info-kermit@ucbvax.ARPA (05/28/85)
From: Frank da Cruz <SY.FDC@CU20B.ARPA> Info-Kermit Digest Mon, 27 May 1985 Volume 2 : Number 30 Departments: C-KERMIT VERSION 4C - More Changes to C-Kermit Version 4C File Sizes Reported by C-Kermit C-Kermit v4 to Apollo C-Kermit 4C Comments/Bugs C-Kermit on TRS-Xenix Kermit 4C Broken "Dial" Command for PC/IX 1.1 C-Kermit 4C (16 May) Local-Remote Bug? Status of 4C on 2.9bsd Wart Mods C-Kermit for Eunice MISCELLANY - Lisp Machine KERMIT Kermit-11 BITnet Distribution: Correction IBM PC/AT Serial Port Compatibility AT Serial Port Compatibility IBM Asynch Adapter at 19.2Kb and Beyond Appeal to the Netherlands Kermit for IAS? IAS AND KERMIT Need Help with Kermit-TSO and 7171 Converting New VM/CMS Kermit to MVS/TSO? Kermit for the NorthStar horizon and USR S-100 modem. Formatting documentation for Xerox 2700 Commodore 64 Kermit Diskette Availability Kermit for Wang VS or OIS? Two Faces of Kermit ---------------------------------------------------------------------- Date: Mon 27 May 85 16:41:12-EDT From: Frank da Cruz <SY.FDC@CU20B.ARPA> Subject: More Changes to C-Kermit Version 4C Quite a few minor changes have been made to the working (but not yet really released) copy of version 4C of C-Kermit. The major functional change is that several of the 'set' parameters have been separated into 'set send' and 'set receive' pairs to allow inbound and outbound packet parameters to be set separately. Some changes have been made to the command parsing functions -- the major result is that the multiline local-mode 'get' command now works as advertised, and in addition can be escaped from if you change your mind in the middle of it. Many minor bugs were fixed, some of which are reported in the messages below. Reportedly, problems still exist on ATT System III and System V based systems in the areas of dialing, modem control, and so on. A special "3bx" makefile option was added for ATT 3B-series systems to accommodate their special handling of uucp lock files. Users of these System III/V systems are encouraged to try out the new files and report problems (and fixes?), so we can finally make version 4C "real". The files are on CU20B, in <CKERMIT>CK*.*, as of 7:30pm EST 27 May 85. Only the Unix and VMS versions have been affected by the recent changes; a new Macintosh release should be ready shortly. ------------------------------ Date: Monday, 20 May 1985 08:26:04-PDT From: d_schullman%sarah.DEC@decwrl.ARPA (Dan Schullman) Subject: File Sizes Reported by C-Kermit When SENDING a file, the displayed size of the file being sent is wrong. It appears that 32kb-->63.99kb, 96kb-->127.99kb, etc. are shown as zero. And that 64kb-->95.99kb, 128kb-->159.99kb, etc. are shown as ( size mod 64kb ). The actual size of the transferred file is fine. [Ed. - The problem is that zchki() returns the size of the file as its value, but zchki() is not declared to be long. Also, it uses a regular int internally to hold the size. The files ckufio.c, ckvfio.c, ckcfns.c, and ckucmd.c have been changed to use longs where necessary, so that file sizes will look right on machines that have short ints.] ------------------------------ Date: Sun, 19 May 85 15:48:00 pdt From: Neal Holtz <holtz%cascade.carleton.cdn%ubc.csnet@csnet-relay.arpa> Subject: C-Kermit v4 to Apollo I finally got around to doing what I said I would in early March - port C-kermit to Apollo (Aegis) environment. I waited for a newer version, 4.2(030). As of this writing, I have only had it running for a few hours, so obviously I have a lot of testing to do yet, but early results are very encouraging. The standard Apollo C-library is very compatible with most Sys III (and some 4.2) -- the only exception is configuring serial lines and the console. I probably could have added about 100 lines to ckxunx.c and ckzunx.c, but I thought they were getting unmanageable with all the IFDEF's, so I made equivalent ckxaeg.x and ckzaeg.c mostly by ripping out most of the compile time control. There were couple of very trivial changes in a couple of other places, but it went very smoothly. When I have had a chance to do more testing, I will forward a few comments. Currently, the TOPS-20 style command parser is being used. That doesn't fit in with the Apollo environment as well as I would like (though it certainly is usable). Haven't