SY.FDC@CU20B.COLUMBIA.EDU (Frank da Cruz) (12/11/85)
Info-Kermit Digest Wed, 11 Dec 1985 Volume 3 : Number 33 Departments: MS-DOS KERMIT 2.28 JRD - MS-DOS Kermit 2.28 jrd Not on BITnet MS-DOS Kermit 2.28 jrd Bug with X Packets Bug Report on MS-DOS Kermit 2.28 jrd on 18MHz PC/AT Suggestion for MS-DOS 2.29 Minor Mod to New MS-DOS Kermit MISCELLANY - Apple-II Kermit Bugs Multics Kermit ---------------------------------------------------------------------- Date: Wed 11 Dec 85 10:13:20-EST From: Frank da Cruz <SY.FDC@CU20B.COLUMBIA.EDU> Subject: MS-DOS Kermit 2.28 jrd Not on BITnet In response to numerous requests from BITnet for Joe Doupnik's improved version of MS-DOS Kermit 2.28, I'm sorry to say I can't put it on BITnet just yet, for a variety of reasons which are too complicated to explain. It looks as though this version will wind up replacing the current one -- one or two minor bugs need fixing, plus checkout is still required on the Wangs, HP's, NECs, etc -- and at that time it will replace the current version on BITnet. I'm also sorry I mentioned the MS-DOS Kermit 2.27 that had VT100 support added to it. In response to numerous requests for it, let me just say that in order to avert utter chaos, I would prefer to get the fixed 2.28 installed as the standard, base version, and then call for volunteers to take the VT100 support and fit it into the new base version. If we don't do it that way, we'll have multiple proliferating divergent versions of the program which will be impossible to support. Those of you who whose phone doesn't ring 500 times a day with MS-Kermit questions might not appreciate why this is important, so you'll just have to take my word for it. Once again, I apologize for the reduced level of support for MS-DOS Kermit and the BITnet Kermit files. Like I said before, it's because the people who used to take care of these things have left Columbia. They will be replaced, but I can't say when. Soon, I hope. ------------------------------ Date: Tue 10 Dec 85 18:02:07-EST From: Frank da Cruz <SY.FDC@CU20B.COLUMBIA.EDU> Subject: MS-DOS Kermit 2.28 jrd Bug with X Packets First real bug -- if you send a generic command to a server which the server does not support, and it sends back an error packet (which is the proper behavior for servers under the circumstances), then the next file sent by MS-DOS Kermit has its name in an X packet rather than an F packet, which of course the server is not prepared to handle. MS-DOS Kermit is apparently setting its "X flag" in anticipation that the generic command will elicit a X packet or a data-bearing ACK from the server, and then not resetting it when the transaction is terminated by an error. In fact, it should not set this flag until such a response actually occurs, and it should reset the flag at the beginning of each transaction. Reportedly, the same behavior crops up randomly (not reproducibly) at other times. Thanks also to others who reported this problem. If anybody develops a fix (it should be real simple, something like setting the X-Flag to 0 at the top of the command loop), please send it in. ------------------------------ Date: 9 Dec 85 08:57:00 PST From: ALEX WOO <wu@ames-aero> Subject: Bug Report on MS-DOS Kermit 2.28 jrd on 18MHz PC/AT The new MS-DOS kermit seems to work fine on a 4.77MHz PC but it coughs on a 18MHz PC AT with the error message "? Not enough memory to run kermit" The PC AT has about 2MB of memory. [Ed. - Beats me, anybody have any ideas? Doesn't sound like anything that would involve timing loops... Do all your other programs work on your souped up AT??? Does the old Kermit?] Also, if there a Makefile for the MS-DOS version of Kermit, perhaps using one of the Info-IBMPC makes. [Ed. - There isn't, sorry. Does anyone know if there is a "make" that will support command line arguments, so you can say "make ibm", "make rainbow", "make tipc", etc? If so, which one is it, where is it?] Alex. ------------------------------ Date: Mon 9 Dec 85 04:18:07-EST From: Joe Smith (415)794-2512 <LSM.SMITH@MARLBORO.DEC.COM> Subject: Suggestion for MS-DOS 2.29 If you do plan to use MSKERMIT 2.28 jrd as the basis of the next release of MS-DOS KERMIT, please add "SET TERMINAL mumble" and "SET DTR mumble" to the system independent code. Define dummy entry points in all the system dependent modules so they can parse "mumble" or say "not implemented". Then the developers could add system specific arguments to