and turns into a news digest) probably is. evidence of when i was last seen by the system is a file i wrote the last time i was logged in (i know i read my mail that time); the file was written at 03:38 on 2/1, and i logged out within a few hours of that. the time of the first msg that is probably mine AND has a readable time is 13:35 on 2/1. this is the ninth msg in foo2 mail; the eighth msg (the news digest with gubbish at the front) does not have a readable time. the news digest canonically comes out at 10:00 PST which means 13:00 EST, but the ostats file has been written to disk 15 (third:) which is currently unaccesssible due to dead disk drive, so i can't tell exactly when it came out on sunday or what msgs i got earlier that morning. however, given the several hours between these times, i think that comsat (or something) walked over the first several msgs sent to me sunday morning and replaced them with the random stuff that is now the first seven msgs in foo2 mail. dave moon says that since i logged out sunday morning ai has had problems (disk drive losing causing locked blocks all over, and hic's memory losing in some way (he wasn't sure of the details when he mentioned this) causing programs to act in very random ways) either of which could have caused the lossage i experienced here. i guess i will have to wait until the third: pack comes back online before i can see what i missed; i may not be able to get the stuff i lost back because of the time delay here. From EDatMIT-AI Mon Feb 2 00:00:00 1981 From: EDatMIT-AI (Ed Schwalenberg) Date: 2 February 1981, 00:00 Subject: No subject Message-ID: What is the effect of linking to a free STY (i.e. Finger shows it as "STY not in use")? I saw someone doing it earlier tonight and wondered. It depends on the state of the STY. If you just link to an unused sty, you get a ?, since the tty you want to link to does not exist. What must have happened is that the sty became unused after the link existed, as for instance if someone logs out from a sty and you link to him just before or after the logout (but before the net connection is closed. I believe that the system doesn't really do anything with stys whose only activity is a link, and they will vanish, breaking the link. From GNU Mon Feb 2 00:00:00 1981 From: GNU (John C. Gilmore) Date: 2 FEB 1981, 00:00 Subject: No subject Message-ID: What is the effect of linking to a free STY (i.e. Finger shows it as "STY not in use")? I saw someone doing it earlier tonight and wondered.