decide whether I will retain it. Will probably decide to do whatever causes the least work and disruption (i.e. retain it). Will gladly supply my changes to whoever wants them. [Ed. - The modules other than ckufio (formerly ckzunx) and ckutio (formerly ckxunx) that need modifications to accommodate Apollo Aegis now have them, and the Apollo-specific modules will be announced when they arrive.] ------------------------------ Date: Wed, 22 May 85 09:55:41 -0100 From: hans@oslo-vax (Hans A. ]lien) Subject: C-Kermit 4c Comments/Bugs I had no trouble building v.4c on our vax 11/780 running 4.2bsd unix. However, i have recognized a few problems / minor bugs. I must apologize for not being very familiar with neither unix nor c, so I have no definitive fixes to offer. But anyway: 1) It seems as some terminal output isn't printed to the terminal while the file transfer is proceeding in the background. As a result, the process stops waiting for terminal output. First of all, there are messages of the kind "Warning: ...", which are printed on stderr. I would very much appreciate that such messages be output similarly to messages like "filename => FILENAME ...", WHICH come through nicely. [Ed. - These messages only appear at the start of a transaction, and they are generally important enough that you should want to see them -- like, "Warning, problem getting exclusive access", meaning that somebody else might be using the same communication line at the same time. Once the line is assigned to you, no further warning messages are issued.] The other problem I have noticed, concerns indication of naks. They are indicated by "N", but unlike ".", they demand the process in the foreground to proceed. Once again, I think such output should be allowed to a job in the background. Hopefully, this works OK if -q (quiet) is set, but I haven't tried. [Ed. - I can't reproduce this, nor can I find any code that seems to exhibit this behavior.] 2) I have recognized a minor error in the Tops20-like command parser, which results in repetition of default items by pressing several <esc> characters. Try for example CKermit>log trans<esc><esc><esc><esc><esc>. [Ed. - It was an error, all right. It's fixed now. Thanks for spotting it.] 3) Regarding command line options, my opinion is that a "bye" option should be included, in addition to, or if the number of options should not grow further, instead of the "finish" (-f) option. I think most people use such an option to end a file transfer session, typically performed with a simple list of commands making up a script. Does anyone (dis)agree? [Ed. - The major restriction on the number of command line options is that the command line help message should fit on one screen.] hans@oslo-vax.arpa Hans A. ]lien, Institutt for informatikk, University Of Oslo, Norway. ------------------------------ Date: Tue, 21 May 85 16:04 MDT From: RMark@DENVER.ARPA Subject: C-Kermit on TRS-Xenix I now have kermit 4C(044) running on TRS-Xenix, a Unix v7 system. Both startup and file transfer are much slower than with kermit version 3. The makefile parameters required are: PROC=_proc DIRECT=-DDIRECT NPROC=_Nproc BOOTFILE=/xenix A fix was required in ckutio.c. Robert Mark, USGS, Menlo Park, CA [Ed. - Fix installed in ckutio.c, and the makefile (ckuker.mak) edited accordingly.] ------------------------------ Date: Tue May 21 1985 14:11:11 From: Marco Papa <papa%usc-cse.csnet@csnet-relay.arpa> Subject: Kermit 4C broken "dial" command for PC/IX 1.1 I encountered the following problems with installing and running Kermit 4C for PC/IX 1.1 on a PC/XT with Hayes compatible modem. 1. Makefile problem. PC/IX's "make" does not seem to like \'s to continue lines. Taking out the \'s at the end of lines, and making one long line, fixes it. [Ed. - A warning to this effect has been inserted in the .bwr file, but it's funny we didn't here this one before, especially since a lot of development was done on a PC/IX system.] 2. Dial feature broken. The "dial" command does not work any more. Executing the following set of commands: > set line /dev/tty1 > set speed 1200 > set modem-dialer hayes > dial 743-0000 will work up to the setting of the modem type. The dial command ALWAYS returns with the message: "Sorry, can't hang up the line" I have not looked up in ckdial.c yet, but I believe the message comes from there. I know that there is no problem with my modem switch settings, since Kermit 4.2 runs just fine. Also, since dial does not work, I cannot test the script command. [Ed. - This has been fixed. There are still reportedly problems with DTR during dialing on some systems.] 