these commands, such as: SET TERMINAL NO-WRAPAROUND SET TERMINAL WIDTH 132 SET TERMINAL ID VT125 SET DTR ON SET DTR OFF SET DTR HANGUP-ON-EXIT We need a general way to get to system-specific functions thru KERMIT's command scanner. /JMS [Ed. - These sound like useful hints for whoever is going to volunteer to fit the VT100 support into 2.28 jrd.] ------------------------------ Date: Sun, 8 Dec 85 01:23 EST From: Larry Afrin <lbafrin%clemson.csnet@CSNET-RELAY.ARPA> Subject: Minor Mod to New MS-DOS Kermit The new MS-DOS Kermit (jrd's rewrite/mod of 2.28) has a minor problem when it invokes COMMAND.COM to do stuff like TYPEing out files, DELeting files, running CHKDSK, and RUNning any other program: when you type in the command line in response to the "Kermit>" prompt, and then hit Enter, only a carriage return is echoed. No line feed. This means that the first line of output (either from COMMAND.COM or from the program loaded in by COMMAND.COM) overwrites the "Kermit>" prompt and the command you typed. Trivial, I admit, but aesthetically displeasing. I made the following mod in the location of the "run3" label in the MSKERM.ASM module in PS:<KERMIT-MS> to output the needed line feed. [Ed. - I've appended your code to the MSKERM.BWR file, but I can't reproduce the problem, either on a PC/AT or a Rainbow.] Other than this minor problem, I haven't found anything wrong yet with jrd's version. I'm running PC-DOS 3.10 on a plain vanilla PC. Still waiting for a VT100 emulator to replace the Heath-19, though... (hey, we can dream, can't we?) [Ed. - Yes, but can we volunteer?] P.S. I'm running MASM version 1.00 (from the Dark Ages of 1981), and in order to get the segments in the right order for the linker, I had to change all references to segment name STACK to segment CTACK. These references are in modules MSKERM.ASM (4 references) and MSXDMB.ASM (2 references). If you don't make this change, the machine winds up taking a long walk off a short pier when you hit Control-] C to come back from on-line mode to command mode. [Ed. - Anyone else have this problem? I assembled the files with Microsoft MASM 1.10 (1981,82) using no switches, and not renaming any segments, and the result worked fine. Also, has anybody ever heard of the /DOS switch that Joe mentioned in his letter?] ------------------------------ Date: Fri 6 Dec 85 01:25:39-EST From: Peter G. Trei <OC.TREI@CU20B.COLUMBIA.EDU> Subject: Apple-II Kermit Bugs In reference to Sam Lam and Bruce Jolliffe's fix for the Apple ][ Kermit's 8th bit quoting bug, here is a way to install it in version 2.1a without doing another download. I have not yet had the opportunity to perform thorough checking and testing of this fix, so be cautious! [This patch only works for version 2.1a] Make a copy of your kermit-65 disk. Boot with the new disk. Enter the following DOS commands: ] UNLOCK KERMIT ] BLOAD KERMIT ] POKE 17303,35 ] POKE 17317,21 ] POKE 17329,9 ] POKE 17399,194 ] DELETE KERMIT ] BSAVE KERMIT,A$801,L$4E9B As soon as I have finished checking this fix, I will ask Frank to put the updated source and hex files into the download area on CU20B for distribution. This may be related to the bug that strips out '&' on reception of files. Peter Trei oc.trei@cu20b ------------------------------ Date: Fri, 6 Dec 85 17:20 EST From: "John C. Klensin" <Klensin@MIT-MULTICS.ARPA> Subject: Multics Kermit As of approximately now, the 'real' Multics kermit is the one with which problems were reported in V3#32. It is a version created at Calgary and being distributed with the system and supported by Honeywell (it is, consequently, unlikely to be released to the Columbia archives). It supports many options that the various older versions and kludges do not, incluing server mode, 8th-bit-quoting, etc., and is structured much more like other Multics commands that do similar things, such as the Internet FTP command. Its default option settings, however, tend to be a little more optimized for Multics<->Multics transfers over X.25 and dial circuits than for PC->Multics or Internet TAC use. Users with problems should probably bug Honeywell over conventional channels, as this is a supported product. Does that help? [Ed. - I guess. Columbia, by the way, has never received this version of Multics Kermit. I suppose if all Multics sites get it automatically, no harm is done.] ------------------------------ End of Info-Kermit Digest ************************* -------