3. Lock file left around. This is related to the "dial" problem above. When dial fails, the "exit" command leaves the lock file around (/etc/locks/tty1 for example), so that one has to manually delete it, before set line will work again. The lock file is correctly deleted on exit, when dial is not executed. Connect, send and receive seem to work just fine. [Ed. - The ttopen() code in ckutio has been reworked to be more careful about leaving lock files around.] Also, ckuus2.c makes the C optimizer run out of space under PC/IX. Marco Papa UUCP: ...!sdcrdcf!uscvax!papa CSNET: papa@usc-cse.csnet ARPA: papa%usc-cse@csnet-relay.arpa ------------------------------ Date: Sat, 25 May 85 11:29:07 pdt From: rag@uw-june.arpa (David Ragozin) Subject: C-Kermit 4C (16 May) local-remote bug? Running C-kermit, 4C (16 May) on 4.2 BSD Unix. Problem: Issue "set file type bin" Then "send file" and on completion do "show para" mode now shows as "local" instead of "remote" Now when I get mode back to remote by "set line", then do a send of a text file, on return the mode is local. Apparantly going into binary file mode flips some flag which controls the mode setting on returns from at least the send type transaction. [Ed. - Oops - this bug was introduced into 4C (the 16 May version) and has just been fixed. It actually had nothing to do with binary mode.] (By the way, I could find no way to "document" this via any of the log functions, since the "session logging" only works for local mode logging while this bug requires that I record what appears on my screen from a remote kermit.) [Ed. - Try "kermit | tee xxx" -- the transcript will go into file xx. Or use script on systems that have it.] ------------------------------ From: Paul Graham <unr70!unrvax!pjg@seismo.ARPA> Date: 23 May 1985 1218-PDT (Thursday) Subject: Status of 4C on 2.9bsd As I suspected in my heart of hearts the transition from the 7th Ed. is not going to quite as simple as I might have liked. Currently the 2.9 version will not act as server or in the remote mode. Nor will it receive files in the local mode (it will get a file from a server). If anybody has solved all of these problems please let me know. Else sometime next week it should percolate to the top of the stack, and I should have some time for it once again. Please drive carefully. Paul Graham 702/784-6007 University of Nevada Reno UUCP (seismo, ucbvax!menlo70)!unr70!unrvax!pjg ARPA unr70!unrvax!pjg@seismo ------------------------------ Date: Sat, 25 May 85 18:24:36 cdt From: knutson@ut-ngp.ARPA (Jim Knutson) Subject: WART mods The following are a couple of context diffs for getting wart running on an MS-DOS machine with the CI-C86 C compiler. The problems that were fixed had to do with 1) not ignoring trailing text on #else and #endif 2) Stripping comments within quotes and 3) not recognizing \f as a whitespace. [Ed. - diffs omitted, changes installed in ckwart.c.] ------------------------------ Date: Wed, 22 May 85 23:01 EST From: Larry Afrin <lbafrin%clemson.csnet@csnet-relay.arpa> Subject: C-Kermit for Eunice I don't remember if I explained this in an earlier message, but Eunice 3.2 is a hybrid of 4.1bsd and 4.2bsd (with a touch of Version 7 thrown in, from what I can deduce from how the time-of-day functions work), that runs under VMS. A new version of Eunice, version 4.0, is coming out "VERY soon", sort of in conjunction with the brand new release of VMS 4.1. Eunice 4.0 is reputedly going to be a full 4.2bsd implementation under VMS. Now, seeing as how Kermit 4C can be installed under VMS directly, and seeing as how Eunice 4.0 is supposedly going to be 4.2bsd straight up and down the line, you may want to defer from adding any Eunice-specific support into the next Kermit version, which I understand from you will be version 4C. You may especially want to avoid using *my* mods since they're meant for Kermit 4.2, which I think you've indicated is now obsolete, sort of. -- Larry Afrin Dept. of Computer Science Clemson University lbafrin@clemson if you're on CSNet lbafrin.clemson@csnet-Relay if you're on ARPANet [Ed. - I agree with Larry -- Eunice support is probably not worth including. This is very much the same situation as Amdahl UTS, whose new version is pure system V, and whose old version is some kind of hybrid. If anyone is desparate for Larry's Eunice Kermit, they may want to contact Larry directly.] ------------------------------ Date: Fri, 24 May 85 14:53 CDT From: "Mark L. Ahlstrom" <Ahlstrom@HI-MULTICS.ARPA> Subject: Lisp Machine KERMIT I now have the Lisp machine Kermit running "pretty well" on Symbolics Lisp machines. I should be shipping it back to George Carrette at LMI in a couple of weeks to have him retest it on LMI Lisp machines and incorporate some recent LMI changes. Hopefully, the first release to Columbia is within a couple months of being "real". By the way, your information should be updated to show that the lmkermit will be written in Zetalisp rather than Common Lisp. The info that it was Common Lisp evidently came from an earlier plan that George Carrette had to develop the code on another machine. --Mark ------------------------------ Date: 21 MAY 85 10:46-EST From: BRIAN%UOFT02.BITNET@WISCVM.ARPA Subject: Kermit-11 BITnet Distribution: Correction I just noticed that I said UOFT01 KERMSRV in my note about getting the new Kermit-11. It should have been UOFT02 KERMSRV. See Kermit Digests numbers 2-21 and 2-26. ------------------------------ Date: 15 May 85 06:52:09 GMT From: Pat Chewning <patch%nsc-pdc.uucp@brl> To: Info-IBMPC@USC-ISIB Subject: IBM PC/AT Serial Port Compatibility I am having trouble with my AT receiving serial data at baud rates over 1200. I suspect that it has something to do with the incompatibility of the serial controller used on the AT versus the PC. Here's the set-up: IBM AT AST ADVANTAGE! multifunction board that has the serial port. (Uses 16450 chip?) Kermit Software (Also tried-out a copy of Crosstalk with same results) [Both these software packages were written for PC] Using the same physical RS232 line, I have no problem doing data transfers at 9600 baud using a Compaq [IBM PC Clone]. The problem shows up at baud rates greater than 1200 (although occasionally at 1200 baud as well). The characters are not wrong, but the problem is missing characters. Transmitting characters works fine, its only on receiving that I have a problem. The manual that came with my serial card indicates that some incompatibility exists between the serial controller chip used in the AT and the controller chip used on the PC. It suggests that some software written for the PC may not operate properly on the AT. Does anyone know what those differences are, and in particular, what specific changes need to be made to Kermit [I have a source] so that it will work on the AT? Pat Chewning [Ed. - See below.] ------------------------------ Date: 16 May 1985 10:22:14 PDT Subject: AT Serial Port Compatibility From: Richard Gillmann <GILLMANN@USC-ISIB.ARPA> Sounds like you have the Professional Graphics display. This has the unfortunate effect of interfering with the serial ports at speeds over 1200 baud. ------------------------------ Date: Thu 16 May 85 13:57:53-EDT From: Charlie C Kim <US.CCK@CU20B> Subject: IBM Asynch Adapter at 19.2Kb and Beyond Using an interrupt based system, as has been indicated, it is definitely feasible to run the IBM serial card at speeds greater than 9600 baud. (In fact, you'll probably need an interrupt based system at any speeds above 1200 or so). On an AT, I've been using Kermit, which is interrupt driven for the serial IO, for file transfer to and as a remote terminal for a 4.2 Unix system (Vax 750 with DMF's--the DMF is a serial/parallel board for the VAX which has a maximum speed ot 19.2KB on serial lines) at 19.2KB without any problems. I've also tried it at 56KB between two PC's (one AT, one PC-1) and between a PC-AT and Macintosh without any problems. I've also be able to use the COM_PKG/INT_PKG pair in the IBMPC library to communicate from a PC-AT to the above mentioned VAX at 19.2KB; again without any problems. So, empirical evidence supports the results. As as side note, I should note that I've used communications packages which were incapable (on a PC/XT) of going above 1200 baud without losing characters; I suppose these were not interrupt driven. Charie C. Kim Columbia University Center for Computing Activites ------------------------------ Date: Thu, 23 May 85 14:26:27 BST From: Cjk@ucl-cs.arpa Subject: Appeal to the Netherlands To anyone in Holland (= The Netherlands): Is there anyone in the Netherlands running Kermit who could help me by providing a mainframe Kermit for demonstration purposes? This would be to demonstrate that a micro-Kermit actually works. If so, please mail me soonest as "cjk @ ucl-cs", on either ARPA or JANET. Chris Kennington. ------------------------------ Date: 24 MAY 85 14:50-N From: DEGROOT@HWALHW5 Subject: Kermit for IAS? 1. Some users keep asking for a KERMIT version for the operating system IAS on the PDP11. I understand that IAS is an obsolete operating system. Previous KERMIT-versions for RSX could be generated for IAS. The current versions of PDP11 KERMIT give not any notice of IAS. Is there a KERMIT for IAS? [Ed. - There is some hope that Kermit-11 will be adapted to IAS. There is at least one very large IAS site that has indicated some willingness to do this, but there is no definite commitment or timetable -- see below.] 2. The file CURRENT.DOC indicates version 3(124) for K10-KERMIT. The right version number for the DEC10-KERMIT should be 2(124). [Ed. - Thanks for pointing out the mistake.] Kees de Groot (DEGROOT@HWALHW5) Computer Centre Agricultural University Wageningen The Netherlands ------------------------------ Date: 24 MAY 85 09:30-EST From: BRIAN%UOFT02.BITNET@WISCVM.ARPA Subject: IAS AND KERMIT I just got a call yesterday from the EPA about IAS and Kermit. They have some 30 systems running IAS version 3.1 and have volunteered to port it to IAS. The main problem with IAS is that RMS-11 version 2 is only supported on IAS 3.2, which many sites seem to have decided not to install. I expect that these folks at the EPA will have Kermit-11 running on IAS soon, so I would suggest that others interested in IAS Kermit to wait a couple of weeks until I get more info on the progress. ------------------------------ Date: Fri 17 May 85 19:34:39-PDT From: Wing Lee <WingLee%ECLD@ECLA> Subject: Need Help with Kermit-TSO and 7171 I am experiencing a problem with the Series/1 version of TSO-KERMIT. It works fine with the Series/1, but here at USC we are replacing the Series/1 with an IBM 7171. I am having problems trying to upload files through the 7171. I am able to download files just fine, but when I upload, TSO-KERMIT stops at the first bad packet. I used the debug option on TSO-KERMIT and when I looked at the debug file, it shows that TSO-KERMIT stops right after a CHECKSUM error. It appears that TSO-KERMIT is unable to notify the micro KERMIT that it has received a bad packet and have the micro resend it, and since the micro KERMIT hasn't received an acknoledement from TSO-KERMIT telling it to that the last packet was good, a transmission lock up occurs. I've tried everything I can think of to solve this problem, and would appreciate any suggestions you can think of. Thanks Wing Lee University Computing Services University of Southern California [Ed. - This is presumably the Version of TSO Kermit whose genealogy is Columbia VM/CMS Kermit v 1.0 -> U of Chicago MVS/TSO support -> U of Toronto Series/1 support. Anyone else out there using it? Also, see next message.] ------------------------------ Date: Mon 20 May 85 16:26:52-EDT From: Frank da Cruz <SY.FDC@CU20B.ARPA> Subject: Converting New VM/CMS Kermit to MVS/TSO? To: Ron Rusnak <SYSRONR@UCHIVM1.BITNET> Hi Ron -- Do you think anybody at U of Chicago will be working on TSO Kermit any time soon? Ideally, I think it would be good if the new CMS release (2.00, which has server mode, binary file transfer, CRC's, etc) could have TSO support added to it in such a way that either TSO or CMS Kermit could be built from it. I don't know much about IBM assembler, but I would hope that this would be possible using macros or conditional assembly for the system- dependent parts. This would simplify life a lot for both of us, not to mention the hundreds of sites that are running TSO &/or CMS Kermit, allowing each to benefit immediately from improvements in the system-independent areas. What do you think? - Frank [Ed. - I never got a reply to this, but maybe someone else who might be interested in TSO Kermit -- U. of Toronto, maybe? -- would be willing to take on the task.] ------------------------------ Date: Sun 19 May 85 19:49:20-MDT From: Jon Albers <JALBERS@SIMTEL20.ARPA> Subject: Kermit for the NorthStar horizon and USR S-100 modem. To: northstar-users@SIMTEL20.ARPA, info-cpm@AMSAA.ARPA, info-kermit@CU20B.ARPA I am looking for a version of Kermit that will work on a Northstar horizon with either the second printer port, or better yet, the US Robotics S-100 internal modem. If you have or know of such a beast, or can perhaps give some help with writing the code to make Kermit work with an S-100 modem board, please reply to me at the below addresses: Jon Albers ARPA: JALBERS@SIMTEL20 /..seismo!dolqci!irsdcp!albers UUCP: ---------<...seismo!dolqci!irsdcp!dcp1!albers \..philabs!sbcs!bnl!bnl44!jalbers ------------------------------ Date: Sat 18 May 85 23:01:53-PDT From: Bob Larson <BLARSON%ECLD@ECLA> Subject: Formatting documentation for Xerox 2700 I had problems formatting the kermit documentation for our xerox 2700 also. The verbatim portions of the manuals have lines to long to print at 10 cpi with the default margins used for the 2700. By making sure that all verbatim areas were printed at 12cpi, and modifying "italics" to do underlining (the only italic font we have is 10cpi) I was able to format the manual so only a few characters where lost. It might work if formatted with a one character left margin. Hope this helps. Bob Larson <BLarson@Usc-Ecl.Arpa> ------------------------------ Date: Tue, 21 May 85 03:02:17 edt From: Robert Scott Lenoil <lenoil@mit-eddie.ARPA> Subject: Commodore 64 Kermit Diskette Availability Enclosed is the text of my message to USENET, re: my ability to supply Commodore 64 Kermit diskettes this summer. Please update the 00readme.doc file and the information that you give out over the phone to reflect this new status. Note that my work phone number this summer will be (206) 828-8080, effective June 1. Subject: Kermit diskette availability Newsgroups: net.micro.cbm Distribution: net.micro.cbm As you may know, I have been providing copies of Commodore 64 Kermit for $7.00 as a service for those who couldn't otherwise obtain the program. However, I will be leaving shortly for the summer and have decided not to take my Commodore with me. Therefore, I do not anticipate being able to fulfill any orders until my return August 25th. There is always a possibility that I find somebody at work with a 64, so if you do need a diskette, send me netmail, and I'll let you know if I can accommodate you. (Send to this address - my mail will forward). Robert Lenoil lenoil@mit-eddie.uucp lenoil@mit-eddie.arpa lenoil@mit-mc.csnet ------------------------------ Date: Friday, 17 May 1985 14:26-MDT From: ritcv!husky!pma%rochester.uucp@Seismo Subject: Kermit for Wang VS or OIS? Does any one know of a "kermit" for a Wang(tm) VS or OIS system ? Please respond to me by email. Thanks, Philip Abram, Eastman Kodak Co., Rochester, N.Y. PATH: {allegra,seismo}!rochester!ritcv!-------\ >------husky!pma {eagle,astrovax,netword}!sun!sunrise!----/ ------------------------------ From: tektronix!stever@Berkeley Date: Tuesday, 21 May 85 13:01:22 PDT Subject: Two Faces of Kermit Kermit is both a protocol and a communications program. One should not ridicule people for their confusion about this point. All communications programs whether commercial or not should be able to do *raw* (i.e. ASCII) data file transfers using system level commands of the computer they are talking to, as well as supporting one or more file transfer protocals. Kermit should not be an exception. This feature should be a part of the Kermit protocal definition. Kermit implementations should not be included in the distribution system of Kermit software unless they have the *raw* file transfer function. steven d. rogers ...ucbvax!tektronix!stever [Ed. - In principle, I agree. In practice, Kermit programs are donated by individuals who are free to implement whatever features they like, so long as the minimum protocol features are present. We're not going to turn down desparately needed contributions (we know when a Kermit implementation is desparately needed when the number of phone calls asking about it exceeds a hundred per day) because they lack a particular optional feature.] ------------------------------ End of Info-Kermit Digest ************************* -------